The Austrian programmer and entrepreneur Peter Steinberger spent 13 years building PSPDFKit before selling it and swearing off ever doing another company. Then agents got good.

Peter’s now known throughout the world as the creator of OpenClaw, the open source AI assistant that runs locally on your machine and talks to you through chat apps. Peter now runs the OpenClaw foundation and works at OpenAI, where he’s shaping the future of personal AI agents.

Last week Peter sat down with a16z speedrun’s Andrew Chen in front of a room full of speedrun founders to talk about what building looks like when the agents write the code.

Watch our full conversation with Peter above, or check out the highlights below:

1) How Peter’s work with agents has evolved

Andrew’s first question to Peter was about how his own usage of agents has been evolving over the last year. His response:

“Agents changed so much. When I started last year, I got excited when they got something right. And now, if something doesn’t go right, I question myself. I think we are very much at the point where you don’t have to write code yourself anymore, but it’s very important to think about: what can I do so my agent can do its best work? How can I design the – call it loop, or graph, or whatever – so that the agent not just writes the code, but then also can review the code until it’s right, and also has all the tools so you can test everything end to end?

“Ultimately you design a pipeline from ‘this is what I want’ to ‘this is the outcome,’ and the better you do that, the better your outcome will be. So this is now part of the role of software engineering: engineering how my agents can do their best job.”

Peter says one of the big mysteries he’s been struggling with is why some people seem to get so much more out of working with agents compared to others:

“Especially last year, there were still so many people AI wouldn’t work for, and there were other people who were very successful. In my head, both existed, and it took me a while to understand why that’s the case – and why I’m in this group and not that group. I think a big reason is because I ran a company before.

“When it transitioned from ‘I write the code’ to ‘I command a group of people who write the code for me,’ I had to accept that the code will not be as perfect as I would write it. I could micromanage them, I could breathe down their necks, and then we would have a really bad time, and they would leave. My agents would not leave, but we’d still have a very bad time. So it’s just accepting that: okay, maybe it’s not exactly how I want it, but maybe that shouldn’t matter anymore. Does it get me to my goal? Yes. Maybe the agent does it in a way I’m not a fan of… but again, my job is to optimize for the agents. I’m not going to read all the code.”

2) How Peter sets up his own loops

Andrew prodded for more details on Peter’s own workflow: “Maybe you can say a little bit concretely about how you set up projects for yourself now,” he asked. “What’s a good loop that you’ve engineered?”

“Even six months ago, when you had a new task, you would clear the session, do the task, and then clear the session again, because the model would degrade, and so on. People already tried orchestrating their sessions, but it didn’t work so well. Now that’s just my default.

“I have so many open source projects, I can’t look at them one by one. There are some bigger ones where I do, because I want to try certain ideas, but the others are basically handled by my open source orchestrator: any time an issue comes in, it already builds a PR and tests it. So when I look at it and I need to make decisions, I’m kind of like the executive. I come in and ask my agent to walk me through decisions, it presents me ideas or problems, and then I give the direction and it goes off and works with the others. It automatically does the releases. That doesn’t work great for everything, but it works great for a lot of the noise… a lot of the little things that don’t really require my attention anymore. And then for important features, I just pair one-on-one with an agent.

“Think of it like interacting with a different team, or a company: usually you interact with the manager, but sometimes something’s important and you pair directly with one of the engineers. That’s how I see my army of agents.”

3) On slop and keeping your typos in

Of course, there are still some hard limits when working with AI, and one Peter says he’s become extremely annoyed by is the inability of AI to write in his own voice. In fact, he’s irritated when others send him AI-written messages as well:

“No matter what you work on, invest in your personal brand. Because in this day and age, the hardest thing is actually getting eyeballs. You have no clue how many people every day try to send me their slop – half a page plus a link, ‘please look at it.’ I see those em-dashes, or all these telltale signs of AI-speak, and I’ll just block you. Sometimes I reply ‘this reads like slop,’ but that’s the only reply you’ll ever get.

“So be very careful about trying to use AI to get your message out, because a lot of people will be triggered. And even more so: if you write something for your website and you want to use AI to make it more professional, it will add those telltale signs. And then people don’t know if your blog post is actually very thoughtful and just a little bit rewritten, or if it’s just a three-word prompt that generates two pages. How could they know, right? As soon as they smell it, people automatically, subconsciously or consciously, degrade the value of it.”

Ultimately, Peter advises, if you want attention, be human. Be different. Show proof of effort, even if that means leaving the typos in.

4) On the challenges of maintaining an open source project

Filtering messages isn’t the only problem Peter has to deal with when running an open source project like OpenClaw. Andrew asked how he handles the influx of ideas from users who want to help make the software better as well. “Do you have the same kind of filtering process for what to take seriously or not,” he asked, “given that there’s an infinite amount of code being generated these days?”

Peter’s response:

“I would say that’s one of my mistakes on OpenClaw as well: saying yes to too many things. It’s so easy to just prompt a new feature into existence. And the feature is never the problem. The problem is the 19 other features that have to interact with that feature, and thinking about how this all fits into your system. That’s such a complex problem that your agent will not be able to help you with it. Agents go in every time knowing nothing about your codebase, and they find their way – they can navigate large codebases very well. But having this understanding of the system – does this really make sense? Would this contradict that thing over there? – that’s a bit beyond what the agents are really good at.

“And it also kind of sucks when people put a lot of time into a PR and then you just say no. But then again, you don’t even know, right? Did people put a lot of time into their PR, or is it just a three-word prompt and the agent put a lot of time into the PR? So again: if you want attention, be human. Be different. For once, I would love a PR that has typos, that’s handwritten. Or submit your logs, so I have proof of effort.”

5) Building for your critics breaks the product

As with any tech product, OpenClaw has drawn its share of criticism and controversy. Peter said the sheer scale of the reaction to OpenClaw at one point led him to learn a lesson about the importance of building for yourself instead of the critics:

“The thing that had a really big emotional impact on me was all the security blog posts, and then the security reports. So many overblown, fabricated claims, with everybody optimizing for clicks… But I suddenly felt this urge of responsibility: oh, now so many people use it, so I need to work really hard to make it secure. And then I worked really hard to make it secure, for months. Basically, that was all I did. And it really damaged the product. At some point, I stopped making something for myself and I started making something for other people. I put things into place that I didn’t really agree with, or I tried to change what it’s for, to be for a different use case… even though I don’t use it like [that]. And of course, then it breaks.

“A lot of the features that were added were making it slower, making it harder for people to use. Definitely safer. But I lost the fun. That’s probably the key thing. If you build something: number one, you should use it yourself, otherwise you’re building the wrong thing. Number two, you should have fun. I always do my best work if I enjoy what I do. And then I can work much harder, longer, and it takes me less energy. But this thing was incredibly draining. I was not enjoying it at all. ... I kind of lost the joy, and then that reflected in the product. Eventually, we got it right. Now there’s a team; now we actually do things the right way.”

That’s it for this week! Huge thanks to Peter for his candor. For more weekly dives into the world of early-stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Over 30,000 teams applied to join the latest class of a16z speedrun, and just over 80 startups were ultimately selected.

This week, the founders of the class of SR007 gathered together for the first time in Los Angeles. The rest of the program runs in San Francisco, but we wanted kickoff week to feel like a retreat before the real work began.

“It’s always special to gather somewhere that isn’t home base for most people,” says Joan DeGennaro, who runs community and events for speedrun. “I think it’s really important to step outside your normal routine and use a kickoff like this as a marker of a new beginning for founders to be fully immersed in it.”

This kickoff week had a more laid-back vibe than SR006’s, a result of some recent tweaks made to the program. After every cohort we dig into founder feedback to see how we might make the speedrun better. The biggest note, by far from the class of SR006 was that founders want more unstructured time with each other.

Samira Behrouzan, Marketing and Ops Partner on speedrun, explains: “Founders value getting to know their cohort above almost everything else, and we’ve seen it over and over that when founders connect early, their whole program experience is better.”

So we cut the programming down. Three packed days became two and a half. The team left more room to breathe in the calendar, with longer lunch blocks so founders could actually sit down and get a meal together (or take a sales call).

And then, instead of two or three fireside chats with different guests, we kept it to just one memorable chat and Q&A with a very special guest.

Marc Andreessen came by for an in-person conversation about how AI is changing the world. He spent most of the session doing Q&A with founders.

“Most founders have seen Marc’s content on YouTube or read nearly everything he’s written,” Joan DeGennaro says. “So was cool to get Marc chatting with the founders about the questions on their minds.”

The extra breathing room did what it was supposed to do.

The social event on night two had an official end time, but many founders lingered well past, swapping stories and intel over drinks and games of Jenga.

Founder introductions are always a highlight. Each company sent one representative to introduce themselves on-stage, in a sort of early, soft-preview of “Demo Day.”

Samira Behrouzan says that what stuck out the most was the variety: “We have founders building satellites that send real-time intelligence from space, using sensors to predict engine failures on cargo ships, and working on missile defense,” she says. “We even have a team turning industrial wastewater into ultrapure water for datacenters.”

On day three, founders wrapped after morning cohort breakouts and got back to building. That’s the mantra for the next 12 weeks: heads down, with regular office hours with cohort leads, and tactical sessions with our operating team.

The next big milestone is week six’s Progress Day: the midpoint of the program and a major milestone, where founders present their companies in a structured format.

The goal for the weeks in between is to help them refine every aspect of their thinking and approach, give them enough reps to pressure test their vision, and help them build meaningful traction along the way.

There will also, of course, be more surprise guests (next up: OpenClaw creator Peter Steinberger).

And that’s a wrap on week one. We’ll bring you more inside looks between now and Demo Day.

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

7️⃣ Our latest batch of speedrun, SR007, kicks off this week! The team is sharing some behind-the-scenes looks.

🏆 Andrew Chen lists everything AI startups are competing against (hint: it’s not other AI startups)

🗺️ Bella Nazzari shares some of her favorite network maps

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

At some point, during each speedrun batch, every founder gets the PR itch. They ask: When do I talk to the press? Or, more bluntly: How do I get a tier-1 publication to write something nice about us?

The idea underlying the question goes something like this: we’ll launch, the launch will generate a wave of coverage, that coverage finds us an audience, and that audience will become customers. There’s a vague sense that getting written about is something that happens to companies that are working.

In his role on speedrun, Lester Chen has worked with hundreds of founders launching new companies. And he spends a lot of that time helping them understand a more practical way to think about PR and media relations.

Below are the tips he shares most often with actual a16z speedrun teams.

What PR is actually for

Lester starts by reframing the purpose of PR. Instead of thinking of press as a distribution channel, think of it as a test.

“PR for early-stage companies is the crucible through which founders discover if others will find value in telling their story for them,” he says.

You have a story about why your company exists and why now, and pitching it to journalists is the fastest way to learn whether that story survives contact with someone who has no reason to be nice to you.

Coverage delivers “outside validation and some form of reach,” as Lester puts it. PR is valuable, he says, as “a forcing function that aligns founders and locks in a point of view. It isn’t a feature set list, a market observation, a preachy set of beliefs, but a packaged message tailored for an intentional audience,” Lester said.

Targeting the right audience

The “intentional audience” part is one of the hardest parts of the test. Many founders default to assuming that their audience is “tech Twitter,” when their customers are actually a much more specific slice.

“Some AI companies are also selling into older industries with customers who have never heard the term ‘new media,’” Lester says. “This rings even truer internationally. The value of a viral Twitter post or TechCrunch feature holds less weight than a story a manufacturing plant owner might read in a national publication during their weekend breakfast.” The former gets you attention, the latter might “drive new inbound or firm up in-flight contracts,” he says.

The inverse holds, too. If you’re building a consumer product for Gen Z, Lester’s read is that a Forbes feature is close to worthless, and that the same energy is better spent on direct founder marketing or creator and UGC channels.

Then there’s the reason a lot of PR happens in the first place: fundraising. Lester argues that announcement coverage carries “immense strategic value for signaling to investors as well as helping in the hiring process, especially if immigration considerations are involved.”

We’d add one more that has gotten harder to ignore: earned media is what the models read. When a customer asks ChatGPT or Claude about your company, the answer is assembled largely out of what credible publications have said about you. Call it SEO, call it GEO. Either way, a good article now has a long second life.

Don’t hit the NOS too soon

When asked about the mistakes he sees most, Lester led with a metaphor we are contractually obligated to print in full:

“I would say the most common mistake I see early-stage founders making is hitting the NOS too soon. (Sorry I just watched far too many Fast & Furious movies on vacation and the analogy is just too good.) Hitting the NOS too soon means going through a premature process when so many variables remain vague, which ends up burning time and money in the process. They think PR will be the PRIMARY unlock for customer acquisition and forgo considering any other alternative. As crazy as it sounds, I hear it a few times every batch.”

There are signals that members of the media look for when deciding whether they care enough to write about a startup, and it’s possible to know whether you meet the bar. Lester lists off the negative signals he sees most often:

“There isn’t a fundraise event, PMF is still questionable, CAC is high while retention remains volatile, product still isn’t publicly launched, the list goes on. Don’t get me wrong, there are instances where articles have been written about companies with these issues, but there is usually some outlier catalyst which triggers this: a stacked team with shiny pedigrees, a viral moment worth talking about, etc. For most, simply raising some pre-seed money isn’t enough. The plot needs to thicken.”

The other trap founders often fall into is simply launching without professional help:

“A second mistake I notice is founders who balk at the idea of getting help from PR consultants or firms,” Lester says. I’ve seen a lot of ‘I’ll do it myself’ scenarios that end with botched pitches, limited coverage, and founders thinking they ‘got PR’ when publishing a GPT-written press release to a newswire.”

There is, Lester says, “no shame in forking over some amount of money to get help with messaging and running a proper process.”

How the playbook is changing

Ask what’s different now versus two years ago and Lester doesn’t hedge: “It has to be founders just going direct.”

The phrase belongs to Lulu Cheng Meservey, who ran communications at Substack, then Activision Blizzard, and now runs her own firm. For years she has argued that founders should stop laundering their message through reporters and simply tell their story themselves. Her “go direct” manifesto has since hardened into something close to conventional wisdom.

The problem, of course, is that many people feel profoundly uncomfortable acting or speaking as public figures.

“For better or worse, it’s almost expected that founders build in public or develop their own online following,” Lester says. “I say worse because most founders suck at doing this or feel deep discomfort when starting. Going direct to your 426 LinkedIn followers feels brutal especially when you see a 21-year-old founder crushing it by making a video every day.”

But the trade is still worth making. “The reality is that there is far less gatekeeping and founders have WAY more channels to generate awareness and find their customers than ever before,” Lester says.

This lands us right back where the whole thing started, on the audience question. More channels does not mean every channel is your channel.

“If you’re selling into Generals at the Department of War,” Lester says, “no amount of LinkedIn thinkpieces will get you a meeting.”

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

FYI - exciting news: we’re running an early acceptance window for SR008 over the next few weeks, ahead of our main application window this fall. If you know a great founder we should back, please encourage them to apply and then send us a referral note at sr-team@a16z.com so we can prioritize their application.

The best intros come from people we trust. That's you! If you believe in a founder, tell us and we'll take it from there.

A few months ago, while reviewing applications for SR007, we asked investors on the speedrun team a question: if a close friend or family member told you they were thinking about raising VC, what would you tell them?

Troy Kirwin said that he’s already approached many of his “exceptional” friends about founding, but then sounded a note of caution:

“I have a great friend who’s running a growth marketing agency. He’s reading all these blogs and wondering if he should be raising. My answer to him is: you have a great, profitable business pumping a million in cash each year, and you’re working with great people. You have your own family and you get to spend time with them. Venture capital is not for everyone.”

– Troy Kirwin on “Advice for applying to a16z speedrun”, on YouTube

Troy Kirwin’s whole job is finding companies to fund, yet he sometimes tells the people he cares about most to reconsider. And he’s not the only one.

While we believe in the power of VC funding to massively accelerate growth for startups, the best investors try to be honest about the types of companies VC funding is and isn’t good for.

So how do you know whether VC is right for you?

First, what venture capital actually is

In January 2019, Josh Kopelman, co-founder of First Round Capital, the first seed investor in Uber, tweeted this extremely succinct description of the product he sells:

“Motorcycles are common (2018: Honda sold 18M motorcycles). Jet planes are rare (2018: 806 Boeing planes). VCs sell jet fuel, which doesn’t work in motorcycles. Bad stuff happens if VCs push jet fuel on a bike owner. Or if a bike owner thinks they can fly.”

He put it even more simply in the New York Times that same day: “I sell jet fuel, and some people don’t want to build a jet.”

The metaphor is obvious: A profitable company doing a million a year in cash is a fantastic motorcycle. But if you fill it with jet fuel, it will explode.

Eric Paley, managing partner at seed firm Founder Collective offered another hardware-based analogy:

“Founders need to think of venture capital as a power tool — a fairly dangerous one — but instead often mistake it for some magical, infinitely renewable resource. In the right hands, power tools can solve some real problems. Used incorrectly, they can chop off your hands.”

The thing you’re signing up for

Fareed Mosavat from the a16z speedrun team says that every founder should ask themselves two questions before they begin: “Should I start a company and should I raise capital for this company?” The second question is one that trips people up most often.

“You are signing up for a very specific kind of company the minute you raise any capital. Are you sure you want to be signed up for that?”

Signed up for what, exactly? Paul Graham spelled out the mechanics twenty years ago:

“Younger would-be founders are often surprised that investors expect them either to sell the company or go public. The reason is that investors need to get their capital back. They’ll only consider companies that have an exit strategy—meaning companies that could get bought or go public.”

The expectation of an eventual exit is one thing to keep in mind. Even tougher is the journey to get there. Venture-backed companies are expected to grow, and preferably quickly. Those on a slower trajectory or with a lower Total Addressable Market find it harder to raise followup funding rounds.

The ego tax

So why do founders sometimes go for VC funding, even why they don’t need it? Emily Bennett shared the following observation:

“There’s an ego element to VC, especially for young founders who think, ‘I need VC to show up credibly in a room.’ There are a lot of businesses that actually don’t need that form of acceleration. The expectation is you’re growing at a certain rate. The expectation is that you can meet this type of market demand, so it is a different kind of structural environment, which isn’t for every kind of business.”

One all-time counterexample to the need to fundraise for credibility is Mailchimp. Ben Chestnut bootstrapped it to roughly $700 million in annual revenue (“give or take,” as he told TechCrunch in 2019) before selling to Intuit for about $12 billion. His reasoning for never taking outside money:

“Today, I have two constituents to worry about: customers and employees… To have a third called investors? No, I can’t do that.”

You need less than you used to

The macro-economic environment is changing the math on this question, too. Per Fareed:

“Now more than ever… you can do more as a small team. You can do more without capital. If you were trying to start almost any company a decade ago, unless you had a bunch of money to self-fund, you probably needed cash just to put a server in the closet. That’s not the case now.”

The evidence to back this up keeps arriving. Midjourney has been turning away investors since 2021. The Information reported VCs “practically begging” David Holz to take their money, chasing warm intros and preemptively sending term sheets, while he built a nine-figure-revenue company with about 40 people and said no.

And Sam Altman, who has raised more money than almost anyone, told Alexis Ohanian:

“In my little group chat with my tech CEO friends there’s this betting pool for the first year that there is a one-person billion-dollar company. Which would have been unimaginable without AI and now will happen.”

If a billion-dollar company can be one person, then “do I need $5 million and a board to find out?” deserves a real answer before you send the deck.

So… should you raise?

If we’re being honest, on some level all this anti-selling is a filter. If one honest conversation with an investor talks you out of raising, the product was probably never for you.

But if you know what the fuel costs, and you’re certain you’re building a jet, we’ve got a link for you: Apply to a16z speedrun

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

🚀 Jon Lai explains how “the best founders today launch in micro beats.”

🕵️‍♂️ Most jobs aren’t posted on LinkedIn. Here’s how to find them.

🖋️ Emily Bennett has some tips for startups working with advisors.

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Today’s essay is a guest post by return contributor Evan Armstrong. Evan is the founder and CEO of The Leverage, a newsletter about how power actually moves in tech.

Evan argues for a reassessment of solo founders, especially in light of new data suggesting that some AI-assisted solo founders are making it bigger than ever before.

For as long as I have been in startups, the first maxim every founder was told is a Zelda quote: “It’s dangerous to go alone!” Cofounders were portrayed as a necessity to build a billion dollar company. Where would Jobs be without Woz? Page without Brin? Simon without Garfunkel? Investors, including the very ones who own this publication, prefer backing up the Brinks Truck when there is a co in front of the founder title.

This idea has never made a ton of sense to me. Pound for pound, the hiring of a cofounder is the most expensive decision you will ever make in the decade long journey of starting a tech company.

For many years, this was a shameful secret opinion of mine. However, we are starting to get data that people are coming around to my view. Solo founders made up 63% of the C corps formed through Stripe Atlas in Q2 2026, an all-time high. There are entire venture programs devoted exclusively to solo founders. Clearly, there is something happening here. But what? And why? And perhaps most importantly, does this trend mean you can keep your company to yourself?

What is happening

My first instinct with a number like 63% is to distrust it. A solo-founder spike looks exactly like what you would get from a flood of throwaway LLCs spawned by ill-considered weekend vibe coding projects. Fortunately, we have an easy signal: these solo founders are making a shit ton of money much more quickly than previous founders. The share of businesses crossing $1 million within a year of going live on Stripe ran about 30% higher for the 2025 cohort than for 2023, and roughly three times higher than for 2019. In Stripe’s proxy index, the share of solopreneurs clearing the highest income thresholds has roughly doubled in the last two years.

This data suggests that the companies being started are of a higher quality, on average, then previous generations of startups.

Why this is happening is AI (duh)

Companies needed cofounders due to skill issues. No one can do it all, so having someone there to help do the things you are weak at is enormously valuable. As AI models get better, they have a deflationary effect on complimentary, generalist skills. These tools enable a new way of work. Call it the revenge of the idea guys.

We also have academic research that shows what happens when a startup fully adopts AI into their operations. A Harvard and INSEAD experiment put 515 high-growth startups in a three-month accelerator. Half the firms got standard entrepreneurship training. The other half (“treated” firms, per the study’s parlance) got case studies showing how AI-native companies have reorganized production around AI. The results were crazy,

“Treated firms complete 12% more tasks, are 18% more likely to acquire paying customers, and generate 1.9x higher revenue. Revenue and investment gains are largest at the 90th percentile and above, consistent with AI expanding the upper range of what firms achieve rather than modestly improving marginal ventures. Despite faster growth, treated firms do not scale inputs proportionally. Their demand for external capital investment falls by 39.5% relative to the control group, while their demand for labor remains unchanged.” [Emphasis added].

—Kim, Hyunjin and Kim, Dahyeon and Koning, Rembrand, Mapping AI into Production: A Field Experiment on Firm Performance (March 30, 2026).

This is about as definitive of proof as we will ever find that a company operating in an AI native way is materially different then its peers. Nearly 2x the revenue with 40% less capital requirements is what we professional analysts would call a whizpopping, awooooga type scenario.

Still, it’s not like someone can sign up for a ChatGPT account and auto escape the permanent underclass. Instead, it is a form of Leverage for the most skilled among us. Stripe data found that the top decile out-earned the median solo founder by about 34 times in their first six months in 2022. By 2025 that gap had stretched to 61 times.

The data does not say that solo founding should be the automatic answer. By month 24 since founding, top-decile multifounder startups generate 53% more revenue than top-decile solo founders. Which intuitively makes sense. The only thing better than one founder with AI is two.

Beyond that, the true argument for a cofounder isn’t merely skills, it is something more fundamental.

What the people writing the checks actually believe

a16z speedrun does not require a cofounder, but Andrew Chen is direct about why the team watches whether you have one anyway: convincing one smart person to join you “is probably step one in validating that it’s a serious effort, idea, or team,” he says. “So we do look at it as a signal.” Recruiting ability is the thing that matters. The cofounder is just the strongest signal of that.

Similarly, speedrun investor Emily Bennett has watched the solo wave build and is openly calibrating against it. She is seeing a surge of founders who think that “because agents are more outcome-focused, they can actually operate solo.” The value, she argues, is having someone “to push back on your ideas, that foil is what ends up being valuable.” Call it the 4o syndrome—a time where you become utterly convinced by an LLM that you are the second coming of Product Jesus. AI agents are trained “to self-perpetuate your own thinking,” so “you actually want a counterpart...who’ll tell you when you’re being stupid.”

So, should you?

Going solo by default is now a defensible choice. If you are immensely talented, you do not need a cofounder to be taken seriously. However, a cofounder who is truly elite, and gives you emotional and intellectual partnership in a way that AI can’t, is perhaps more valuable than ever. The issue is 1) those people are incredibly rare and 2) why the hell would someone that great want to partner with you?

Recognizing the value of cofounders is not a license to grab the nearest warm body and hand them half the company. The most expensive mistake you can possibly make is bolting on a cofounder because you are “supposed to.” If you do bring someone on, the bar goes up, not down. And do not confuse staying solo with working alone: the best solo founders are recruiting machines, pulling in early hires and a network easily and early.

For the rare right person, the answer to the “should I have a cofounder” question is an emphatic, bet-the-decade, yes. But for other cases? It’s more of an open question.

That’s it for this week!

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

🌉 Submissions for SF/LA Tech Week 2026 are open! Rose Johnson has the details.

❤️ Clair Health (SR006) announced they’ve raised $11.6M to build the world’s first continuous hormone monitor.

🇸🇪 speedrun is coming to Stockholm this week! Kenan Saleh shares how to join us.

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Every founder thinks their challenges are unique, until you’ve seen a hundred startups make the exact same mistakes.

I’m Macy Mills and I lead GTM at a16z speedrun. My job is to give our speedrun founders an unfair advantage by helping them accelerate their traction. We do this with large-scale events, customer introductions, and hands-on advice.

Today, I want to share the seven mistakes I see most often at the early stage, and how to fix them fast so you can build real traction.

Watch the full video below, or scroll down for all seven mistakes:

1. Don’t Wait for Perfection Before You Start Selling

You don’t need a perfect product to start selling. You need a hypothesis about a real problem that you think you can solve. If you’re waiting until your deck is beautiful and your product is flawless, you’re honestly waiting too long.

Selling is a contact sport. You learn by doing. Get out there, start talking to customers, get feedback, and adjust. The input you get from real conversations is way more valuable than making sure you show potential customers the perfect pitch. Feedback beats stealth mode every single time, and there’s a reason why everyone is building in public.

Early customer signals are what is going to help you get to product-market fit, specifically by allowing you to diagnose when something isn’t working as soon as possible. It’s critical you understand your customers’ pain points deeply, and by waiting until you feel fully “ready” you may find yourself too far down the wrong path.

2. Stop Underpricing Yourself

This is a huge one. Founders are afraid to charge what their product is worth. I can’t tell you how many times I’ve seen founders charging one or two thousand dollars a month. That’s pocket change for big organizations. I personally don’t even think it’s worth asking for money if you’re only asking for that much. At that point, you should just move forward with a “design partnership” or a customer you don’t charge anything in return for product feedback, marketing case studies, and references. Honestly, at some point those intangibles have more value for you.

Founders think, “If I make it cheap, it’ll be easier to close,” or “I don’t have enough value here to ask for money yet.” Wrong. You should always start by asking for an amount that scares you just a little bit. You may be surprised at how many times you hear yes. And if they say no way, then fine: ask them what they feel is fair. Have the confidence to throw a number out there and see what happens.

3. Find the Person Who Holds the Budget

You’ve got an enthusiastic champion inside the company. Awesome. But unfortunately, enthusiasm doesn’t sign contracts. You need to find the person with budget authority early.

Ask things like, “Who else needs to weigh in before we move forward?” or “Whose budget would this come from?” The fastest way to stall a deal is to think you’ve got an opportunity when it turns out to be a fan instead of a financer. You need to be positive you’ve found the budget authority, also known as the Economic Buyer.

4. Listen More Than You Talk

The best founders learn to listen. They ask the right questions, like “What’s the process for getting this deal signed?” and “How high of priority is solving this problem for you right now?”

I like to give the guidance of an 80/20 rule: 80% listening, 20% talking. Ask open questions and write down the customer’s exact words. Once you have a good understanding of their priorities and problems they are hoping to solve, you can then deliver a pitch that resonates. They will immediately trust you more, because they know you were listening to them, and seem to understand the problem.

5. Don’t Fall in Love with Just One Customer

I see this every week. A founder gets one hot lead, one big logo, and suddenly all roads lead to that one customer. But if you’re hanging your fundraising outcome on one deal, you’re setting yourself up to fail.

Founders should be running many design-partner conversations in parallel. The reality is that a pipeline is just that, and many will say no. As a founder,you have to get used to hearing no and moving on. You should actually see it as a blessing so you can stop wasting your time. If all your potential design partners want similar things, that’s great news. But don’t get your hopes up with that one big logo. Make sure you’re spreading the love across a larger pipeline. And celebrate the no’s! :)

6. Your Pitch Should Evolve with Every Conversation

One of the biggest mistakes I see is founders treating their pitch like it’s permanent. Your pitch is supposed to get better with every single conversation. The first ten times you explain what you do, you’re really learning how people hear it. Watch for where they light up, where they look confused, where they interrupt you with questions. Those are great signals to take as learnings as you continue to hone your pitch.

Your story should evolve as fast as your product does. After every call, I recommend taking sixty seconds to jot down what resonated and what didn’t. That’s how you build a pitch over time that actually converts. And when you’re in a live conversation, stay flexible. If a buyer reacts to a certain product feature, lean into it. The best founders know how to read the room and adjust on their feet.

7. Don’t Hire a Salesperson Too Early

Founder-led sales is so important at the early stages of building your business. For technical founders, it can be tempting to try and hand off the sales job so you can focus on building the product.But the truth is, if you can’t sell it, a salesperson definitely can’t.

Founders have to run those early laps themselves. You’re discovering the sales motion, along with selling a product. You’re learning what customers are saying firsthand and allowing that to influence your product roadmap. No one can influence the product roadmap faster than you! You might as well get used to it because being a CEO involves a lot of selling. Customers are actually going to take you more seriously if you’re the founder, because you’re there selling the vision that you are obviously passionate about it. Once you’ve closed many repeatable deals and know the playbook, then you can hire that salesperson to help you run it. But not before.

Those are the seven biggest mistakes I see founders make when chasing their very first deals. If you avoid even half of what we covered here, you’ll move faster, close smarter, and build real traction: the kind that investors like to see on paper.

If you’re in the middle of early-stage selling, let us know how it’s going.

That’s it for this week! Thanks to Macy for sharing her hard-earned sales wisdom.

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

❓ Emily Bennett shares one of her favorite cofounder diligence questions

💰 Fearn (SR005) announced they’ve raised a $5.5M seed to help inventors file high-quality patents in mere minutes.

☀️ Orbital (SR007) raised a $5M pre-seed to build solar-powered intelligence

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Every desperate founder, betting their last dollar on a pivot, leans on the lore: Slack was a game studio. Instagram was a check-in app. Twitter was a podcast directory. These iconic stories of the web 2.0 era give all founders hope that unimaginable riches are just one more iteration away.

But are these stories even relevant anymore? AI model capabilities shift every quarter, which means that a new pivot is theoretically possible every quarter. Should every company be pivoting accordingly? Knowing when to pivot, how fast, and toward what has quietly become a core founder skill in the AI era.

So we asked speedrun investors Emily Bennett and Troy Kirwin what advice they give to founders considering a pivot.

First, what a pivot actually is

In basketball, a pivot means one foot stays planted. Move both feet and it’s traveling. (Go Knicks.) Startups work the same way. A good pivot keeps something planted, oftentimes the team and the investor’s cash, while everything else moves. The term originated from author Eric Ries and his 2011 book A Lean Startup, though he initially blogged about it in 2009.

Ask a speedrun investor and they’ll tell you that sometimes the worst failure comes due to refusing to pivot at all. As Emily Bennett puts it:

“Pivots are really hard to do. But what I always say is that the worst thing to do is to run with obstinate determination at the wrong thing, once it’s already become very clear that it’s the wrong thing.”

There’s even a case that the founders best built to pivot are the ones with the least to unlearn. Troy Kirwin argues that inexperience can be a feature:

“One of the ways young, first-time technical founders are empowered versus more senior founders is that they don’t come with any experience. So they’re not backed into a corner about what industries they feel they need to build for. They can be more of a heat-seeking missile to go test and validate any market. That’s the double-edged sword of founder-market fit.”

Three recent pivots worth studying

Clay: pivot toward the heat. Kareem Amin co-founded Clay in 2017 with a beautiful, somewhat abstract mission to make the power of programming accessible to everyone. The team’s product was a spreadsheet that could talk to the whole internet. Five years later, revenue was still close to zero. (Ouch.) The handful of customers they did have were using it for everything from accounting to sales. The unlock was noticing that some of their most active users were growth teams chaining data sources together to find customers. So, Clay rebuilt its entire identity around go-to-market. Two years after that pivot they hit $30M ARR. By early 2026 they were at $100M+ ARR.

Lovable: pivot toward insight. In June 2023, Anton Osika spent a weekend building a command-line tool that generated codebases from plain English. He then open-sourced the project and it pulled 50,000 GitHub stars in about two months. The beacons were lit; customers were calling for product. He and co-founder Fabian Hedin rebuilt the idea as Lovable, a closed, polished tool for non-technical builders. It hit $100M ARR in roughly eight months. Plenty of teams have turned repos into companies, Databricks sold Spark back to the same engineers already running it, and LangChain, Ollama, and ComfyUI are monetizing their own user bases right now. Lovable ran a stranger play. While the repo proved there was demand, the founders realized that the real market would be the 99% who couldn’t use a command-line tool at all.

Cursor: pivot toward yourself. Before Cursor, Anysphere’s four founders spent nearly a year building AI autocomplete for CAD software. CEO Michael Truell calls that period “wandering in the desert,” and his diagnosis of what went wrong is sorta obvious. They weren’t mechanical engineers, they weren’t passionate about the space, and there wasn’t much CAD data to train on. That idea might be a good one, but they weren’t the founders to build it. So, they swung to a market where they were the customer, where they actually cared, and where the internet was full of training data—aka coding.

The playbook

Across these stories and our investors’ answers, the same principles keep showing up.

When should you pivot?

Emily warns is that early stage founders sometimes fall in love with their own idea that they fail to notice the lack of market pull for it:

“Very early in the founder journey, you’re instinctually looking for market pull. You’re looking for signal from users that there’s real, unrestrained demand for what you’re pitching them… a commercial appetite underlying it, a willingness to pay. I’ve seen some founders get so excited by their concept, by the market they think they’re building in, that they’re blind to the lack of signal on that front.”

And when the signal isn’t there, she pushes founders past the comfortable explanation:

“You can either say, well, they’re not getting it. Or, the more likely answer, is that there’s something you’re not getting.”

Troy turns this into a cheap, time-boxed test that forces the truth out fast:

“For founders who are thinking about pivoting, I say: Why don’t you go try to fill your calendar with another fifty conversations with customers over the next two weeks? If you can’t even get those conversations, then that itself is a signal… If it feels like pulling teeth to get them to hop on another call, or they ghost you, then that’s a signal. You’re never going to have one hundred percent conversion, but if you can’t get fifteen or twenty percent, then that’s a problem.”

Troy says that sometimes fear can keep founders from acting on the signals they’ve gotten from the market:

“Folks are sometimes afraid of the truth, because the truth can lead toward sunk-cost fallacy and realizing you just burned the last three to six months on something you now need to shutter… But the most successful founders are truth-seekers.”

How should you pivot?

Remember the basketball rule: a good pivot keeps one foot planted. What stays planted, Emily argues, is everything you already learned, including the pile of conversations with customers a founder is tempted to write off as wasted:

“Look at all the conversations you just had that you now may think are totally moot. You’re not wasting time. Nothing is wasted. Look at the prior concept you were trying to launch, and the main positive insights you can parse from that experience then become the foundations to enable you to move faster at a new idea. Usually there’s a kernel of something worth keeping.”

From there, Emily runs a three-step playbook. It starts deliberately small:

“Find ten people who are willing to go incredibly deep with you and share all of their thoughts. Make that a representative pool… really use their input to inform your new market direction.”

Only once that depth of insight is in hand does she widen the funnel, running side-by-side A/B tests on messaging and archetypes to layer quantitative signal on top of the qualitative.

As for the conversations themselves, Troy says the goal is to learn:

“This is Sales 101 stuff. You’re not necessarily pitching on the call, but doing discovery on the call. State your hypotheses of a problem you’ve identified with other customers and ask if that resonates.”

And in 2026, there’s no excuse to delay the test. Emily’s point lands especially hard given what Lovable became:

“Now it’s so easy to build, you can literally pull up Lovable and create a rapid, interactive prototype and put that into the world. So your excuse should never be that you need to do product debt in order to get customer feedback.”

Anything else founders should think about before pivoting?

The hardest pivots tend to be for companies that aren’t clearly dead, but aren’t growing fast enough. Here Emily makes founders confront whether they’re building a venture-shaped company:

“There are good businesses that don’t grow unbelievably rapidly. They’re good lifestyle businesses, they’re actually profitable. But it just means you probably aren’t going to be able to command scaled venture dollars… It all goes back to the venture business model.”

For an idea that truly isn’t working, though, she thinks founders have the fear backwards:

“What’s scarier is building a product for nobody and wasting a lot of money.”

And for anyone dreading the conversation with their investors, Troy offers the reassurance that a pivot is a green flag, not a red one:

“Sometimes founders are nervous to come to us with the potential of a pivot. And to us, sometimes I’m excited to hear that, because it just means they’re moving fast. Getting to conviction that a space isn’t working is as strong a sign as showing that it is working. To me, it’s a sign of a great team.”

That’s it for this week! Thanks to Troy and Emily from our investing team for sharing their insights.

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

New York Tech Week just wrapped, and it was a big one, with 50,000 attendees across over 1,500 events.

The a16z speedrun team was on the ground hosting events and meeting with founders and investors, so we polled the team for hot takes and observations about their time in NYC.

One theme that came up was the cultural difference between SF and NYC, especially when it comes to AI adoption. a16z speedrun investor Troy Kirwin observed:

“NY is becoming AI pilled, but there is definitely a 4-6 month lag to counterparts in SF. I spoke with a number of non-tech knowledge workers whose companies are just now adopting Claude Code/Cowork while most people I talk to in SF are now jumping to Codex.”

Bella Nazzari from the a16z speedrun talent team said that people love the NYTW events hosted by New Yorkers:

“New Yorkers think Tech Week is just San Francisco coming to New York yet New Yorkers are coming to events, hosting them, and trying to find the offshoot culturally. They want tech hosted in New York by the people who know New York in all of its angles.

“There’s a want for something more anonymized. The gossip hotline, posters around the city, the dinner you get invited to with no partiful. People in New York are more likely to attend with the strong referral of one person versus the scan of a whole list on a platform.”

Tom Hammer’s take on the brunch this year:

“The topic of culture has an unfortunate tendency to entropy into Linkedin-worthy platitudes, especially in front of a public audience. Making this conversation explicitly private was intended to let the founders get a little bit spicier... and it worked 😀.”

Many of the best events were focused on serving founders. HubSpot for Startups featured a blind pitch event where a16z speedrun investor Kenan Saleh listened to and gave feedback on pitches from founders hidden behind a partition:

Student events and hackathons were a big draw too. Cindy Morand co-hosted a “vibe code & tea” event that was sponsored by Google Developers Group of NYC, FinOptiv, and Columbia Entrepreneurship Organization.

“We had 1.1k requests, accepted 224, and 143 vibe coded an actual app,” Morand says.

Lejla Johnsen said a lot of the magic of Tech Week was the chance for startup founders to meet buyers and decision-makers from top cos:

“Incredible energy, and met so many senior and C-suite execs from Fortune 500 companies. So magical to see win-win partnerships take place due to Tech Week!”

Events with creators and media tended to draw some of the largest lines and best discussions. Substack tended to come up a lot in these discussions.

“I heard multiple times that everyone is now leaning into Substack, that Instagram has gone LinkedIn and TikTok is too infantile, so Substack is the middle road for creators.”
—Bella Nazzari

The Tech Week team says they’re already gearing up for TW SF and LA this October.

You can submit your own event for listing on the Tech Week site. Hope to see you there!

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

🖌️ Sekai (SR003) raised a $23M series A for its AI app creation tool

💻 ZeroDrift (SR006) just announced they’ve raised a $10M seed to enforce compliance on AI communications.

🗽 David Haber says the best founders don’t end up in New York by accident.

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Building in public used to be simple. You’d launch on Product Hunt with some story, ask (aka beg) your friends to share it, tweet about it all day, and then hopefully some customers would appear.

The second wave was all about bombast, with indie hackers posting revenue dashboards and selling to the audience watching. But as a general backlash against hustle culture sets in, there are new and better ways of building in public emerging. The new growth meta is to be publicly excellent, making the company itself your content.

Here are three founders who are running this playbook right now.

1) Ship the frontier

As founders, you can live on the edge of the possible with AI models. Whenever a new model drops, it should, hopefully, enable new use cases for your product.

Tom Krcha exemplifies this at Pencil (SR005), the design tool built to live where you code. When Pencil crossed 100,000 users in May, the announcement was about Swarm mode: six AI design agents working a single canvas in parallel. It’s a visually enticing orchestra: a new way of visualizing AI tech at the very cutting edge.

It helps that Krcha is a four-time builder who sold companies to Miro and Google, so these updates read as a frontier operator reporting from the edge rather than a marketer hyping a release.

His video about swarm mode is a perfect example, six agents on one canvas could not have shipped a year ago, which is the line between building in public and posting a changelog.

2) Bet the company in public

If you’ve read any tech-adjacent content you’ll have seen debates on how SaaS companies should adjust their business strategies in the AI era.

One way to build in public is to take a hard stance on this question: pick a business model on the edge of possible, something newly enabled by AI, and tell the story about your ongoing transformation.

Chuck Ganapathi is doing exactly this at Gainsight. In May he told Upstarts that the company, on hundreds of millions in ARR, was going all-in on AI-native services over SaaS, spinning up an agent product called Atlas as its own P&L. A profitable company publicly declared its own core model obsolete! Talk about pivoting in public.

“There would be a revolt if we said we were going to kill that business,” Ganapathi told Upstarts. “But our fundamental belief is that Atlas, and the AI native services business, has a larger market.”

Now there’s a real storyline here worth following: Will Gainsight pull it off? Every update they post from here has added narrative weight because they’ve been so public about the stakes.

3) Hire weird, on purpose

The third version turns who you hire, and on what terms, into the headline. The most legible signal a company sends is the caliber of person willing to join it and the seat they take. When that seat breaks convention, the announcement becomes a build in public moment for staffing.

For example, Jean-Michel Lemieux just ran this at Spellbook, the “Cursor for contracts” startup that raised a $50M Series B from Khosla last October. Lemieux was CTO of Shopify, where he scaled engineering from 50 people to roughly 4,000, and before that ran engineering at Atlassian, helping both companies go from single product to platform. He is a heavy hitter who you would expect to join Spellbook as a CTO. Instead, he joined as something he calls an Executive IC. “It means I’m here to build and be hands on in every part of the company,” he wrote. A person who has held the top engineering seat at two iconic companies took one with no title leverage and no reports, in public, on purpose. This is weird! It’s also cool.It’s a handy way to grab some attention and get a kick ass employee at the same time.

Steak before sizzle

The rule underlying all three of these ideas is that you only earn the right to “build in public” if you’re doing something worth talking about. If you are, your content and your company become the same thing. You have to actually be shipping at the frontier, actually betting the business, actually landing the hire who only makes sense if you’re right. It isn’t about posting grindslop hustleporn about how much you work (spoiler, we are all working hard). It’s about how you’re working differently. The results will speak for themselves.

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Calling all operators, executives, and the AI-curious! Join us at New York Tech Week for our first ever east coast AI Faire. Meet 20+ speedrun founders and explore live demos of the AI tools they’re building.

One joy of working with founders through speedrun is seeing how quickly teams evolve. Some teams get accepted into the program and end up pivoting completely. They go down a rabbit hole and end up chasing an adjacent idea.

Others stick to their guns and iterate, becoming ever more refined in the process. Concorda (SR006) is one of those. Today, they’re kindly allowing us to show their transformation in public.

Sam Oh and Ke Ma first met in 2021, in an elevator in Chicago in a building where they both lived. Sam was a software engineer at a startup and Ke was attending law school at Northwestern. The two immediately hit it off, and since it was Ke’s birthday, he invited Sam to his party that same day. “We’ve been best friends ever since,” says Ke, now the COO and cofounder of Concorda.

Before founding a startup together, the two did a whirlwind tour of the legal world. Ke worked at McKinsey & Company and as an Assistant State’s Attorney at the Cook County State’s Attorney Office, before leading operations at Sage, a YC-backed startup in the healthcare sector. Sam did tours of duty at Fish & Richardson, Latham & Watkins, and Avantech Law.

The Founder Journey Begins

Sometime in 2025, Sam and Ke got together and began to develop an idea for a startup that would apply agentic AI tech for trial lawyers. The duo found their way to speedrun through a referral to a speedrun investor, Emily Bennett, late in the application cycle for SR006. “We prepared thoroughly for our interviews, pitched our backgrounds, and showed that we had a deep understanding of the market,” Ke says. Emily loved the pitch and extended the team an offer.

From the jump, Concorda had a fairly clear vision and story to tell about what they wanted to build. But that story evolved as they went through speedrun and ultimately pitched at SR006 Demo Day.

To show that evolution, below we’re sharing both the decks Concorda submitted when originally applying to speedrun and when presenting at Demo Day at the conclusion of the 12-week program. (Some numbers redacted for confidentiality.)

Concorda Original Pitch Deck

640KB ∙ PDF file

Download

Download

Concorda Demo Day Pitch Deck

1.19MB ∙ PDF file

Download

Download

Slide 1: Positioning for Ambition

The level-up in the branding is obvious, thanks to the Concorda team leveraging speedrun’s “Brand Lab” program to develop a new logo, color and font system, and style guide. Our own Lester Chen can tell you more about Brand Lab. Onto the content:

The initial pitch of “Agentic AI for Complex Litigation” is defensible and reasonable: exactly what you would expect from former lawyers. But the new version, “AI Operating System for trial lawyers” is more clear about who the end-user is, and the vision is simply bigger. The implication is that ultimately Concorda aims to build everything a trial lawyer needs.

“We learned to be both more precise with our internal plans, and more ambitious in our overall vision for the company,” says Ke.

Slide 2: Dress for Success

One thing that speedrun GP Andrew Chen likes to tell founders is “brag about yourself more.”

If the title slide introduces the vision, the team slide shows why the founders are the best people on the planet to pull this off. Ideally, you establish why you have unfair insights about your target market and why you have the technical chops to build a product that takes advantage of those secrets. “Engineers who try cases. Lawyers who ship code.” is a pithy way to accomplish exactly that.

Slide 3: Show ‘em the Money

Ideas are great, but money talks. When teams apply to speedrun, many don’t have revenue. But by the time you get to Demo Day, it helps to show some progress. Concorda’s real deck had unredacted numbers in straightforward, plain language.

Slide 4: From Mission Statement to Problem Statement

Originally, Concorda’s deck had a slide with a mission statement with callouts annotating which words mean what. But this was a little redundant with the title slide, so Concorda swapped it out for a more compelling problem statement. When presented on stage or on a call with investors, you can extrapolate on a slide like this with anecdotes and specific examples about why it’s a problem that, for example, “generic legal AI has no context.”

In Concorda’s case, they could defend this conclusion by pulling from Sam’s story of being the AI specialist at a litigation boutique and Ke’s background at Cook County State’s Attorney Office. The goal is to show the market the gap you understand better/different than everyone else. Problem statements describe a world the investor already knows is broken. And in this case, it’s a world the founder has already audited firsthand.

Slide 5: TAM and the Limits of the Possible

Let’s be honest: TAM is mostly just a vibe. Nobody actually knows what the TAM for your market is. And besides, generational startups invent markets. They change how work is done.

When making your TAM slide show the key constraints restricting market size, and then show how your product unlocks those constraints. You can be more realistic about it, it makes you more believable.

Slide 6: Send it Home

Concorda took Andrew Chen’s advice to brag more to heart, and ended their Demo Day deck with a humdinger of a brag. Great walk-off line.

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Next week, Tech Week lands in Boston, followed by NYC. If you’re a founder, check out Pitch Day, a live event at the a16z NY offices where selected founders will have the chance to pitch their company and get feedback from a16z investors.

This is our biggest East coast Tech Week yet, with 2,000+ events happening across both cities. Boston is from May 26 to 31 and NYC is right after from June 1 to 7.

Check out the full calendar and build your own personalized schedule here.

The official application window for SR007 has closed! Investors are now reviewing applications and doing followups to select teams for our Summer/Fall cohort.

If you applied and haven’t heard back, hang tight. We’re reviewing apps on a rolling basis and expect to get back to everyone within ~4 weeks from today.

In the meantime, we know many of you are wondering: what goes on behind the scenes while investors are looking at apps? To answer, we hopped in the booth with Kenan Saleh, Emily Bennett, Fareed Mosavat, and Troy Kirwin from the a16z speedrun investing team to talk about:

00:00 - patterns we’re seeing in apps for SR007
04:50 - our process for reviewing apps
08:20 - traction signals we look for
09:10 - on teams that are a little too early for speedrun
14:26 - surprising things we’ve seen in interviews
20:46 - why you SHOULD NOT take vc funds
25:16 - why should founders pick speedrun?

Watch the full roundtable here:

We’ve put out a lot of content over the last month to explain the speedrun program and what we hope to see from applicants.

Scroll down below to hear directly from 18 different partners from the firm as they explain:

• How the a16z speedrun program supports founders

• The types of startups investors throughout a16z are hoping to invest in

• The big ideas our investors are obsessed with in 2026

If you have questions, shoot us a note at sr-team@a16z.com.

How a16z speedrun Supports Founders

Pitching to VCs is stressful. So we’re turning the tables and pitching to you instead.

The a16z speedrun team has a stacked crew of operators to help you build, scale, and launch your business. Hear from five of our partners below as they explain how they give speedrun founders unfair advantages over the competition.

The Global Founders Program helps you build in America

By Tom Hammer (Head of People, a16z speedrun) - X | LinkedIn

30 of the 70 companies in the last Speedrun batch used our Global Founders Program for help navigating visas and immigration. And not all of those founders were based abroad. For founders on H-1B, TN, or a number of other visas currently working in the US, there’s a path to founding a company and having your own startup sponsor you.

The GTM Team kicks down doors to help you scale

By Macy Mills (Head of GTM, a16z speedrun) - X | LinkedIn

What if I told you I had over 50 executive buyers across Fortune 500 companies, who regularly engage with founders. This year we launched Speedrun Insiders. The exact people you’re trying to get in a room, who will actually take meetings, give feedback, and explore real deployments. We already have them lined up for you, ready to engage.

Brand Lab and Launch Lab make your marketing soar

By Lester Chen (Head of Creators, a16z speedrun) - X | LinkedIn

Every early-stage startup needs attention. Your weekend vibecoded website isn't cutting it. You watch videos but don't know how to make one. 'New media' seems distant, inaccessible. You can't tell your story, so others won't either. Here at speedrun, we’ve designed brand, story, and launch programs to help fix this.

The Talent Network helps you hire

By Bella Nazzari (Talent Partner, a16z speedrun) - X | LinkedIn

We’re the team that helps you hire from the moment you get accepted to speedrun, until well after you finish our program. We’re on track to make over 100 placements across our portfolio this year. Founding engineers, designers, marketers, sales, and all of the new job titles coined this year. No referral fees, no agencies. Direct free hires.

The speedrun marketplace unlocks $7M+ in credits

By Emlyn Thompson (Ecosystem Lead, a16z speedrun) - X | LinkedIn

This is the a16z speedrun Marketplace. Your operating stack, curated and pre-negotiated.We’re talking millions in savings across the tools you actually need: banking, AI, infrastructure, cloud. Everything you need, already onboard.

a16z Team Requests for Startups

The speedrun team partners deeply with the investors on other funds at a16z to find, fund, and support startups. Parters from our Bio + Health, Infrastructure & AI, and American Dynamism teams hopped in the studio to talk about the sort of startups they hope to see.

Early-stage skin cancer detection

By Eva Steinman (Investing Partner, a16z Bio + Health) - X | LinkedIn

Getting to a dermatologist takes three to six months, and in rural areas it often simply doesn't happen at all. We understand how to treat early-stage melanoma. The problem is that we are systematically failing to detect it in time.

A space warfighting prime

By Christian Keil (Investing Partner, American Dynamism) - X | LinkedIn

Combine these ingredients: ubiquitous launch, hundreds of billions of assets in space, and increasingly hostile activity from America’s adversaries… and the opportunity is self-evident. We need the ability to defend our valuable assets on orbit, which is why I want to fund a space warfighting prime.

Video intelligence companies

By Jason Cui (Investing Partner, Infrastructure & AI) - X | LinkedIn

Over the past few years, we've seen remarkable advances in video generation, editing, avatars, and creative tooling. An increasingly important problem is teaching machines to search, analyze, reason over, and extract insight from video across massive libraries and live streams alike.

Infinite healthcare apps

By Jay Rughani (Investing Partner, a16z Bio + Health) - X | LinkedIn

Today, we don't have enough healthcare professionals in America. Wait times to see a doctor average over a month. We're short 600,000+ doctors and nurses. And nearly half of the clinicians we do have are burned out, buried in paperwork. This is a math problem. And AI can fix the equation.

Big Ideas from the Investors

The investors on the speedrun team are always keeping an eye out for the next big bottleneck holding back the future. Below are eight big ideas about what the world needs next from startup founders.

Proactive AI agents

By Kenan Saleh (Investing Partner, a16z speedrun) - X | LinkedIn

Today’s AI products are reactive. You give the model a prompt, it responds with an answer. These are useful, but I’m excited about products that take this further and shift the paradigm from “ask to answer” to “observe to act.”

GUIs for agents

By Jonathan Lai (General Partner, a16z speedrun) - X | LinkedIn

We’re still in the MS-DOS era of agents toda: CLI, terminal sessions, file directories deleted by openclaw etc. while a small slice of silicon valley are power users. At speedrun, we’re looking for bold founders excited to bring the power of agents to normies everywhere.

We’re entering a world of infinite output

By Lejla Johnsen (Investing Partner, a16z speedrun) - X | LinkedIn

What would it look like to create tools that train people to attach real probabilities to their beliefs… and see, over time, where they tend to overestimate or underestimate?Or systems that can make decision-making visible, trackable, and improvable?

Come for the agent, stay for the network

By Troy Kirwin (Investing Partner, a16z speedrun) - X | LinkedIn

In the past, suppliers thrived off of the offline nature of these markets with an agentic platform, the demand side can be aggregated and the power balance flipped. You can start to become the interface buyers default to, the channel suppliers need to be on, and the owner of the richest pricing dataset in the industry.

Tools for the 100x solo founder

By Emily Bennett (Investing Partner, a16z speedrun) - X | LinkedIn

For decades, company size has been dictated by coordination costs. If you wanted to grow revenue, you hired people. More customers meant more support reps. More revenue meant more finance and ops. Scale meant headcount. AI is starting to break that relationship. At speedrun, we’re interested in startups building tools that allow a single founder to operate what would previously have required entire teams.

The next great software distribution company

By Josh Lu (General Manager, a16z speedrun) - X | LinkedIn

For the last decade-plus, building a strong software business was mostly a game of scarcity. The hard part was assembling the capital, taste, engineering, and distribution to build one enduring app. The discovery systems we built matched that world: app stores, SEO, search bars, rankings, reviews, ads, marketplaces. AI has blown up these constraints.

Agent-native products are coming

By Fareed Mosavat (Visiting Partner, a16z speedrun) - X | LinkedIn

Every product on the internet was built for a human with eyes, a cursor, and a credit card. Agents have none of those things. Most companies are teaching agents to pretend to be humans. That's a hack. The real opportunity is products designed for agents from scratch.

New UI/UX paradigms

By Marcus Segal (Investing Partner, a16z speedrun) - X | LinkedIn

The hardest part of software used to be building it, but now it's designing it. Hamburger menus, nav bars, buttons, chat fields... None of these were designed for a world where software participates.

And, lastly, if you’re wondering what actually happens after you hit “submit” on your application:

What happens after you apply to speedrun:

By Ryan Rigney (Marketing Partner, a16z speedrun) - X | LinkedIn

We’re looking for amazing founders. What makes you special, what you understand about this market that other people don't, and evidence that you know how to ship: these are the things that make applications stand out. Traction with customers and an already-assembled killer cofounder team go a long way too.

That’s it for this week! For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Just fyi that startups applying to speedrun should do it before Sunday, May 17th. If y’all have been waiting for whatever milestone (the demo to be better or metrics or whatever) just apply now. The best founders we fund usually applied too early by their own standards!!

If you’re on the fence, this is the call. Send it in.

More from the team below, plus some of the questions I get asked the most about speedrun.

–A

Andrew Chen
General Partner, a16z speedrun
2865 Sand Hill Rd, Suite 101. Menlo Park CA 94025

__

The questions I get the most

“How do I know if I’m actually cut out to be a founder?”

Honestly, I think most people secretly know if they’re founders or not. Some of you can never be happy working inside a giant company, writing docs, in endless meetings. Deep down, you already know you’re supposed to build. If that describes you, the question isn’t whether you should start a company. It’s just when. There are very few moments in tech I’d rather be a founder in than right now.

“Am I too early to apply? My company isn’t even incorporated.”

That’s actually the sweet spot. In our last batch of 70 companies, probably a dozen were people who hadn’t quit their jobs yet when they applied. They had to turn in their badges, and we couldn’t even wire the money over because they didn’t have a company to incorporate yet. We want to find founders as close to day one as possible. If you’re a couple of friends with a real idea and the willingness to drop everything to go work on it, that’s a speedrun application.

“Do I need a cofounder, or can I apply solo?”

We don’t require it, and we do accept solo founders. But convincing one smart person to join you is probably step one in validating that it’s a serious effort, idea, or team. So we do look at it as a signal. If you have someone in mind but haven’t asked them yet, this is a good week to have that conversation.

“Do I need a specific background to get in?”

There isn’t a checklist. We have a team in the current batch that’s all 18 and 19 year olds who decided to skip college, and we also have cofounders of billion-dollar companies in the same batch. What we look for is something unique in your background. That could be a GitHub repo with thousands of stars growing fast, deep technical work at one of the top AI labs, a serious athletic background on the way to tech, or any number of other things we wouldn’t have predicted. Pedigree is one signal, but it’s not the only one.

“Is speedrun only for AI companies?”

No, it’s a horizontal program. The current batch has B2B, B2C, infra and dev tools, healthcare, gaming, govtech, and a growing number of robotics teams alongside the AI applications. There are also famous examples of ideas changing all the time. Slack started as a browser-based video game company. So if you’re an AI founder today and you end up pivoting in week three, that’s a pretty old Silicon Valley tradition. We care about the founders more than the category.

“I’m outside the US. Can I apply?”

Yes. We select founders from all over the world. This batch has founders applying from the US, India, UK, Nigeria, Canada, and dozens of other countries. If you get in, you and your cofounder will need to come to San Francisco for the 12 weeks of the program, and we help with visa support to make that possible. Don’t let geography stop you from applying.

“Do I have to move to SF?”

For the 12 weeks of the program, yes. You and your cofounder come to San Francisco. We have a welcome guide with recommended neighborhoods, help on Airbnbs, and access to the a16z office and a couple of co-working spaces. Most teams end up rooming with their cofounder for the duration, working together 24/7. After the program ends, plenty of teams go back to their home communities (Austin, Chicago, NYC, Pacific Northwest, wherever they came from, including other countries) and keep building. Plenty stay in the Bay. Both are great outcomes.

“What’s the biggest mistake founders make in the application?”

Being long-winded and vague when you talk about yourselves. A lot of founders bury the lede on their own team. If you started a successful company before, tell us you went from zero to X million in revenue. If you worked on a well-known product, tell us how many millions of users you served. If your side project hit a million downloads, say so. Specific facts and figures help. You can call this bragging. I call it the only good way to pitch a team in writing. (I wrote a longer piece about this called The Dinner Party Jerk Test if you want the full version.)

“What happens if my company doesn’t work out?”

Realistically, a lot of startups don’t. That’s just venture. What we care about is what happens next, and we think of the relationship as a lifelong one, not a batch-long one. The best outcome is we back the founder again on a new idea. We just did that last week with a founder whose first idea didn’t pan out. The other thing we see a lot is speedrun founders hiring each other when one team doesn’t make it, because the people coming through this program are unusually good at working on hard, unstructured problems with the latest AI tools. We want to be in business with these founders for their whole career.

The deadline to apply for SR007 is Sunday, May 17th at 11:59pm PT. Apply here.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

⏰ Applications for SR007 close this Sunday, May 17th, at 11:59pm PT! Apply now

🇺🇸 a16z speedrun GP Andrew Chen had a great conversation with Matt Perault about Little Tech on the a16z AI Policy Brief Podcast

🤖 Lejla Johnsen argues that in a world of infinite output, judgment about what to build is what matters

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Today’s newsletter is a guest post by a16z speedrun investing partner Emily Bennett. She argues that spotting early markets is a skill: one that founders can learn.

Everybody loves a good market map. It’s seductive to think that you can compress the world into tidy grids with logos neatly sorted into categories, each square representing a company that has raised meaningful capital. For investors, these maps are useful tools to help think through investment opportunities.

They are also, unfortunately, dangerous for founders.

By the time a market is legible enough to be mapped, the most valuable entry points have already closed. The best founders are not waiting for a sector to mature enough that a VC can toss it on a 2x2 it. Generational companies are built by recognizing patterns of behavior that precede the sector map entirely. In my work at a16z speedrun, the most promising opportunities almost never arrive looking like opportunities. They arrive looking strange.

Behavior Comes Before Categories

Markets begin as behavioral shifts in small, overlooked communities. It may be a few hundred developers who can’t stop talking about a new tool. Or maybe there are a few thousand users engaging with an app in ways that seem totally indecipherable to outsiders. The usage numbers are modest, and if you were just glancing at the product, you would probably ignore it.

What matters is the intensity of the behavior. There is a meaningful difference between an app that people open often and an app that changes how people operate. A worldchanging company has to, by its very definition, induce structural new ways of working/being/creating/existing.

Consider OpenClaw. An open-source project built for developers, it required terminal knowledge and real technical fluency. It was not used by everyone. But for those who adopted it, OpenClaw became a fundamental part of their workflow. It reinvented the way they conducted work.

What also made OpenClaw significant was its downstream effects. At speedrun we have already started seeing a new generation of pitches built on the conceptual foundation it propagated. Founders now have proof that agents could be held towards outcomes, not merely tasks completed. Within a week or two of OpenClaws launch, we had pitches for making these systems accessible to non-engineers. Then as the ideas permeated, we started having companies proposing agents that could autonomously run entire business functions, from legal to HR. The agentic workforce thesis now discussed at every conference accelerated across a small developer community restructuring their work around a tool most of the industry had never heard of.

But as a founder, you don’t have to wait for something as popular as OpenClaw to come along.

How to Spot an Early Market

Working with really early-stage companies at speedrun, you start to get a gut sense for whether something is being pulled in by real demand or just pushed out, well before the usual metrics tell you much.

What we try to get founders to look for is simple: are there signs that demand exists even before the product is fully there? The specifics vary, but a few patterns show up a lot.

First is what happens before launch. It is not just about racking up signups. Plenty of subpar products with good marketing campaigns can do that. But is there real intent behind the interest? You can see this in everything that forms around the waitlist. People share it, talk about it, and bring others in without being asked. That energy is coming from the market.

Second is early community. Sometimes people start gathering around a product before it’s even live, and the conversation keeps going on its own. No one from the company is propping it up. People show up because it connects with something they’ve been looking for. Users start creating explainers, tutorials, or threads about the product before there’s any official documentation. They’re doing the company’s marketing without being asked, because they want others to find it.

Third is how intensely a small group uses it once they get access. You’ll see a handful of users spending a surprising amount of time with the product, not because they have to, but because they want to. From there, one of two things usually happens. Either it replaces tools they were already using, or they start bending it into new use cases the founders did not plan for.

Enough of these happen and you know you have something primordial and powerful on your hands.

Building for What You Cannot Yet See

One common failure mode I see from founders is when a team gets stuck building “one more feature” before showing their app to users.

The founders I have seen get this right do something that sounds limiting but is actually the opposite. They pick ten users, maybe fewer, and they go deep. They become almost unreasonably attentive to how those specific people work and what they actually need. And the insights that come out of that kind of closeness are usually generalizable. What ten power users care about tends to map, at least directionally, onto what ten thousand eventual users will care about. Narrowing your focus and shipping live features constantly It is how you earn the right to go broad.

The Map and the Territory

Market maps will keep getting published. They’re useful abstractions, but they’re definitionally lagging. By the time something is legible enough to map, the underlying behavior has already stabilized and the earliest forms of leverage have already been captured.

The next market worth building in is forming right now in a community, or a product, or a pre-order page that no one has yet thought to categorize. Whether you see it depends entirely on whether you are paying attention to behavior or waiting for a label.

For founders, that means building the habit before you need it. You have to haunt the weird corners before they’re legible, the Discord servers with no business model, the subreddits full of complaints about a workflow nobody has named yet. You’re looking for people who have already changed how they hope to work and just haven’t been handed the right product yet.

Thanks to Emily for sharing her insights!

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Applications for SR007 close in just two weeks! It only takes five minutes. Apply now.

This Week’s Big Ideas 💡

🗽 The Boston and New York Tech Week calendars are live! It’s the biggest East coast Tech Week yet, with 1,500+ events happening across both cities. Boston is from May 26–31 and NYC is right after from June 1–7. Check out the calendar and build your own personalized schedule here.

🇺🇸 Tom Hammer explains how speedrun supports founders from around the world.

🤖 Fareed Mosavat says agent-native applications are coming.

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Should you hire two engineers or give one engineer unlimited tokens and nicotine?

The answer to this question is an existential one for startups to answer. Most seed stage startups only have 12-18 months to live. So deciding whether to spend $300,000 a year on agents or using that cash to hire a second engineer is a do or die decision.

So what is a founder to do? Should you be token-maxxing?

The case for going long on tokens

The structural arguments for token-maxxing are that tokens scale linearly with throughput. You can 10x your fleet on a Tuesday afternoon. You cannot 10x your headcount in a day. Additionally, tokens are pure variable cost, with no equity dilution, no demands for green juice and complimentary massages, and no managing someone through a PIP. If the models get better in six months (as they have been) your engineering pipeline gets better.

It is also significantly faster. Founders are building more product in 12 weeks than they used to build in a year. Non-technical founders are shipping enterprise-ready platforms in just a couple of months. The upper bound of what a small team can do has moved, and it has moved fast.

Andrey Starenky at Sentra (SR006) puts a number on it. His team uses Claude Code, Cursor, and OpenCode in whatever combination each engineer prefers, with cloud agents wired into Slack to triage bugs and ship small features the moment a complaint comes in. He pegs the productivity boost at a minimum of 5-10x. The annualized Cursor bill alone, he told us, “is more than the salary of an average engineer.”

The aggressive use cases are even more lopsided. Freddie Wollen from Coalition Systems (SR006) runs agents overnight, attacking his own architecture in search of security flaws. “One workload where spraying tokens at the problem seems like the right answer,” he calls it. The Rust compiler and a strong test suite give the agents a closed feedback loop, so they iterate without supervision while the team sleeps. No human team would do this work, at this scale, for any price.

The point is that the upper bound of what a small team can do has moved, and it has moved fast.

The case for going long on people

Now the other side, just as hard.

When token-maxxing, it can be exceptionally hard to understand what caused the AI to go haywire, and to fix it going forward. With people you can just say “stop doing that or you are fired.”

Simo Rachidi of Safeworld (SR006) builds robotics, where weirdly stupid bugs have physical consequences. He has postponed hiring around internal tools, QA automation, and lightweight full-stack work, but draws a hard line on the roles that carry accountability with tasks like robotics safety, systems architecture, simulation fidelity, and customer trust. “AI gives every engineer more hands,” he says. “It does not replace ownership.”

Token costs are also a moving target, and they have moved in both directions. Frontier model pricing has been reset multiple times in the last two years. A $300K-a-year agent budget in Q1 can become a $500K budget by Q3 if the wrong tier gets repriced. Hiring a person commits you to a salary you can model. Hiring a fleet commits you to a vendor’s roadmap.

The complication

It may be that this is a false binary.

It’s possible founders are not actually foregoing hires because of tokens. Token spend is up. Headcount is also up. Founders are doing both at the same time, or doing neither for unrelated reasons like the funding environment or the difficulty of recruiting in 2026.

Jackie Lunger from Panorama (SR006) told us that her team’s first instinct when agentic tools arrived was to spin up hundreds of agents and let them cook. They coded fast and made the wrong choices. “Coding is actually the easy part,” she says. “The hard part has always been knowing why you are doing what you’re doing.” Now the team invests heavily in planning, and a single agent with clear instructions can one-shot the work. “Spending more tokens usually just means the clarity of thought wasn’t there when the task was started,” Jackie says. “More tokens signal less precision.”

Simo from Safeworld arrives at the same place from a different direction. “I would not call us token-maxxers,” he says. “I would call us failure-mode-maxxers. We use AI aggressively, but the thing we are really maximizing is confidence before deployment.”

If that hypothesis holds, then “token maxing versus hiring” is a false binary, and the real question is something more specific. Which categories of work should be automated, and which categories require a person who can be looked in the eye?

We can all stop asking “tokens or people” and start asking “what kind of work am I actually trying to get done.”

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Each Tuesday we’ll show up in your inbox with content created to be useful to startup founders and investors. Sometimes it’s a guide or strategy memo from an investor on our team. Other times we go straight to founders or successful creatives to get their advice.

Some recent popular posts:

• Emily Bennett on building financial models

• Troy Kirwin on how to write an investment memo

• Marcus Segal on building processes to find and keep product-market fit

Welcome! And drop us a line by replying to this email if you have any questions that you’d like us to address with future posts.

–The a16z speedrun team

Want to secure up to $1M for your startup? Applications for a16z speedrun SR007 are open now through May 17th. Apply now for earlier consideration.

This Week’s Big Ideas 💡

🗽 Tech Week is coming to Boston and New York! Check out the full calendar to sign up for events and join us as at the biggest decentralized conference in tech.

🎬 We just put out a fresh trailer for speedrun that asks: Why not you?

🤖 Troy Kirwin’s new thesis: Come for the agent, stay for the network.

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Today’s essay is a guest post by return contributor Evan Armstrong. Evan is the founder and CEO of The Leverage, a newsletter about how power actually moves in tech.

Evan argues that startups winning in attention marketplaces are investing more in their written product communication, and breaks down the strategies the best founders are using.

For one brief, deliriously delicious moment in the 2010s, cupcakes were the center of the dessert universe. Sprinkles installed 24-hour cupcake ATMs in six cities, each one playing a jingle while a mechanical arm delivered a $4.50 treat. Crumbs Bake Shop went public on NASDAQ, with the stock peaking at $13 a share. The world craved one thing above all else: baked circles topped with 6 tons of powdered sugar.

Then, suddenly, poof.

Crumbs’ stock fell from $13 to four cents. All 65 stores shuttered overnight. Sprinkles, the company that invented the cupcake vending machine, closed permanently on New Year’s Eve 2025. By making cupcakes ubiquitous, the industry had gorged the public on the product until everyone’s taste refined past what most shops could deliver. Only the best cupcake shops have survived.

This is now happening to writing in tech.

AI has made prose cheap. Every seed-stage founder with a pulse and ChatGPT account can produce a competent product update in 90 seconds. Just give an agent access to your codebase, prompt it with “write an email summarizing recent updates we’ve made and the value they’ll offer users” and away you go. This should be the employment apocalypse for the fools who love to wrestle with the written word (aka people like me).

And yet the companies winning in attention marketplaces are investing more in their written product communication, not less. Why? Because when the floor drops to zero, when the world is filled with shitty cupcakes, the ceiling is the only thing that matters.

What a bad feature update looks like

You already know this post. You have written this post:

“We’re excited to announce [Feature]. Here’s what it does. [Screenshot or a GIF if you are feeling fancy.] We can’t wait for you to try it!”

If you stripped away the logo and brand colors, it would be impossible to tell which company in your market wrote that update. If you can swap your name for a competitor’s and the post still reads fine, you have written the cupcake equivalent of a grocery store dessert. It’s fine. Serviceable even. It’s just that nobody will care about it. All of those expensive tokens, all that overpriced engineering time will be greeted with a collective shrug as your customers move on to more interesting things.

Here are three things that you can do to immediately improve your startup’s public facing writing.

1. The why matters over everything

Your customers do not care about your features. I need you to really sit with that and feel it in your bones. They. Do. Not. Care. Most people today have dozens of software products all of which are releasing updates on a near-weekly cadence. The drumbeat of “new feature!” notifications have become white noise, a hum they have long since tuned out. To get them to pay attention, you have to connect it to why you built it.

For an example of this, see how Linear launched Linear Agent this past March. They open the post with a thesis: as execution accelerates, the bottleneck in product development shifts toward judgment about what to build and where your time and tokens are best spent. The product (an AI agent) was the consequence of that strategic claim.

You can easily imagine the version ChatGPT would spit out: “We’re excited to announce Linear Agent: it is not just a new agent, it’s a new way to delve into work.”

In defense of that phrasing, it is accurate! It’s also boring. There is no strategy, no stakes. Customers are betting on startup’s products to wrench them into the future. If you just list out what you released there is no forward momentum to it.

The key test is that if your product update could have been written by anyone with access to your feature spec and a chatbot. The update should contain insights or secrets that only you know—whether that was a customer conversation that sparked the feature build, or your vision for the future.

2. Have an Opinion

When 37signals launched HEY, it would’ve been easy to announce “a new email client with filtering capabilities.” Jason Fried went with the more challenging approach of opening with an argument about the broken state of email: “You started getting stuff you didn’t want from people you didn’t know. You lost control over who could reach you. An avalanche of automated emails cluttered everything up. And Gmail, Outlook, Yahoo, and Apple just let it happen.”

Good products are built off of spiky opinions. Why should good product announcements be any different? The feature was called The Screener, which forced every first-time sender into a purgatory where the user was empowered to decide whether to ever hear from them again. DHH later said The Screener was “a pretty controversial feature to mandate internally” but it became “perhaps the highest profile feature since launch.”

You can disagree with HEY’s take on email! I certainly do. But you cannot confuse it for someone else’s product announcement. Most product updates hedge everything. You can feel the softening of the language, the committee of product managers, marketers, and sales staff that qualified every sentence. Great writing is expensive because it has a point of view the author is willing to defend. Cheap writing reads like it was reviewed by four people who each added a caveat.

You’ll note there isn’t anything about AI tells in this section. LLMs shine in helping you discover what your opinion is, even if the writing they actually suggest is usually mediocre. The only caveat is that they make it incredibly easy to convince yourself into any position. Use AI tools to make sure your argument and opinion are clear. Beware the temptation to use it as a source of validation.

3. Show What You Chose Not to Build

This is my personal favorite, and a perpetually underused tool in the writing tool kit. Write about what you didn’t launch! By walking your users through the very hard trade-offs every product represents, you create a story. Customers view your product as something built with real risks and choices. It draws them in.

When Superhuman built their AI features, CEO Rahul Vohra wrote about the decision process in detail. “There was no shortage of ideas,” he wrote. Should they automatically summarize long emails? Prewrite drafts for you to review? All great ideas that all ended up being rejected. Vohra’s reasoning was that those features are all preemptive. The user sets them up once and they run automatically. Preemptive features cost more to run and must meet a higher quality bar. “Imagine if automatic summaries or prewritten drafts were low quality. You would tune them out, or just turn them off,” Vohra wrote. So they started with on-demand features instead.

I distinctly remember getting this email. Despite getting a daily drum beat of feature releases from my personal software suite, I remember admiring the choices they made, even though I didn’t necessarily think they had made the right one.

There is a temptation to be Steve Jobs with feature releases. You descend from the mountaintop, enrobed in black turtleneck of a product geniushood and tell your customers what they want. Unfortunately, you are not that guy. It is much better to show your customers you are human, the hard choices you struggled with, and where you landed. And again, this type of email is something an AI draft will never produce on its own, because AI was not in the room when you decided to ship one thing instead of another.

How to Actually Use AI to Do This

Everything above can be summarized in one sentence: good product writing contains information that only you know, written in a way that only you could write. The practical question is how to get that information out of your head and into the post, using AI tools, without the result reading like it was written by a chatbot with access to your Jira board.

Mess over spec. If you give Claude your release notes and say “write blog post good,” you will get a steaming pile of meh every time. Instead, give it all the background data, pulling in slack threads, customer support tickets that sparked the idea, etc. Then, ramble into it for five minutes in voice mode about why this blog posts matters in the context of the company. AI can put it all together better than you can.

Use AI to find your argument. Ask: “What is the strongest case for why we built this instead of the three other things we could have built?” Use the answer to find your angle. Then rewrite until it sounds like something you’d say to a smart friend. If you would be comfortable sending it as a board update without rewriting first, you’ve messed up. (Or your board is probably a little too chill).

Write ugly first. The best AI-assisted product writing starts with a terrible human draft, full of half-thoughts and run-on sentences, and uses AI to clean it up. This is the opposite of what most people do. Starting ugly and polishing with AI produces writing that sounds like a person. Starting clean and trying to add personality produces writing that sounds like a chatbot in a tophat.

Your company can write better than it currently is. Use this post to get started.

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Last September, the founders of 60 startups were sitting right where you are now, looking at a message inviting them to apply to join the next batch of a16z speedrun.

Last Monday, those same founders stepped onto the stage at the Palace of Fine Arts in San Francisco to pitch their companies to over 1,000 in the audience and over 10,000 viewers online.

As the first attendees attendees arrived, a16z’s Andrew Lee climbed up on a chair to rally the crowd.

Breakfast was served as founders set up their booths and chatted with guests.

Quickly, the main theater filled up with guests eager to kick off the day’s main event.

Demo Day presentations are intense: a back-to-back onslaught of 90-second pitches from the 60 speedrun companies from the class of SR006.

But the founders came prepared. Founders like Sam Oh of Concorda (shown below) had already done a test-run in this venue two weeks prior, during “Alumni Demo Day.”

The range among this year’s companies was wild: insurance solutions, next-gen entertainment apps, laundry-folding robots, and AI agents for just about everything.

Once the pitches wrapped, every founder had time at their booths to field questions.

Here, as onstage, all the practice and time spent building pays off when you see a founder walking an investor through the product for the first time, watching the lightbulb go on.

Some founders came ready to be remembered. Paypath’s team showed up in custom “Modern Debt Rebuilt” merch, just the kind of branding instinct that sticks in an investor’s memory at 6pm after 60 pitches.

That’s a wrap on SR006. If you’re building something interesting, apply to join SR007. We can’t wait to see you on that stage.

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

There are still a few slots open at our invite-only café in San Francisco!

RSVP on Partiful using the links below to reserve your spot on one of the five days:

Monday (4/20) | Tuesday | Wednesday | Thursday | Friday (4/24)

This Week’s Big Ideas 💡

🏛️ Today is Demo Day for the class of SR006! We’ll be sharing lots of photos on our social accounts. Signups are closed for this one, but if you’d like to join us at the next Demo Day, register here.

💰 Shout out to Rork (SR005) on their $15M seed round!

🏋 Andrew Chen explained what the bar is to get into speedrun on a recent episode of George Bandarian’s Agentic Podcast

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Qasar Younis is the co-founder and CEO of Applied Intuition, a $15 billion AI software company for autonomous systems. Previously, Younis was an automotive engineer at General Motors, built two startups (the second acquired by Google, where he ended up reporting near the top of the org), and served as COO at Y Combinator.

In a fireside chat with a16z speedrun investing partner Kenan Saleh, Younis laid out how he thinks about decision-making, culture, recruiting, work ethic, and what it actually takes to sustain a company for the long haul. His philosophy, which he calls “radical pragmatism,” boils down to stripping emotion from decisions and refusing to borrow frameworks wholesale from anyone else, including him.

Scroll down for our top five takeaways from the conversation:

1. Don’t Borrow Other People’s Frameworks

Younis’s central philosophy is what he calls “radical pragmatism”—a term so core to his identity that radicalpragmatism.com literally redirects to his website. The idea came from years of watching smart people give confident advice that didn’t transfer. At YC, he was surrounded by successful CEOs who all had strong, often contradictory views on how companies should run, and it started to make him feel a little crazy. Applied Intuition, his third company, became the place where he could finally test his own instincts. The results gave him his sanity back.

“When you think about your company, you can’t take ideas from me or other people. You have to make decisions based on the reality of who you are as a person, who your co-founder is, the stage and the resources you have available, the state of your competitors and the state of the market.”

He drew an analogy to Bruce Lee, who studied every martial art and then created his own—one built for practical use, not tradition. Younis sees startup advice the same way: most of what gets said in interviews is seventy percent entertainment and thirty percent practical usability. “It’s like watching Breaking Bad for life advice,” he said. “High school teachers who get cancer just die. That’s the sad reality.”

This is to say: Take all the advice that follows with a grain of salt.

2. Compensate Against Your Values

When Applied Intuition was about a dozen people and had landed $10 million in business, Younis and his team went to a house in Santa Cruz on a Friday afternoon and wrote down why they thought they had early success. That became the company’s ten values—less a statement of aspiration than a forensic analysis of what was actually working. The hypothesis: as you grow to thirty, fifty, two hundred people, you regress to the mean. The values were meant to fight that.

“A lot of companies have values. So why do some companies maintain this very prickly shape where they’re so distinct? One of the key reasons is we compensate and we promote against adherence to our values. That’s what makes them actually real.”

The mechanism is specific: each of the ten values has roughly five observable behaviors underneath it—about fifty in total. Every individual contributor rates their manager on a one-to-five scale for each behavior, and you can’t pick three. Questions like “My manager is decisive,” “My manager is a good teacher,” “My manager is the most technically competent person on the team,” and even “My manager works harder than me.” When you ask a thousand engineers that last one, Younis pointed out, you find out very quickly which managers actually work hard—because answering “yes” means admitting you’re the lazier one.

3. Do the Opposite of What Big Companies Do

Younis has a simple heuristic that has shaped much of Applied Intuition’s culture: whatever large companies do, don’t do that. Big companies love org charts? Don’t have one until fifty people. Big companies love titles? No titles. Big companies love LinkedIn? No employee was on LinkedIn until five years into the company.

“We would have competitors—rest in peace—and I’d tell the recruiting team: go create the org chart for them. They would literally go to every one of their employees’ LinkedIn profiles, see their job titles, see what they’re doing, and reverse-engineer their org chart. And then it was like, if we can do that to them, they can probably do it to us. Everyone’s off LinkedIn.”

The no-titles policy also serves as a hiring filter: when someone interviewing is really title-oriented, that tells you something. Younis also kept his calendar public to the entire company for ninety-five percent of its history, as did his co-founder and leads. The logic: if your calendar isn’t full of product reviews and customer meetings, you should be embarrassed. “Shame is a great moral compass,” he said.

4. Raise Money as a Weapon, Even If You Never Spend It

One of the biggest things Younis said he’d changed his mind about is fundraising. He used to believe that if the company was genuinely great, everyone would just know. He now considers that a fantasy.

“I had this fantasy that if the company is really good, everyone’s going to know. They won’t. Raising money does send messages. It’s sending a very clear message to your employees, other investors, your competitors, and your customers that you’re doing well.”

Applied Intuition has raised a billion dollars and has never used any of it. The fundraising itself is the product: it’s a signal to the ecosystem, done at minimal dilution (about one percent in the last round). Younis was candid that this was a belief he had to be dragged into. He credited Marc Andreessen with pushing him to get the word out, telling him that nobody knew about this great company because of his “hermit-like view from Sunnyvale.”

Similarly, he begrudgingly admits that hyping your company matters.

“I hate to say it because I feel like I’m in middle school trying to get you to smoke or something. You’ve got to hype the company. But if I was sitting alone thinking, what do I believe today that I didn’t believe ten years ago? That’s one of them. I just felt like the good companies will get known. No.”

5. You Have to Enjoy It, or You Won’t Last

Younis has worked seven days a week since he was fourteen. And the companies Younis admires most—Berkshire Hathaway, Nvidia—are led by people who have been at it for decades. Jensen Huang has been CEO for thirty years. “A lot of people can do this for two or four years,” Younis said. “But to figure out how to do it while entire markets change and countries change and culture changes—that’s operating at the highest level.”

“One of our core values is “laugh a lot.” If you’re not having a good time and you’re not enjoying yourself, you’re not going to make it. You’re just not going to make it. The ego is such a powerful thing—you’ll rationalize all the reasons you want to give up.”

“This is a tough job,” Younis says. “You get phenomenal return for it. I still think, risk-adjusted, reward-adjusted, I would do this over Taco Bell every day of the week. That’s why I did it.”

That’s it for this week! Thanks to Qasar Younis for speaking with our founders.

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

SPEEDRUN APPLICATIONS ARE OPENING SOON! And we’re coming to SF.

Join us at the pop-up speedrun cafe (presented by Mercury) in SF during the first week of the SR007 application window: April 20th–24th! Each guest can snag a drink, a food item, and a limited-edition hat (while supplies last).

RSVP on Partiful using the links below to reserve your spot on one of the five days:

Monday (4/20) | Tuesday | Wednesday | Thursday | Friday (4/24)

This Week’s Big Ideas 💡

💻 Andrew Chen comments on the growing wave of non-technical founders using AI coding tools to build companies: “their iteration speed is 10x.”

🪞 Mirror Mirror AI (SR006) founder Yusan Lin explains why she’s building infrastructure for people to own their own likenesses in the AI era.

🛠️ Bella Nazzari explores job openings for the growing discipline of “gtm engineering.”

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Every founder is trying to crack distribution. They’re posting on X, debating whether to start a podcast, and wondering if doing the latest viral trend on TikTok is worth the effort.

Meanwhile, LinkedIn is sitting right there—wildly underutilized and practically begging early-stage founders to use it. Eventually, this channel will be saturated like everything else. But for now, for the willing founder, we believe there are customers to be had.

The LinkedIn meta is driven by two factors:

• The audience is already there, pre-qualified. Every other social media fight is a slugfest between political takes, culture war brainrot, an algorithm that changes weekly, while you’re earnestly posting about your latest feature. In contrast, LinkedIn’s feed is overwhelmingly professional. The people scrolling it are decision-makers, investors, hiring managers, and operators looking for “work” content. If you’re selling to businesses or raising money, your buyers are literally browsing between meetings. manager.

• The algorithm is generous. LinkedIn’s distribution is still in growth mode. A post from a 7,000-follower account can get 850 likes and 184 comments—that’s what happened when Salar Shahini of Sweatpals (SR004) announced his team’s $12M raise. That’s a 12%+ engagement rate! The platform actively rewards original posts, especially from personal accounts (not company pages).

What actually works on LinkedIn

After watching hundreds of founders go through speedrun, here are some tips for making performant posts:

Post your actual numbers.

The highest-performing LinkedIn posts from speedrun founders are almost always the ones sharing real metrics. No one wants to read your thoughtslop pieces about vibe coding. Just follow the example of Daniel Dhawan of Rork (SR005)—he posted that Rork users had created over 500,000 projects with 2,000+ apps published to the App Store, hitting #2 in the US Dev Tools category” and let the numbers do the work. Specificity is the currency of trust on LinkedIn. Vague “excited to announce” posts get polite likes from your mother. Actual numbers get inbound leads.

Build in public, but make it a narrative.

Taylor Offer of Atrios (SR005) makes each post build the narrative around his company. He posts about the emotional arc of founding (”Don’t forget how bad you wanted what you currently have” got 44,000+ likes), the tactical realities of scaling, and meta-observations about the platform itself. His audience feels like they’re following a character in a show. Treat LinkedIn as a serialized story to build the kind of following that actually converts into quality leads. As we’ve discussed before, early sales is about people loving the founders first and the product second.

Use your pre-founder credibility.

Your career before your startup is a feature, not baggage. Eric Berman of Lil Snack (SR002) was a Hulu and Crunchyroll executive before founding an AI gaming company. Paddy Lambros of Dex (SR005) ran talent at Atomico, so when he posts about building an AI recruiter, his network of recruiters and hiring managers actually believes what he says. If you spent years in an industry before founding a company in that space, your LinkedIn network is your first customer list.

Don’t overthink the format.

The posts getting the most engagement from speedrun founders are almost comically simple. A photo with a short story. A screenshot of a metric. Michael Lingelbach of Hedra (SR002) posts about AI video breakthroughs in a few short paragraphs—no carousels, no “10 things I learned” threads, or graphic design. His $32M Series A announcement was straightforward and still generated dozens of comments from exactly the audience Hedra wants. LinkedIn rewards “authenticity” over production value. A real photo from your office beats a Canva graphic every time.

How to survive the trough of cringe

Let’s be very clear: the first few LinkedIn posts feel terrible. You’ll write something, hover over the post button, and wonder if your college roommate or former boss is going to think you’ve lost your mind. Call it the “trough of cringe.” Every founder who’s now crushing it on LinkedIn went through it.

The discomfort is a competitive advantage. It’s the barrier that keeps your competitors from doing this. If posting on LinkedIn felt natural and easy, everyone would do it and it wouldn’t work. So start with what you know is true. Write about something you genuinely experienced this week: a customer interaction, a technical decision, a number that surprised you. The bar for “good LinkedIn content” is lower than you think because most people aren’t posting anything at all.

The cringe fades. And unlike paid ads or PR agencies, it’s free. The only cost is your willingness to hit “post.”

That’s it for this week!

For more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

Liquid Death is a healthy beverage company that looks like a beer brand, markets like an entertainment company, and grows like a tech startup. Founded in 2019 by Mike Cessario, a former ad creative with a background in making funny videos on no budget, the company has expanded from water to sparkling water, iced tea, and now energy drinks, doing hundreds of millions in revenue along the way. The company ran a Super Bowl ad this year which they produced internally for $300,000.

In a conversation with a16z speedrun, Cessario laid out the thinking behind the brand: why he assumes every product is a commodity, why comedy is the last unoccupied competitive moat in consumer goods, and what it actually takes to get a can of water onto a shelf at Kroger. Most of his advice applies far beyond beverages.

Watch the full conversation below, or scroll down for our top five takeaways from the conversation:

1. Comedy Is the One Thing Coke and Pepsi Can’t Do

Cessario’s career started in ad creative, and he’s always had a knack for making people laugh. He built Liquid Death’s competitive strategy around humor and entertainment in part because it’s structurally inaccessible to the biggest players in the market.

“Let’s say you want to do athlete marketing. Anybody with a checkbook can buy any athlete they want. If you’ve got the big enough number, you can get any snowboarder, any NFL athlete. So if you try to have a sports marketing theme for your brand, you’re just competing with every other giant brand that can just spend way more money than you can to, to, to buy more athletes or whatever. Comedy is the one thing that Coke and Pepsi can’t really do. Comedy is the one thing that Coke and Pepsi can’t really do. Even if they wanted to, true comedy can’t really survive in that bureaucracy.”

Cessario pointed to the experience of A-list comedians who’ve done deals with major brands, like Ben Stiller doing a Pepsi campaign. The story is always the same: everything funny got killed in the approval chain. “You actually had to work to make this person not funny,” he said.

2. Start by Assuming Your Product Is a Commodity

Cessario’s framework for thinking about brand starts from a premise that most founders resist: your product is probably not that different from your competitors’. He argued that founders get so deep in the weeds that they overweight tiny product differences that the broader market doesn’t notice or care about.

“Does Nike ever talk about how its materials last twenty days longer than Adidas? Any mustard on the shelf is probably all the same f***ing mustard. Can anybody pick out their favorite beer in a blind taste test? Probably not.”

His advice: start from the assumption that product differences are slim to none, then figure out how you win. Either you win on price and become the cheapest option, or you win because your commodity is far more memorable than everyone else’s. This thought exercise, he argued, is what forces you to invest in the things that actually create differentiation: brand, entertainment, and emotional connection.

3. An Entertainment Company That Monetizes via Beverage

Cessario described Liquid Death’s model inspired by the Red Bull playbook, with one substitution: comedy instead of action sports. The ethos is entertain people first, sell beverages second. That’s not a tagline. It’s the operational reality of how the company allocates resources.

“I made a video for fifteen hundred bucks that got three million views and generated more page followers than Aquafina with almost no investment. That’s proving out the model: if you can entertain people in commoditized categories, you’re going to have this huge competitive advantage because no one has any real product advantage.”

The economics have held at scale. Liquid Death’s Super Bowl ad cost $300,000 to produce internally, much lower than industry averages.

“We’re trying to make people laugh on the internet. As long as you can do that, you’re going to attract people and they have a reason to tune in. We’re occupying a place in their mind when they have five seconds to choose what to buy walking down the aisle.”

4. Cool Company, Not Just Cool Product

Asked to define what brand actually means, Cessario drew a distinction most founders miss. A lot of products are cool. Very few companies are cool. The strongest brands are the ones where people think the company itself is cool, independent of any specific product.

“There’s a lot of really cool products out there. But there’s very few cool companies. If you ask people, ‘What’s a really cool company?’ That’s where you start getting down to the people who are really good at brand.”

He used ChatGPT as a current example. Every LLM does more or less the same thing. But one brand dominates because it occupies a place in people’s heads, the way Google did with search.

Cessario also pushed back on the idea that mission alone drives loyalty. He told the story of an Uber driver who fell in love with Liquid Death’s “death to plastic” message. But the real reason she’d keep buying, he argued, is that the can gives her permission to participate in a counterculture she wouldn’t otherwise access. The mission is how she justifies it. “If that were really the only reason, the minute there was canned Smartwater, she’d buy that because it was cheaper.”

5. Distribution Is the Real Challenge

For founders who’ve only operated in software, Cessario’s description of beverage distribution is a wake-up call. Getting authorized in a retail chain is relatively straightforward. One meeting with a Walmart buyer can open four thousand doors. The problem is what happens next.

“As a small brand, you’re authorized in two thousand Krogers, but you’re only on the shelf in seven hundred, because the distributors just aren’t actually getting the product to the retailer. They don’t care, and there’s nothing you can do about it.”

The beverage industry runs on DSD (direct store delivery) networks owned by Coke, Pepsi, and the beer companies. You can’t use Coke’s or Pepsi’s distribution unless they own you. The beer networks are different because of the three-tier system: alcohol laws prevent producers from owning more than twenty percent of their distribution. So hundreds of independent distributors exist, and a new brand has to win them over one by one.

But even then, it’s a challenge to get distributors to get you priority:

“As a new brand, you’ve got to go one by one to all these guys and convince them to distribute your product ,to take your product and sell it to Walmart and Target and Kroger and 7-Eleven and everywhere else. But they’re also selling Budweiser and Bud Light and Michelob Ultra and all these other brands. So when you go in as a small brand, you’re so far down the totem pole. The rep is going into a Kroger and he’s only going to be there an hour. He needs to go to the next store. Most of his time is going to be restocking Bud Light.”

This, says Cessario, is why managing distribution is far and away the hardest part of scaling a new ready-to-drink beverage company.

Thanks to Mike Cessario for sharing his insights with us. And for more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.

🏰 Jon Lai proposes the “minimum viable moat” as the way to win in AI

🤖 Andrew Chen has a simple rubric for telling the difference between AI-native products and products with AI bolted on as an afterthought

💰 Emily Bennett has advice on fundraising: don’t lead with your “dream investor.”

💼 Join our talent network for more opportunities. If we see a fit for you, we’ll intro you to relevant founders in the portfolio

Somewhere around 2003, in a room full of investors, a video game crashed.

The game was called Republic: The Revolution, and it was supposed to be a generation-defining political simulation. Its creator, a young Londoner named Demis Hassabis, had pitched it with the kind of total conviction that makes people open their checkbooks. The demo failed. The product barely functioned. But Hassabis forged ahead, telling investors a story about what the game would become.

Ultimately, the game shipped to middling reviews, and Hassabis’s company, Elixir Studios, folded. Hassabis went quiet for a few years, then started a small research lab called DeepMind. Everyone in Silicon Valley knows the story from here. Google acquired DeepMind and the company has gone on to develop technologies that power Google’s frontier AI models, and which have earned Hassabis the Nobel Prize in Chemistry.

Sebastian Mallaby has been writing about entrepreneurs like Hassabis for twenty years. His book More Money Than God went inside hedge funds, The Power Law chronicled venture capital, and his new book, The Infinity Machine, tells the story of Hassabis and, through him, the story of how modern AI came to exist.

We spoke with Mallaby about what he’s observed across all three of these worlds, and why the Hassabis story might be more useful to founders than they realize.

Conviction Vs Delusion

Hedge funds, startups, venture capital: in each case, a small group of people has decided that everyone else has it wrong, and is betting real money on that belief. These are all fields where the gap between conviction and delusion is essentially invisible from the outside. The only thing that separates them is the outcome.

We asked Mallaby: Is this tension between conviction and delusion the real throughline connecting your work?

“I think that’s a fair characterization of the whole investment boutique entrepreneurship ecosystem,” Mallaby said. “In hedge funds, you’ve got market prices which you think you can outguess, and those market prices are the weighted average of intelligent people’s judgment. So you need to be quite bold to think you know better. In the same way, a startup entrepreneur is trying to disrupt incumbents who are not idiots. You have to be bold.”

But the line between conviction and delusion can at times appear thin. And as Mallaby tells it, even the most successful founders have found themselves on both sides of that line.

The Fluency Test

One of the early stories told in The Infinity Machine recounts how the dev team on Republic: The Revolution slipped into a pattern of delusional belief about the game’s chances at success. “The engineers knew he [Hassabis] wouldn’t accept anything short of ‘yes, we can do this,’” Mallaby told us. “So they said that. And that made Hassabis even more determined not to accept no for an answer. Because he was hearing yes from his own team, he got even more confident. There was this feedback loop.”

What matters about this story is what happened after. Mallaby frames the Elixir collapse as the event that made DeepMind possible.

Years later, facing a similar crossroads with AlphaFold, Hassabis found himself in the position of deciding whether to push a team past the point where its own leader said the goal was impossible. The head of the AlphaFold effort, Andrew Senior, told him flatly in 2018: we’ve built the best protein-folding system in the world, but we cannot do what you’re asking us to do. Instead of overriding him, Hassabis began sitting in on meetings with the broader team and listening closely for what Mallaby calls “the fluency of the conversation.”

“If lots of ideas are coming up, he figures, well, it’s not like they’re stuck,” Mallaby says. “If they had more time or resources, they could test out some of these ideas they’re kicking around in the conference room. Fluency was his metric for whether to go ahead.”

A room full of people generating ideas looks fundamentally different from a room that’s gone quiet. The first means there’s potential. The latter means the problem might be a dead end. The fluency test was how Hassabis learned to tell the difference, and it was a lesson hard-won by a founder who’d experienced the downsides of charging ahead when the room was telling him what he wanted to hear.

Mallaby put the takeaway plainly: “I think it’ll resonate with a venture capital firm that second-time entrepreneurs are often the ones you most want to back.”

On Communicating Conviction

For many practical and structural reasons, the dominant way that Silicon Valley founders talk is prospective. You announce what you’re going to build before you’ve built it. You generate excitement around a future state that may or may not arrive. Some of the most visible figures in tech have operated this way for years, tweeting about ambitious new capabilities before the technology can deliver them. It works as a fundraising mechanism. It also, as Mallaby noted dryly, sometimes involves announcing things that aren’t yet certain to work.

Hassabis, at least in his mature period, worked in the opposite direction. After DeepMind’s Atari system started producing real results in 2013, his instinct was to take the concrete achievement and wring every possible drop of credibility out of it. The target he picked was Nature magazine. “Nobody else on his team thought that publishing a paper in Nature was even conceivable,” Mallaby said. “Nature didn’t really do computer science. But Hassabis cultivated an editor at Nature for a year, then got him to agree to look at a submission from DeepMind, then got the DeepMind video team to design cover art so Nature would put it on the front.”

These kinds of details accumulate throughout the book. Hassabis comes across as someone who wastes very little, who finds a use for whatever’s already lying around, and who treats reputation as a resource to be deployed strategically rather than as a mere byproduct of the work.

“He was absolutely as focused on communication as anyone else in AI,” Mallaby said. “But he had this retrospective approach, of playing up actual achievements, as opposed to announcing future ones.”

On Timing and “Scientific Taste”

Marc Andreessen likes to say that one of the main reasons great entrepreneurs fail isn’t because their idea is wrong, but because they were simply too early. Predicting when the market is ready for something is one of the hardest jobs of the entrepreneur, and a thread running through Mallaby’s book is that Hassabis foresaw a lot of things before others, including the possibilities of reinforcement learning and the possibility of delivering something meaningful for the protein folding problem.

Mallaby told us that Hassabis uses the phrase “scientific taste” to describe the ability to sense when a field is about to tip, even when most observers still see stagnation. It’s a phrase that does a lot of quiet work, treating taste as pattern recognition operating at the boundary between what the data supports and what it merely suggests: the ability to read straws in the wind when everyone else is staring at the ground.

“With DeepMind’s Atari agent, which is an early experiment in combining deep learning with reinforcement learning, Hassabis sees when something is poised to advance,” Mallaby said. “And then he repeats the same trick with the decision to go after AlphaGo, even though Sergey Brin had told him it’s crazy. After Atari he feels you can do Go. Later, after Go, he feels you can do protein folding.”

David Silver, one of DeepMind’s key researchers, gave Mallaby a metaphor for this: Hassabis had understood “the ladder.” Each project was grander than the last, and each one funded the next: financially, and in terms of credibility. To get to the moon, the DeepMind team figured out which rung to stand on next.

Internally, the DeepMind team ran two modes at once: a cheap exploration layer, where researchers pursued open-ended questions, and an expensive exploitation layer, where strike teams poured everything into a single target. Hassabis’s job, as Mallaby described it, was knowing when to shift from one to the other. “He’s willing to have people exploring different avenues for AI, doing their thing,” Mallaby said. “But then he’s going to have the scientific taste to know which one to push on. And when he’s pushing, it’s a totally different organizational structure.”

How to Predict the Future

We asked Mallaby whether, after spending years embedded in this story, he felt he understood where AI was actually going. He gave a modest yes.

“My view is, if you look at what’s happened since ChatGPT came out, the number of major jumps that the technology has made is flabbergasting,” he said. “You go from something which basically hallucinated nonstop to something that pretty much got that under control. Longer context windows. Multimodal. Reasoning. Now it’s starting to be agentic. Next is world models. And it’s only been three and a half years.”

Every time someone declares a ceiling, Mallaby pointed out, a new axis of scaling appears to replace the one that supposedly ran out. He’s bullish on infrastructure. The application layer, he admitted, is anyone’s guess.

But the question we kept circling back to was the one about conviction and delusion, and whether there was any reliable way to tell them apart before the outcome settles the matter. Mallaby’s answer, when it finally came, had the weight of someone who has spent two decades in rooms with people who stake everything on their own judgment.

“You don’t necessarily get the balance correct the first time,” he said. “But the only way you learn is by doing.”

The Elixir game crashed, and the company died, and then the same person started something else.

Thanks to Sebastian Mallaby for speaking with us. And for more weekly dives into the world of early stage startups, subscribe below.

You are receiving this newsletter since you opted in earlier; if you would like to opt out of future newsletters, you can unsubscribe immediately.

This newsletter is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. This newsletter may link to other websites and certain information contained herein has been obtained from third-party sources. While taken from sources believed to be reliable, a16z has not independently verified such information and makes no representations about the enduring accuracy of the information or its appropriateness for a given situation.

References to any companies, securities, or digital assets are for illustrative purposes only and do not constitute an investment recommendation or offer to provide investment advisory services. Any references to companies are for illustrative purposes only; please see a16z.com/investments. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund which should be read in their entirety.) Past performance is not indicative of future results.

Charts and graphs provided within are for informational purposes solely and should not be relied upon when making any investment decision. Content in this newsletter speaks only as of the date indicated. Any projections, estimates, forecasts, targets, prospects and/or opinions expressed in these materials are subject to change without notice and may differ or be contrary to opinions expressed by others. Please see a16z.com/disclosures for additional important details.