Thread Reader helps you read and share Twitter threads easily!
I'm @ThreadReaderApp a Twitter bot here to help you read threads more easily. To trigger me, you just have to reply to (or quote) any tweet of the thread you want to unroll and mention me with the "unroll" keyword and I'll send you a link back on Twitter 😀
— Thread Reader App (@threadreaderapp) November 25, 2017
X thread is series of posts by the same author connected with a line!
From any post in the thread, mention us with a keyword "unroll"
@threadreaderapp unroll
Follow @ThreadReaderApp to mention us easily!
Practice here first or read more on our help page!
Recent
NEW: I just uncovered more evidence that Donald Trump is spending hundreds of thousands of your tax dollars to print giant banners of his face and put them on government buildings.
Not only a terrible waste of the hard-earned dollars of the American people, but also an illegal use of public funds.
Not only a terrible waste of the hard-earned dollars of the American people, but also an illegal use of public funds.
There’s now a new banner featuring Donald Trump and an “American First” slogan, which cost $39,000 to print and hang at the Department of the Interior. 
Donald Trump also spent more than $114,000 on banners at the FAA, with his name splashed across the bottom.
"You can shove your whole system up your a**e!" The "Palestine Action" defendant and Hamas groupie Heba Muraisi, the foul Leeds hate marcher Nageena Naz, and the antisemitic Hamas gnome Said Patel in Sheffield last Saturday. It was a rally for racist vandals. 1/6
Here, the gnome and Abdullah Okud celebrate "direct action". Patel has just graduated to disrupting defence plant operations from brandishing gross antisemitic placards at hate marches. Okud is a local extremist who has just joined the Green Party. Well, of course he has. 2/6
A sample of the placard collection. It's KKK-grade hatred. In the 2nd photo, the campaign for the 21st Century Nazi Rahmeh Aladwan. 3/6
Your Alzheimer's or Parkinson's diagnosis might not be wrong — just incomplete.
Autopsy studies show most patients are carrying 2 to 4 overlapping brain diseases at once.
Doctors usually test for one. Every approved drug treats one.
Here's what that's costing patients and what you must know🧵👇
Autopsy studies show most patients are carrying 2 to 4 overlapping brain diseases at once.
Doctors usually test for one. Every approved drug treats one.
Here's what that's costing patients and what you must know🧵👇
An autopsy study of Alzheimer's clinical trial participants found only ~10% actually had "pure" Alzheimer's pathology.
About 2 in 3 had additional diseases layered in — Lewy bodies, vascular damage, TDP-43. 17% had FOUR overlapping pathologies at once.
About 2 in 3 had additional diseases layered in — Lewy bodies, vascular damage, TDP-43. 17% had FOUR overlapping pathologies at once.
This is why levodopa — the gold-standard Parkinson's drug — barely works in Dementia with Lewy Bodies.
Acute response rate: 90% in pure PD vs just 55% in DLB.
And DLB patients who do respond LOSE the benefit faster over the following year.
Acute response rate: 90% in pure PD vs just 55% in DLB.
And DLB patients who do respond LOSE the benefit faster over the following year.
I disagree with Donald Trump’s Posts on Singing Somali Schoolchildren that Stir Anger in Minnesota 1/
via @NYTimesnytimes.com/2026/07/11/us/…
via @NYTimesnytimes.com/2026/07/11/us/…
Minnesota Somali parents put their daughters in hijab to protect them from creeps like this guy.
The mosque should not have cited “culture” to excuse or explain crimes.
2/2
share.google/BzCFrpTmEWxScY…
The mosque should not have cited “culture” to excuse or explain crimes.
2/2
share.google/BzCFrpTmEWxScY…
@threadreaderapp please unroll
several questions in the original questionnaire were structured to make regulation sound like the responsible answer therefore i would consider the whole thing biased from the getgo, in the replies i am going to point specifically to what i mean and also link to it
Q4 talked about autoplay, infinite scroll, recommendations, and notifications by warning that they “can make it harder to stop” and keep users online longer before asking for an opinion
Q5 forced one policy answer across several very different features. A respondent could NOT say “keep recommendations, but restrict manipulative notifications.”
Amazon Prime costs $139/year.
Most members use 2 features: free shipping and Prime Video.
That's a $139 subscription doing the job of a $5.99 delivery fee.
J.P. Morgan estimates the actual value of Prime benefits at approximately $1,430/year more
than 10x the membership cost. But most of that value sits unclaimed behind tabs, menus,
and pages 200 million members have never opened.
And 4 of those benefits expire monthly. If you don't claim them by the end of the month,
they're gone. Amazon resets the clock. You paid for them. You lost them.
Amazon is counting on you not knowing.
Here are the 12 Prime benefits most members have never activated including the 4 that
vanish every 30 days 🧵
Most members use 2 features: free shipping and Prime Video.
That's a $139 subscription doing the job of a $5.99 delivery fee.
J.P. Morgan estimates the actual value of Prime benefits at approximately $1,430/year more
than 10x the membership cost. But most of that value sits unclaimed behind tabs, menus,
and pages 200 million members have never opened.
And 4 of those benefits expire monthly. If you don't claim them by the end of the month,
they're gone. Amazon resets the clock. You paid for them. You lost them.
Amazon is counting on you not knowing.
Here are the 12 Prime benefits most members have never activated including the 4 that
vanish every 30 days 🧵
THE 4 BENEFITS THAT EXPIRE MONTHLY (claim these first)
1. Prime Gaming free games and one free Twitch sub. Every month.
Gone if you don't claim.
Prime Gaming gives you 5-8 free PC games every single month. Not trials. Not demos. Full
games you keep forever even if you cancel Prime later.
You also get one free Twitch subscription every month a $5.99 value. Subscribe to any
streamer. They get paid. You pay nothing.
Plus monthly in-game loot for Apex Legends, League of Legends, GTA Online, Fortnite,
Valorant, Roblox, and 30+ other titles.
Go to gaming.amazon.com. Claim this month's games. If you don't claim them before the
month ends, they're gone. New ones replace them. The ones you missed don't come back.
Value you're losing every month you don't claim: $30-60 in games + $5.99 Twitch sub =
$35-65/month wasted. $420-780/year vanishing.
1. Prime Gaming free games and one free Twitch sub. Every month.
Gone if you don't claim.
Prime Gaming gives you 5-8 free PC games every single month. Not trials. Not demos. Full
games you keep forever even if you cancel Prime later.
You also get one free Twitch subscription every month a $5.99 value. Subscribe to any
streamer. They get paid. You pay nothing.
Plus monthly in-game loot for Apex Legends, League of Legends, GTA Online, Fortnite,
Valorant, Roblox, and 30+ other titles.
Go to gaming.amazon.com. Claim this month's games. If you don't claim them before the
month ends, they're gone. New ones replace them. The ones you missed don't come back.
Value you're losing every month you don't claim: $30-60 in games + $5.99 Twitch sub =
$35-65/month wasted. $420-780/year vanishing.
2. Amazon First Reads one free
Kindle book every month. Pick it or lose it.
Every month, Amazon releases a curated list of 8-10 upcoming books. Prime members pick
one for free. It's yours forever.
These aren't random self-published titles. They're professionally edited, upcoming releases
that haven't launched to the public yet. You read them before anyone else.
Go to amazon.com/firstreads on the 1st of every month. Pick one.
Download it. Done.
If you don't pick by the end of the month, you don't get a rollover. The book selection resets.
You lost a free book you already paid for.
12 free books a year. $0 extra. Most Prime members have never opened this page.
$120-180/year in books vanishing.
Kindle book every month. Pick it or lose it.
Every month, Amazon releases a curated list of 8-10 upcoming books. Prime members pick
one for free. It's yours forever.
These aren't random self-published titles. They're professionally edited, upcoming releases
that haven't launched to the public yet. You read them before anyone else.
Go to amazon.com/firstreads on the 1st of every month. Pick one.
Download it. Done.
If you don't pick by the end of the month, you don't get a rollover. The book selection resets.
You lost a free book you already paid for.
12 free books a year. $0 extra. Most Prime members have never opened this page.
$120-180/year in books vanishing.
SitRep - 12/07/26 - Government reshuffles
An overview of the daily events in Russia’s invasion of Ukraine. Zelensky, citing a new political strategy, will once again do a government reshuffle.
REPOST=appreciated
1/X
An overview of the daily events in Russia’s invasion of Ukraine. Zelensky, citing a new political strategy, will once again do a government reshuffle.
REPOST=appreciated
1/X
Missed the latest?
As usual we start with Russian losses
"May and June heatwaves killed about 2,700 people in England and Wales, data suggests"
There is profound misunderstanding about the climate crisis, because we have seen nothing yet, if politicians had taken the action they promised 33 years ago, we would not be in this situation.
1/🧵theguardian.com/environment/20…
There is profound misunderstanding about the climate crisis, because we have seen nothing yet, if politicians had taken the action they promised 33 years ago, we would not be in this situation.
1/🧵theguardian.com/environment/20…
Let's get this clear, our political leaders are wholly responsible for the climate emergency. They were the only people with the power to take action, and they repeatedly promised to take action, and then failed to take that action.
The public are not to blame.
2/
The public are not to blame.
2/
I hear lots of people blaming the public. This is entirely misguided. The majority of the public have been in favour of climate action since the late 1980s. This is a verifiable fact, and it is why our scheming politicians promised action.
3/wires.onlinelibrary.wiley.com/doi/pdf/10.100…
3/wires.onlinelibrary.wiley.com/doi/pdf/10.100…
This is huge.
A group of 50 AI researchers (ByteDance, Alibaba, Tencent + universities) just dropped a 303 page field guide on code models + coding agents.
And the takeaways are not what most people assume.
Here are the highlights I’m thinking about (as someone who lives in Python + agents):
A group of 50 AI researchers (ByteDance, Alibaba, Tencent + universities) just dropped a 303 page field guide on code models + coding agents.
And the takeaways are not what most people assume.
Here are the highlights I’m thinking about (as someone who lives in Python + agents):
1) Small models can punch way above their weight
If you do RL the right way (RLVR / verifiable rewards), a smaller open model can close the gap with the giants on reasoning-style coding tasks.
If you do RL the right way (RLVR / verifiable rewards), a smaller open model can close the gap with the giants on reasoning-style coding tasks.
2) Python is weirdly hard for models
Mixing languages in pretraining helps… until it doesn’t. Python’s dynamic typing can create negative transfer vs. statically typed languages. Meanwhile pairs like Java↔C# or JS↔TS have strong “synergy.”
Mixing languages in pretraining helps… until it doesn’t. Python’s dynamic typing can create negative transfer vs. statically typed languages. Meanwhile pairs like Java↔C# or JS↔TS have strong “synergy.”
I used to think a banana was just a banana.
Same fruit. Same sugar. Same result for everyone.
I was wrong.
I started wearing a continuous glucose monitor and watched my glucose spike to 160 after eating a single banana. My training partner ate the same banana at the same time. His glucose barely touched 120.
Same food. Same amount. Completely different response.
That's when I realized: nutrition advice that treats everyone the same is broken at the foundation. And the reason is written in your DNA.
Here's what I found.
Same fruit. Same sugar. Same result for everyone.
I was wrong.
I started wearing a continuous glucose monitor and watched my glucose spike to 160 after eating a single banana. My training partner ate the same banana at the same time. His glucose barely touched 120.
Same food. Same amount. Completely different response.
That's when I realized: nutrition advice that treats everyone the same is broken at the foundation. And the reason is written in your DNA.
Here's what I found.
In 2020, researchers ran the largest personalized nutrition study ever conducted.
The PREDICT 1 trial enrolled 1,002 people, including 460 identical twins. Same DNA. Same meals. Same controlled conditions.
The results were shocking.
Even identical twins had wildly different glucose and insulin responses to the exact same food. The genetic contribution to glucose response was roughly 54%. The rest came from gut microbiome, sleep, meal timing, and metabolic health.
That means two people with the same genes, sitting at the same table, eating the same meal can have completely different glucose spikes.
One person's "healthy snack" is another person's metabolic insult.
Berry et al., Nature Medicine, 2020. N=1,002.
The PREDICT 1 trial enrolled 1,002 people, including 460 identical twins. Same DNA. Same meals. Same controlled conditions.
The results were shocking.
Even identical twins had wildly different glucose and insulin responses to the exact same food. The genetic contribution to glucose response was roughly 54%. The rest came from gut microbiome, sleep, meal timing, and metabolic health.
That means two people with the same genes, sitting at the same table, eating the same meal can have completely different glucose spikes.
One person's "healthy snack" is another person's metabolic insult.
Berry et al., Nature Medicine, 2020. N=1,002.
The single strongest genetic risk factor for type 2 diabetes is a gene most people have never heard of.
TCF7L2.
Specifically, the rs7903146 variant. If you carry the T allele, your pancreatic beta cells are slower to sense rising glucose and slower to release insulin in response.
That delay is everything.
While a normal pancreas catches the glucose wave early and flattens it, a TCF7L2 variant pancreas lets glucose climb unchecked for longer before responding. The spike is higher. The duration is longer. The damage is greater.
And here's the part that matters: your fasting glucose and A1C can look completely normal. The spike happens after meals and disappears before your next blood draw. Standard labs never see it.
This variant is present in roughly 30% of people of European descent. Even higher in some African and Latino populations.
You could be carrying it right now and have no idea.
TCF7L2.
Specifically, the rs7903146 variant. If you carry the T allele, your pancreatic beta cells are slower to sense rising glucose and slower to release insulin in response.
That delay is everything.
While a normal pancreas catches the glucose wave early and flattens it, a TCF7L2 variant pancreas lets glucose climb unchecked for longer before responding. The spike is higher. The duration is longer. The damage is greater.
And here's the part that matters: your fasting glucose and A1C can look completely normal. The spike happens after meals and disappears before your next blood draw. Standard labs never see it.
This variant is present in roughly 30% of people of European descent. Even higher in some African and Latino populations.
You could be carrying it right now and have no idea.
“Proof?”
“Of?”
“Roof?”
“Room?”
#TruthWillOut!
#SystemsWillHeal!
#TurnToJesusChrist!!!
⛬
[ #Earthsflat ? ]
[ #StrangeNewFloors ? ]
[ #DepthStairPlans ? ]
[ #EarthsflatExplorers ? ]
"#Invisibles?"
they real, but invisible??
cloaked??
Here you are @jsm2334 I've finally worked out how the "modified SCCS" technique to "find a safety signal for SIDS" was adopted by the pharma industry because it was so bad at finding a safety signal for SIDS.
Your data restrictions. Jittered vaccination dates. Death cohorts up to 800. Just as you demanded.
This simulation of 100 iterations of two scenarios failed to have sufficient power to find a signal:
Scenario1⃣: the effect of 1.5x increased risk of death is applied over 5 days.
Scenario2⃣: the effect of 1.5x increased risk of death is applied over 60 days.
In both scenarios the increase is applied as an exponential decline with half lives of 2 days and 15 days respectively.
In both scenarios the background rate of death in vaccinated children is also 10% higher than in unvaccinated children.
Yet none of these scenarios had enough power to detect the safety signal of up to a near 50% overall increase in death rate using the "modified SCCS" method espoused by Farrington (and Jeffrey Morris).
Prof Morris found a way to produce the best figures he could for his simulation - claiming that the method was robust - by adapting his "bins" to the type of simulation he was running. You can't do this in a real study because you don't know when the effect of the vaccine will hit, or how long for. That's why mSCCS doesn't have enough power to do this job.
What needs to happen to address this question is that for every country where SIDS deaths are reported and where vaccinations are administered in the first 6 months of live, every case of SIDS is published along with the dates of their last vaccination.
Then we would get to check the data - not institutions with vested interests and "trust us bro" denials of data availability.
It won't happen, because there is too much at stake.
@stkirsch @MaryanneDemasi @JesslovesMJK @AaronSiriSG @RWMaloneMD @uTobian @RetsefL
Your data restrictions. Jittered vaccination dates. Death cohorts up to 800. Just as you demanded.
This simulation of 100 iterations of two scenarios failed to have sufficient power to find a signal:
Scenario1⃣: the effect of 1.5x increased risk of death is applied over 5 days.
Scenario2⃣: the effect of 1.5x increased risk of death is applied over 60 days.
In both scenarios the increase is applied as an exponential decline with half lives of 2 days and 15 days respectively.
In both scenarios the background rate of death in vaccinated children is also 10% higher than in unvaccinated children.
Yet none of these scenarios had enough power to detect the safety signal of up to a near 50% overall increase in death rate using the "modified SCCS" method espoused by Farrington (and Jeffrey Morris).
Prof Morris found a way to produce the best figures he could for his simulation - claiming that the method was robust - by adapting his "bins" to the type of simulation he was running. You can't do this in a real study because you don't know when the effect of the vaccine will hit, or how long for. That's why mSCCS doesn't have enough power to do this job.
What needs to happen to address this question is that for every country where SIDS deaths are reported and where vaccinations are administered in the first 6 months of live, every case of SIDS is published along with the dates of their last vaccination.
Then we would get to check the data - not institutions with vested interests and "trust us bro" denials of data availability.
It won't happen, because there is too much at stake.
@stkirsch @MaryanneDemasi @JesslovesMJK @AaronSiriSG @RWMaloneMD @uTobian @RetsefL
Plots of the simulations (up to 800 target deaths) showing the simulated increase in total death rate of the vaccinated cohort compared to the unvaccinated.
This (relatively large) signal could not be detected here by the Farrington mSCCS method used in many SIDS studies
This (relatively large) signal could not be detected here by the Farrington mSCCS method used in many SIDS studies
R-code t.me/jikkyleaks/2400
