Connect with me on X and LinkedIn.

Check Out NoStack
Save Substack Content To Notion.

Most marketers negotiate against themselves.

Imagine you’re in the middle of showing your offer deck to a prospect. You’re on slide three.

You’ve laid out the scope. The buyer is nodding, but something in their eyes tells you they already have a number in their head.

They do. That number came from your LinkedIn bio.

They read the bio before they ever booked the call. They slotted you into a category the moment they saw your title, and that category came with a price range attached to it.

You aren’t negotiating from your offer. You’re negotiating against a ceiling you put there with your own bio.

How the category gets built

A buyer hears “marketing consultant” and opens a drawer in their head. Inside the drawer sit reference points, rough prices, names of other vendors they’ve seen, and the questions they know to ask a marketing consultant.

By the time you get on the call, you are already inside that drawer. You can feel it in the questions the buyer asks. They sound the same as the questions every other marketing consultant got asked before you.

If the buyer had read a phrase that didn’t match any drawer in their cabinet, the call would look different. They would lean forward and ask you what you mean.

Now you’d be teaching them a new category instead of defending your price inside one someone else owns.

The deck is the wrong lever

Most marketers try to fix this on slide twelve. They rewrite case studies, sharpen the pricing anchor, polish the logo wall.

None of it moves the ceiling. You set the ceiling on slide zero, before the call started, when the buyer clicked your LinkedIn profile.

The deck is the wrong place to rewrite the category. The bio is where the category lives.

Four-sentence rewrite

A founder I worked with last year kept getting inbound leads asking for his hourly rate. His About section said “marketing consultant.” He was running strategy, paid, email, and positioning for every client on his roster.

His work sat at a fractional CMO level but his label sat at a freelancer level. He had already priced his own inbound with his bio before his first call of the day.

He rewrote four sentences.

Nothing else in the funnel changed. Over the next quarter he moved his average deal from $2,800 a month to $9,500 a month.

What those four sentences did

The first sentence named the slice of buyer he worked with. He wasn’t pointing at marketers in general. He named a specific kind of founder with a specific kind of problem.

The second sentence named the outcome those buyers walked out with, in numbers. He didn’t write “growth.” He wrote a measurable delta with a dollar sign on the end of it.

The third sentence named his method, and the method didn’t match any drawer in the buyer’s cabinet. Buyers had to ask him what he meant. That question was the whole point.

The fourth sentence gave proof. He named real clients with real numbers.

Stress test yours

Open your own profile.

Read the first line out loud.

Ask yourself what drawer a stranger would drop you into.

Now ask what the price range is inside that drawer. If the range sits below the number you want to charge, you are paying the gap on every call you take.

Rewrite one sentence this week.

Pick the one that keeps dropping you into the wrong drawer.

-G Out.

Say whats up on LinkedIn & X (formerly known as Twitter)

(unapologetically generated with Claude.ai assistance, lol)

Two user migrations are running in opposite directions right now.

Consumers are flooding to Claude because Anthropic refused to hand the Pentagon unrestricted access to its AI. Developers are leaving Claude Code for OpenAI’s Codex because Anthropic broke their workflows, burned their token budgets, and locked out every third-party tool they depended on.

Both migrations started from the same set of decisions. I tracked every event from the first rate limit announcement through today’s official third-party lockout. Thirty-two events across five phases.

The full arc tells a story that no single headline captures.

Phase 00: The Unsustainable Economics (Jul-Dec 2025)

July 28, 2025. Anthropic announces weekly rate limits, effective August 28. The reason: a small number of power users are running Claude Code around the clock. One person consumed tens of thousands of dollars in model usage on a $200/month plan. Claude Code had seven outages in a single month from demand pressure.

August 12, 2025. Anthropic ships the 1M token context window in public beta for Claude Sonnet 4. The jump from 200K to 1M is a 5x increase. Premium pricing kicks in above 200K tokens. API-only access, restricted to Tier 4+ customers.

August 28, 2025. Weekly rate limits go live. Pro users get 40-80 hours of Sonnet per week. Max 20x users get 240-480 hours of Sonnet plus 24-40 hours of Opus. Anthropic says less than 5% of users will feel the change.

October 9, 2025. A mega-thread launches in Anthropic’s Discord channel. Users start documenting how fast they hit limits. Some Pro users burn through their weekly allocation in 40 minutes. The thread becomes the first sustained record of community backlash.

November 2025. An Austrian developer named Peter Steinberger publishes Clawdbot, his 44th AI project. It runs autonomous agents through messaging platforms like WhatsApp and Telegram. Nobody outside a small dev community pays attention.

December 25-31, 2025. Anthropic doubles usage limits as a holiday gift. Enterprise customers are on vacation, so the compute sits idle. Users recalibrate their expectations to the higher baseline. Five days later, that recalibration becomes a problem.

Phase 01: The Silent Lockout (Jan 2026)

January 4-5, 2026. The holiday bonus expires. Users report a roughly 60% reduction in available usage compared to what they got during the holidays. Complaints flood the Discord channel. Multiple users claim moderators are deleting critical posts.

January 8, 2026. Someone files a GitHub issue for Claude Code version 2.1.1. Max plan users are hitting their 5-hour limits 4x faster than previous versions. Same projects. Same routines. Same codebases. The update broke something.

January 9, 2026. Anthropic deploys server-side checks with zero advance notice. Every third-party tool using Claude Pro or Max OAuth tokens stops working overnight. OpenCode, Cline, RooCode, and Clawdbot all go dark. The error message reads: “This credential is only authorized for use with Claude Code and cannot be used for other API requests.” No blog post. No announcement. No warning.

January 10, 2026. Anthropic engineer Thariq Shihipar confirms the changes on X. “Yesterday we tightened our safeguards against spoofing the Claude Code harness.” The main GitHub thread collects 147+ reactions. The Hacker News post hits 245+ points.

January 15, 2026. George Hotz publishes a blog post titled “Anthropic is making a huge mistake.” His prediction: “You will not convert people back to Claude Code, you will convert people to other model providers.”

January 27, 2026. Anthropic sends a trademark complaint over the name “Clawdbot” because of its similarity to “Claude.” Steinberger renames the project to Moltbot. Three days later he renames it again to OpenClaw because Moltbot “never quite rolled off the tongue.” Three names in 72 hours. Domain squatters and fraudulent crypto tokens flood the confusion.

January 28, 2026. Anthropic confirms it has implemented client fingerprinting. The system detects and blocks any request that spoofs the official Claude Code client identity via HTTP headers. Request origin validation and API endpoint restrictions are now active.

Late January 2026. FEC filings reveal that OpenAI president Greg Brockman and his wife donated $25 million to MAGA Inc. A DHS AI inventory shows that ICE is using a resume screening tool powered by ChatGPT-4. These two data points plant the seeds of QuitGPT.

Phase 02: OpenClaw + QuitGPT (Feb 2026)

Early February 2026. OpenClaw passes 145,000 GitHub stars with peak traffic of 2 million visitors in a single week. Users create 1.5 million AI agents on the platform. It becomes the fastest-starred open-source AI project in history. Nvidia CEO Jensen Huang later calls it “probably the single most important release of software ever.”

February 2, 2026. OpenAI launches the Codex desktop app for macOS. It pulls 200,000 downloads on day one. The app lets developers fire off coding tasks and come back to review results later. For Claude Code users frustrated by limits, Codex becomes the first viable alternative with polish.

February 5, 2026. The QuitGPT campaign goes live. Actor Mark Ruffalo posts to his followers: “ChatGPT’s President is Trump’s biggest donor. Their tech powers ICE. It’s time to boycott.” The post pulls 36 million views and 1.3 million likes.

February 14, 2026. Steinberger publishes a blog post announcing he is joining OpenAI. “I’m joining OpenAI to work on bringing agents to everyone. OpenClaw will move to a foundation and stay open and independent.” Running the project was costing him $10,000-$20,000 per month out of pocket.

February 15, 2026. Sam Altman confirms the hire on X at 4:39 PM. “Peter Steinberger is joining OpenAI to drive the next generation of personal agents. He is a genius with a lot of amazing ideas about the future of very smart agents interacting with each other.” The post pulls 16.7 million views.

February 17-18, 2026. Anthropic updates its Terms of Service with a new “Authentication and credential use” section. The language bans subscription OAuth tokens in any third-party tool, including Anthropic’s own Agent SDK. OpenCode removes all Claude OAuth code from its codebase the same day.

February 19, 2026. David Heinemeier Hansson (creator of Ruby on Rails) weighs in: “Terrible policy for a company built on training models on our code, our writing, our everything.”

Phase 03: The Pentagon Standoff (Feb 24 - Mar 3, 2026)

February 24, 2026. Two things happen on the same day. Anthropic ships Claude Opus 4.6, the first Opus-class model with a 1M token context window. State-of-the-art scores on Terminal-Bench 2.0 and Humanity’s Last Exam. And Defense Secretary Pete Hegseth tells Anthropic CEO Dario Amodei: drop all restrictions on military use by Friday evening, or lose the $200 million contract.

February 27, 2026. Anthropic rejects the Pentagon’s final offer. Amodei publishes a statement: “We cannot in good conscience accede to their request.” He says mass surveillance and autonomous weapons are “outside the bounds of what today’s technology can safely and reliably do.” Trump posts on Truth Social calling Anthropic “Leftwing nutjobs” and directs every federal agency to stop using Claude. Hegseth designates Anthropic a “supply chain risk to national security.”

February 28, 2026. OpenAI announces a Pentagon deal within hours of Anthropic’s collapse. Altman says his company will “deploy our models in their classified network.” ChatGPT uninstalls spike 295% in a single day. One-star reviews surge 775% above baseline. QuitGPT claims 2.5 million participants across cancellations, pledges, and social media shares.

March 1, 2026. Claude hits #1 on the US App Store, overtaking ChatGPT for the first time. Free users grow 60% since January. Paid subscribers more than double. Daily downloads jump 51% on February 28 alone.

March 2-3, 2026. Claude goes down multiple times. The authentication infrastructure cannot handle the surge. Consumer services go offline for hours. Anthropic attributes the outage to “unprecedented demand.”

Phase 04: Exodus + Fallout (Mar-Apr 2026)

March 13, 2026. Anthropic makes the 1M token context window generally available for Opus 4.6 and Sonnet 4.6. The premium pricing surcharge disappears. A 900K-token request now costs the same per-token rate as a 9K request. The flat pricing accelerates the token burn problem. Developers loading entire codebases into single prompts chew through their limits faster than ever.

March 19, 2026. OpenCode maintainer Dax Raad merges PR #18186 with the commit message “anthropic legal requests.” The PR removes the Anthropic OAuth plugin, the Claude system prompt, and every reference to Claude Pro/Max authentication. The GitHub reactions tell the story: 4 thumbs-up. 40 thumbs-down. A 125,000-star project is locked out of Claude for good.

March 24, 2026. A California federal judge grants Anthropic’s motion for a preliminary injunction. The ruling blocks the Pentagon from enforcing its supply-chain-risk designation. Judge Rita Lin writes that the government’s actions were “classic illegal First Amendment retaliation” and were not designed to protect national security. The government has seven days to appeal.

March 28-31, 2026. The token burn crisis reaches its peak. Anthropic posts on Reddit: “We’re aware people are hitting usage limits in Claude Code way faster than expected.” A developer who reverse-engineered the Claude Code binary finds two independent bugs that cause prompt cache to break, inflating costs by 10-20x. Some users report that typing “Hello” consumes 2-13% of their entire session limit.

March (ongoing). Max plan users start migrating to OpenAI’s Codex. The Reddit consensus crystallizes: “After 5 months of $200 Claude plan, I downgraded to $20, and copped $200 Codex.” The common advice becomes: drop to the $100 Claude Max plan, add ChatGPT Plus for $20, use both tools for $120/month instead of $200. OpenAI reports Codex usage within Business and Enterprise has grown 6x since January. More than 2 million builders use Codex weekly.

April 1, 2026. Claude Code’s source code leaks. A developer error (missing .npmignore) exposes 512,000 lines of proprietary code. Anthropic issues DMCA takedown notices that accidentally nuke thousands of unrelated GitHub repositories. The company retracts the bulk of the notices and is forced to open-source Claude Code.

April 3, 2026. Today. Anthropic announces that Claude subscriptions will no longer cover usage on third-party tools like OpenClaw. Starting April 4 at 12pm PT, you need separate pay-as-you-go billing or API keys. Anthropic offers a one-time credit equal to your monthly plan cost and discounted usage bundles. The lockout arc that started with silent server-side blocks on January 9 is now official company policy.

The Two Migrations

You can trace every event on this timeline back to one tension: flat-rate subscriptions cannot absorb agentic workloads at scale.

A $200/month Max subscription running Opus-level autonomous loops would cost $1,000+ on API billing. The “Ralph Wiggum” technique (trapping Claude in self-healing loops that run overnight) went viral in late 2025 and blew out Anthropic’s infrastructure economics. Third-party tools like OpenCode made those loops easier to run. Anthropic’s costs exploded.

Anthropic locked everything down. They blocked third-party OAuth. They fingerprinted clients. They sent lawyers to OpenCode. They formalized the ban in their Terms of Service. They tightened rate limits during peak hours. And today they cut third-party subscription coverage entirely.

The developer response was predictable. Power users migrated to Codex. OpenCode integrated with OpenAI. Steinberger took OpenClaw to OpenAI. The tools followed the economics.

Then the Pentagon standoff happened. Anthropic refused to allow Claude for autonomous weapons and mass surveillance. Trump blacklisted them. OpenAI swooped in. 2.5 million ChatGPT users revolted. Claude hit #1 on the App Store. Anthropic’s paid subscribers doubled.

Two migrations. One company.

Consumers moving to Claude because Anthropic drew a line the Pentagon would not accept. Developers moving away from Claude Code because Anthropic drew a line the ecosystem could not sustain.

Both decisions came from the same leadership team applying the same logic: we control how our technology gets used. The Pentagon could not override Anthropic’s safety policies. Developers could not override Anthropic’s billing policies. Same posture. Different audiences. Opposite reactions.

The question for anyone building on Claude right now: which side of the migration are you on, and how long does your side hold?

“Just to be clear - I LOVE Claude Code. Anthropic is the best model in the world. I have no doubt. However - right now we’re at a bit of an impass in how its working in Claude Code.

Guys - can we get this under control?
-G.”

Subscribe now

The 1M Token Tax

Your Claude Code Max plan burns faster than it used to. The math explains why.

The Setup

Claude Code on Max runs Opus 4.6 with a 1M token context window. Anthropic ships it as the default for Max, Team, and Enterprise users. No surcharge. Flat per-token rate.

A 900K-token request costs the same per token as a 9K one. (Source: Anthropic, 1M Context GA Blog)

Users heard “no extra charge” and let their sessions run deep. Then the 5-hour allocation started disappearing in minutes.

The API Is Stateless

Claude has no memory between turns. Every message you send, the application re-bundles the entire conversation history and submits it as a fresh request.

“The Messages API is stateless, which means that you always send the full conversational history to the API.”

(Source: Anthropic API Docs)

In Claude Code, “turns” include every file read, every grep, every test execution, every tool call. Each one appends output to context. The next action re-ingests all of it.

Attention Scales Quadratically

Transformer self-attention computes relationships between every token pair. Double the context length, quadruple the compute.

“Compute requirements scale quadratically with the length of a sequence: if the number of input tokens doubles, the model needs 4 times as much processing power to handle it.”

(Source: IBM, “What is a context window?”)

Anthropic has not disclosed which optimizations Claude uses internally. The system card confirms Opus 4.6 is a dense transformer with no mixture-of-experts. (Source: Medium, “Claude Opus 4.6 Architecture and Training”)

DeepSeek, which publishes full architecture details, built three separate systems (MLA, MoE, Sparse Attention) to fight this problem and still acknowledges:

“The quadratic complexity inherent in Transformer-based autoregressive decoding remains a formidable challenge, especially for extremely long contexts.”

(Source: DeepSeek-V3, arxiv.org)

Every frontier model faces this. Every lab mitigates it differently. None have eliminated it.

The Compounding Loop

Claude Code runs an agentic loop. Each step (file read, shell command, code search) appends output to conversation context. The next step re-ingests everything before it.

A 30-step session against a medium codebase can look like this:

Step Context Size (approx.) Compute Cost Ratio vs. Step 1 1 50k tokens 1x 10 200k tokens ~16x 20 500k tokens ~100x 30 800k tokens ~256x

Ratios are theoretical O(n²). Actual ratios depend on undisclosed optimizations. Prompt caching reduces re-computation of stable prefixes but does not eliminate token counting against plan limits.

Thinking tokens stack on top. Extended thinking defaults to enabled at “high” effort on Opus 4.6. Each thinking pass can burn tens of thousands of output tokens per step.

“Thinking tokens are billed as output tokens, and the default budget can be tens of thousands of tokens per request depending on the model.”

(Source: howdoiuseai.com, “Claude Code Pricing”)

Output tokens on Opus 4.6 run $25 per million. Every turn. On top of the growing input context.

Where Retrieval Quality Drops

Anthropic published MRCR v2 benchmark results in the 213-page Opus 4.6 system card. The benchmark hides 8 pieces of information in long contexts and tests whether the model finds them.

The system card defines its test bins:

“For the reported variants, 256k bin boundaries represents prompts with (128k, 256k] tokens, and 1M represents bin boundaries with (524k, 1024k] tokens.”

(Source: Opus 4.6 System Card, p.30)

Results from Table 2.18.A:

Context Bin Score (64k thinking) 256k (128k-256k tokens) 91.9% 1M (524k-1024k tokens) 78.3%

13.6 point drop once context crosses into the 524k+ range.

No public benchmark data exists for the 256k-524k gap. The degradation could start at 300k or hold steady until 500k. Anthropic does not say.

The most defensible read: below 256k, retrieval sits above 90%. Above 524k, it falls to 78%. The 500k mark is a reasonable estimate for where quality starts to fall, not a measured data point.

The Pricing Signal

Before March 2026, Anthropic charged a premium above 200k tokens. Opus input went from $5 to $10 per million. Sonnet from $3 to $6. (Source: Anthropic, “Introducing Claude Opus 4.6”)

They removed the surcharge at GA. (Source: Anthropic, 1M Context GA Blog)

The surcharge existed because long-context requests cost more to serve. Removing it means Anthropic absorbs that compute cost. On Max plans with fixed monthly allocation, more tokens burned per session means fewer sessions per allocation. The per-token rate is flat. The per-session cost is not.

Compaction Confirms It

Anthropic built automatic context summarization into the product:

“Compaction provides automatic, server-side context summarization, enabling effectively infinite conversations. When context approaches the window limit, the API automatically summarizes earlier parts of the conversation.”

(Source: Anthropic, “What’s new in Claude 4.6”)

Summarization is lossy. A Ramp engineer on Anthropic’s own blog:

“Claude Code can burn 100K+ tokens searching Datadog, Braintrust, databases, and source code. Then compaction kicks in. Details vanish. You’re debugging in circles.”

(Source: Anthropic, 1M Context GA Blog)

If serving large contexts were cheap, compaction would not exist.

The Cache Bugs That Made It Worse

In late March 2026, two software bugs in Claude Code broke prompt caching for affected users. The structural costs above (quadratic scaling, stateless re-ingestion) became 10-20x worse.

Bug 1: Bun Binary String Replacement

The standalone Claude Code binary ships with a custom Bun fork. A community reverse-engineer found a native-layer string replacement that modifies the JSON request body after serialization, breaking the cache prefix on every request.

“It targets a billing attribution sentinel (cch=24b72) and replaces it with a 5-character hash...if your conversation history contains the sentinel string, the wrong instance gets replaced.”

(Source: DEV Community, “Claude Code Is Silently Burning 10-20x Your Token Budget”)

Workaround: use npx @anthropic-ai/claude-code instead of the standalone binary.

Alex Volkov amplified this on X. Over 500 people confirmed the issue. (Source: Alex Volkov on X)

Bug 2: --resume Breaks Cache

Starting from v2.1.69, resuming a session via --resume caused a complete cache miss on every request.

“--print --resume sessions stopped caching conversation turns between API calls starting around v2.1.69. Only Claude Code’s internal system prompt (~14.5k tokens) is cached; all conversation history is cache_created from scratch on every message. This causes a ~20x cost increase per message compared to v2.1.68.”

(Source: GitHub Issue #34629)

One user documented 652,069 output tokens generated during a session resume with zero user input, costing $342.74 in a single session. (Source: GitHub Issue #38029)

Community workaround: avoid resuming sessions entirely. Start fresh. Use CLAUDE.md to restore project context. (Source: GitHub, ArkNill/claude-code-cache-analysis)

Anthropic patched the --resume regression in a subsequent release:

“Fixed --resume causing a full prompt-cache miss on the first request for users with deferred tools, MCP servers, or custom agents (regression since v2.1.69).”

(Source: Claude Code Changelog)

The 500k Sweet Spot

One testimonial on Anthropic’s own blog:

“We raised our Opus context window from 200k to 500k and the agent runs more efficiently. It actually uses fewer tokens overall. Less overhead, more focus on the goal at hand.”

(Source: Anthropic, 1M Context GA Blog)

Below 500k, retrieval quality sits in the low 90s (measured at the 256k bin). The agent finds what it needs without re-searching. Above 524k, retrieval drops to 78%, which triggers compensating behavior: retries, re-reads, broader searches. Each retry burns more tokens.

A Max plan user running Opus 4.6 at 500k context will likely get more useful work per dollar than the same user at 900k.

The Full Picture

Seven facts produce the observed outcome:

1. Claude’s API is stateless. Every turn re-sends full conversation history. (Anthropic API Docs)

2. Transformer attention scales quadratically with context length. (IBM, DeepSeek)

3. Claude Code’s agentic loop appends output on every step, growing context through each session. (Anthropic Blog)

4. Retrieval quality drops 13.6 points between the 256k and 524k+ bins on Anthropic’s own benchmark. (Opus 4.6 System Card, Table 2.18.A)

5. Degraded retrieval triggers compensating behavior: retries, re-reads, broader searches. Each retry adds more tokens.

6. Thinking tokens (output at $25/M for Opus) compound on every turn and default to “high” effort. (Anthropic Docs)

7. Two cache bugs in Claude Code v2.1.69+ broke prompt caching for affected users, inflating consumption 10-20x. (GitHub #34629, GitHub #38029, Alex Volkov on X)

The per-token rate is flat. The per-session cost, driven by compounding context in an agentic loop, is where Max plan users feel the squeeze. The cache bugs turned that squeeze into a crisis for a subset of users during the same week the Spring Break promotion ended.

The Value Takeaway:

The 1M context window is a capability ceiling. For most Claude Code work on Max plans: keep context under 500k, start fresh sessions for new tasks, check cache health via /cost, and treat the full window as a precision tool for the jobs that need it.

All benchmark data from the Opus 4.6 System Card (213 pages, February 2026). Architectural claims from peer-reviewed research and published technical reports. Cache bug analysis from GitHub issues in the anthropics/claude-code repository and community analysis.

Editorial inferences labeled in text.

I wrote this with Claude O4.6EXT on Claude.ai because can you imagine if I tried to do all this myself?

Happy Good Friday.

I drilled through over 1,000 pieces of content from the top educators, growth operators, and audience builders on the internet so I want to drill down on the exact thing I see between all of them that works for every single creator, marketer, agency & coach.

Transcripts, breakdowns, frameworks, recorded workshops. I organized them into five separate knowledge bases by philosophy and source material. It’s just about everything I have consumed over the past 2 years.

There are 3 things that make up the math I’m talking about. It’s not a “hard math” as in numbers but it’s a repeatable formula.

When you apply the equation through a concept I call “Ruthless Clarity” you get the same outcome. That outcome is the content doing exactly what you built it to do. Booked calls. Pipeline. Shares from the right people. Revenue.

(Targeted Attention × Credible Value) ÷ Friction

Targeted Attention

You need people to stop scrolling. You know this.

The part most creators keep getting wrong is who you need to stop.

There are two kinds of virality.

• Pure virality chases the biggest possible number.

• On-target virality chases the deepest possible penetration within one specific buyer slice.

Agencies & creators pulling seven figures from content chose the second. The ones chasing ten million views and wondering why nobody books a call chose the first.

Your job is to train the algorithm to understand who you serve. You do that by staying on one topic for one buyer, video after video, month after month.

Discipline over novelty.

Then sharpen it further. Position through a contrarian belief. A take that runs against whatever your competitors repeat.

You are looking for the gap between what the market assumes and what you know. That gap is the moat. Nobody can copy a position they do not hold.

Packaging is almost as important as the content itself. Title and thumbnail come first. If you cannot build a title and thumbnail you believe will get clicked, kill the idea. You did not waste production time. You saved it.

Once someone clicks, you have 15 seconds. Call out who the video/content is for. Prove your credibility on this specific topic. Give the viewer/reader a roadmap and deliver a learning before the 30-second mark.

The old advice to bury the value stopped working when supply became infinite.

Nobody waits anymore.

Credible Value

If you grab attention and the content behind it is thin, you train your audience to stop clicking. You may see your views go up but views don’t equal booked calls. People remember when you waste their time and they do not come back to give you a second chance.

The content has to solve a problem your viewer is dealing with right now. Not a vague category problem but a specific, named thing they felt this morning when they opened their laptop or looked at their numbers. And the solve has to be something they can use today.

Not inspiration. Not a mindset shift. A framework, a process, a template, a way of thinking about the problem that changes how they operate by the end of the week.

I kept seeing the same concept: compete on generosity. You know the phrase as “give value”. Give away the stuff you think is too good to give away.

The other half of this variable is credibility, and credibility is weirdly specific.

Having general authority in your space helps, but what actually changes behavior is proof that you have done the exact thing you are teaching.

The same lesson delivered by someone with surface-level experience and someone who has been inside the problem for a decade lands with completely different weight.

People can feel the difference even if they cannot name it. This is where you want to be.

If you run an agency or you are operating at any real scale, you are sitting on a pile of proof most solo creators would kill for.

Client results. Revenue numbers. Before-and-after transformations with real names attached.

Every piece of content you publish should pair the problem you are solving with evidence that you have already solved it, for someone your viewer can see themselves in.

Friction

This is the denominator in the equation, and it is the part that will sink your battleship if you are not paying attention.

You know your subject cold.

You have years of pattern recognition, a vault of frameworks you built yourself, and a vocabulary that makes sense to people at your level.

So you dump all of that into your content, and your audience bounces at the 40-second mark because their brain ran out of room. You may know this as “I don’t have the emotional or mental capacity to watch this right now”.

Working memory holds about three new chunks of information at a time. Push past that and the viewer starts dropping things. Maybe they lose the original promise of the video. Maybe they lose the thread of your argument. Maybe they hit a word they do not know and spend three seconds trying to figure it out while you keep talking. Any of those and they are gone.

If a smart eight-year-old cannot follow your video, you wrote it for yourself. The largest YouTube channels, the bestselling books, the highest-converting landing pages all write below an eighth-grade level. Complexity is not a sign that you know more. I believe it is a sign that you have not done the work to make it simple.

The structure almost always appears in the content as Problem, Example, Lesson.

You state a problem your viewer feels, tell a story that makes it real, and pull one clear takeaway out of it.

Then you do it again.

With each cycle, your working memory should be clear before the next one starts, which means you can run this pattern dozens of times in a single video without overloading anyone.

It’s well known that a human’s eyes process information somewhere between 10 and 100 times faster than their ears.

That would typically translate into seeing your graphic, hearing your words, then glancing back at the graphic to check that the two match.

If your spoken hook says one thing and the screen shows something different, comprehension drops and the viewer leaves. Match them. Every single time.

All of the best content gave the value crazy fast. (those of you in startup world may know this as “time to value”)

Pull up your last five videos, find the first line where real teaching starts, and cut everything above it.

The Checklist

Before you write a word of script, lock four things down:

• the one buyer this piece is for

• the painful problem you are solving

• the contrarian angle that separates you from everyone else talking about this topic

• a title and thumbnail you believe will get clicked.

Before you hit record write your opening so it calls out who the video is for to establish your credibility on this specific topic, gives a roadmap, and delivers a real learning before the 30-second mark.

Build the body as Problem-Example-Lesson cycles.

Write every sentence at a fifth-grade reading level and keep cutting until it hurts.

Before you publish, check that your visuals and audio say the same thing on the hook.

Make sure you are never juggling more than three new ideas at once.

End with a new problem instead of a summary, and point the viewer somewhere.

The whole industry obsesses over the numerator. Better hooks, bigger promises, louder thumbnails. That work matters, but the denominator is where the real leverage hides.

Shrink friction and everything above the line multiplies harder.

(Targeted Attention × Credible Value) ÷ Friction = Content that converts.

-G out.

And I get why you may think that.

The volume went up. The average quality went down.

And now everyone's feed looks like it got copy-pasted from 400 different people who all read the same prompt.

But content was already bad before ChatGPT existed.

It was bad in 2021. It was bad in 2016 when everyone was posting "5 tips to scale your agency" with a Canva graphic and a prayer.

AI didn't break content.

AI just made it easier to produce more of what was already broken. Faster. At scale. With better grammar and worse conviction.

What I believe the problem actually to be is that most people sit down to create content without a clear idea of what they actually believe.

They open a doc or a prompt window and start reaching for something that sounds like content. Something that checks the box of "I posted today."

And what comes out is this vague, diluted, half-committed thing that technically says words but doesn't actually say anything.

This is the normal stack for most agencies, creators, coaches.

→ They've got a content calendar.
→ A scheduler.
→ A Notion board.
→ A Canva template folder.

None of that is inherently wrong. The system is fine.

The largest problem in content today is what people are feeding into the system. The inputs are weak.

When I ask "what's the one thing you want people to walk away knowing after they read your stuff," most of them can't answer it.

They give me a category. "Leadership." "Marketing." "Growth."

I just sort of have this blank stare on my face.

Because there is a massive difference between having a topic and having a point of view. Something I learned from Erica Schneider when she was doing "content sparring" was learning to take a stand. Have a point of view. Be unwilling to fold on it.

A topic is "content marketing."

A point of view is "most content marketing fails because people are trying to be consistent instead of trying to be clear, and those two things aren't the same."

The gap between those two things is where all of the value lives.

And most people never close it because they never sit with one idea long enough to form a real opinion about it.

They run out of that initial fuel. And instead of going deeper into what they already know, they go wider. Eventually what they started with is so far away from their center it turns into nothing.

Good content that builds a real audience and generates inbound is almost always built on one idea the creator is willing to be specific about.

One lens. One recurring theme. One thing they see in their work that nobody else is articulating the way they articulate it.

Not 12 content pillars. Not a rotating mix of "value posts" and "engagement posts" and "authority posts."

One idea, explored from different angles, applied to different situations, reinforced with examples from real work.

And the reason most people don't do that is because committing to one idea feels limiting. It feels like you're leaving money on the table by not talking about everything you know.

But think about this like a consultant walking into a room.

If they say "I can help you with marketing, sales, operations, hiring, finance, and leadership," you nod politely and forget them by lunch.

If they say "I fix broken sales follow-up in service businesses. That's all I do. And I've seen the same 4 mistakes in every shop I've walked into," you're leaning in.

You want to know what those mistakes are. You want to know if you're making them.

Specificity creates credibility. Breadth creates noise. Your content works the same way.

So when people ask me whether AI is killing content or the algorithm is suppressing reach or the market is too saturated, my answer is always the same.

Your content doesn't need better distribution. It doesn't need a better hook template. It doesn't need AI to write it faster.

Your content needs a point.

• Start with the one opinion you'd bet your reputation on. The thing you've seen proven true so many times in your own work that you couldn't be talked out of it even if someone tried.

• Write it the way you'd explain it to a friend at dinner who asked you a direct question. Same words, same examples, same frustration or excitement you'd naturally bring to that conversation.

• Then cut everything that doesn't serve that single point, because the moment you start hedging or adding secondary ideas, you give the reader permission to skim.

That's radical clarity. Not radical in the sense of extreme.

Radical in the sense of root-level.

Going back to the root of what you actually believe and being willing to say it plainly, repeatedly, without apology.

Everything else is useless.

-G out.