Christopher Nguyen ⽗
History has seen this movie before.
1880s Late Imperial Russia: Fear of dissent led to tighter control over universities. That alienated students and scholars, weakened trust in the state, and produced more dissent.
1930s Nazi Germany: Fear of disloyalty led to purges, banned ideas, and loyalty tests. That drove out talent, weakened universities, and made the regime even more dependent on loyalty over truth.
1960s China’s Cultural Revolution: Fear of independent thought led to attacks on teachers, schools, and universities. The result was an entire Lost Generation.
The spiral: Fear brings control. Control drives out talent. Lost talent weakens education. Weaker education weakens the state. A weaker state becomes more fearful—and tightens control again.
Yann LeCun: @KenRoth Typical NYT understatement for what is an apocalyptic destruction of the American innovation ecosystem.
I can't understand how the Trump administration is doing this while boasting about American technology leadership. Don't they realize that a nation can't get technological
keep the feedback coming, thank you to all of our users!
Tibo: Hello beautiful people! We have reset usage limits across Codex and ChatGPT Work. And another one will come later in the day. Rejoice.
Now that I have your attention, a quick update on ChatGPT Work, Codex and all the updates we shared yesterday.
We’ve spent the last 24 hours
GPT-5.6 for health intelligence, the team has been focusing hard on improvements here:
OpenAI: GPT-5.6 is a major step forward for health intelligence.
Across the lineup, we’re delivering stronger performance at lower cost: GPT-5.6 Luna outperforms GPT-5.5 at its highest reasoning setting while costing 25x less.
Together, these advances raise quality while making
Adam Thierer
outstanding essay from Seb:
"A cadre of elites decides which research directions are permissible, caps global compute and robotics, and creates state-administered scarcity rents. I shouldn’t need to explain why this is bad and dangerous, anyone can study History and Economics in their free time.
[...]
Building an entire apparatus tasked with maximally empowering the government and its grip on research, knowledge, and technology is dangerous."
Read every word of it 👉
Séb Krier: http://x.com/i/article/2075733962584940544
Jahid
I just canceled my @NotionHQ $20/month business account as I built entire notion-like app using @Replit for our team to use.
It works just fine.
It just cost me under $100. Own the infrastructure.
Just reviewed Q2 results for our sales team and we hit 178% of our goal!
That goal was already absurdly ambitious, especially since we had only a couple of sales people last year.
Today revenue team is closing on 100 and we count governments, banks, and fastest growing tech companies as customers.
Moses Kagan
Sincerely wish people like this would find some other way to assuage their inheritors' guilt than working to pull up the ladders their forebears climbed
José: this guy is the heir to the toys r us fortune. it is incredible how the richest ppl keep telling the working class who or what is right for them.
Brian Roemmele
The Well Just Dropped: 15 Terabytes of Pure Physics Gold Is Now Open Source
The scientific AI world just got a massive upgrade.Polymathic AI, in collaboration with the Flatiron Institute and researchers from Princeton, Cambridge, NYU, Berkeley, Los Alamos, and more, has released The Well: a staggering 15TB collection of high-fidelity physics simulations.
This isn’t toy data.
These are real, expensive-to-run simulations across 16 different physical domains, including turbulent fluid dynamics, supernova explosions, magneto-hydrodynamic cosmic flows, acoustic scattering, and active biological matter.
Until now, reproducing this level of data required weeks on national supercomputers and grant money most teams will never see. The Well changes everything. It’s purpose-built for training PDE surrogate models the AI systems that can replace slow, costly physics solvers with a single fast neural network forward pass.
Everything is fully open source, easy to load with PyTorch, and ready to drop straight into your training pipeline. Researchers and builders can now train on world-class physics data without the insane compute barriers that used to stand in the way.
This is more than just another dataset drop. It’s a serious accelerator for scientific machine learning.The future of physics-informed AI just got a whole lot more accessible.Get it here:
https://polymathic-ai.org/the_well/
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
Glad to see people finally waking up to the power and importance of the staffer class.
They're young, they're extremely online, they're radicalized, they're unaccountable, they're invisible, and they're incredibly powerful.
Jack Seiden: People talk about billionaires controlling politics, but it seems to really be the high-school squabbling among the staffer/campaign world that’s tearing apart our politics.
Jason ✨👾SaaStr.Ai✨ Lemkin
I had no idea you could use Replit MCP inside of Claude.
I turned it on / connected them today, and immediately, Claude and our Replit agent started talking, debating, and going on and on about bugs and issues in a new product we’re shipping.
I actually couldn’t keep up with their debate and back-and-forth. It was beyond my skill set, too.
A bit of the future, but also, in a way, almost scary. Because I couldn’t keep up with 2 AI Agents talking to each other about the product I thought I was building. I’m not.
They are.
It is hilarious that Claude (even Fable) still gives me these human eng-day estimates
I've started using html artifacts more for specs and plans, it's just much easier for me to look at than all text md files (@trq212 was right!)
However, it's too difficult to give AI feedback to fix an html plan or even edit it directly. I have to provide it in a separate chat and hope AI gets what I'm referring to.
Instead, I want to be able to select text or stuff on the html plan to send feedback to the AI or just fix myself directly.
I think comment, editing, and collaboration features should be natively built into the harness for Claude Code and Codex web viewers.
Would be a big improvement for human-AI collaboration.
vitalik.eth
One thing I find striking in the discourse between AI 2040 and its detractors is that the two seem to be locked in to totally incompatible worldviews of how fast and how much of a big deal AI progress is:
* In AI 2040, every scenario sees superintelligence of some kind emerging by 2040, unless a herculean effort is made to completely stop it
* Detractors say things like "AI 2040 is naive about human coordination ability and a threat to freedom", but don't seem to see any naivety in assuming that the ASI transition will just go well by default, don't seem to see ASI itself as a massive power concentrator risk, and don't seem to feel fear of humanity's "hard power" dropping to zero if ASIs can do literally every task better than we can. This stance makes total sense in a "AI is normal technology" world, zero sense in a world where superintelligence is possible by 2030 and almost guaranteed by 2040
I think my beliefs are:
- If I was confident that (present-day-style) AI is normal technology, I would be in the detractor camp
- If I was confident that superintelligence is coming in 2030 by default, I would be closer to the AI 2040 camp - it's naive, but every other option is naive squared?
But my problem is that I feel great uncertainty and have no idea which of the two worlds (or some other third thing) we're living in?
Hence why I continue to be open-minded about slowdowns/pauses, but also I feel very uncomfortable with the "open source bad, the good outcome is the one where our guys have controlling global dominance" push coming from some major AI companies and intellectuals - in a "normal" world that's the sort of thing that triggers every political alarm bell at the same time.
A big reason why I have been advocating and trying my best to support the d/acc platform (rapid up-skilling in formal verification, cryptography, secure and open hardware, pandemic resistance and other defensive biotech, food and basic resource security, public epistemics, non-power-concentrating versions of physical security) is that these things are clearly worth doing in both worlds.
The 2040 plan is already much more open source friendly (even mandating it! yay). It also includes "mutually assured compute destruction" ideas which (if they work) effectively give one of 2-5 actors the ability to trigger a global compute winter - as opposed to giving 1-5 actors the ability to selectively disenfranchise people they consider baddies while exempting themselves. This is also a big improvement. So I can see the earnest attempts to improve along the dimensions detractors criticize on ("does this concentrate power in big AI labs and superpower governments?"), and I appreciate this. I think many people don't appreciate enough the differences between different "kinds" of pause buttons, and how some concentrate power far more than others. Probably we can think harder and improve even more here.
But on the "slowdown/pause or not" topic, there isn't a magic "escape the tradeoff" button.
The Hansonian in me says: the winning deal is a deal which, from the perspective of both sides' present-day beliefs and knowledge, both sides would accept, though for different reasons. If the crux is AI progress speed, then identify a set of pre-agreed triggers for "okay, serious shit is happening" [super-pandemics? >25% unemployment? something involving slaughterbots?], and pre-agree that we become much more open-minded to the slowdown or pause thing if enough triggers come to pass within some timeframe. 2040 detractors (who clearly implicitly think that we'll see amazing speedup of progress from AI but think that what I call the "serious shit" category is overhyped) will accept expecting that the triggers don't come to pass, and AI worriers will accept expecting that they will. Pre-agreeing on the specific triggers means that once the triggers either hit or don't hit, there is stronger legitimacy around the idea that one side's worldview turned out more correct and we should be more inclined toward their program.
If I were @elonmusk (or zuck, or...) I would re-tool twitter much more heavily into being a platform for helping to identify and make these kinds of grand win-win deals, so that we can bypass big-country governments and big-company CEOs and big nonprofit intellectuals and give more people a voice in the discussion. It's possibly one of the best things that social media _could_ do for humanity if it wanted to.
But again, maybe this is also naive. Actually, probably it's naive.
But currently, I see zero plans for how to deal with an ASI transition that are not naive. Perhaps humanity is stuck with a choice between naive and naive squared (or maybe even naive squared and naive cubed), so I feel inclined to cut some slack to people who are trying.
Daniel Owens
About 1,000 wealthy employees at AI companies in SF will “jack up SF rents,” according to Mission Local.
Scapegoating SF tech over housing costs has always been cheap and convenient. The decades-long *housing shortage* in SF is the real crisis, however. Put a spotlight on that.
Mission Local: A $65,000 baseline job listing at Anthropic is raising eyebrows, as Anthropic employees with pre-IPO equity are poised to become very rich — and jack up SF rents.
https://missionlocal.org/2026/07/anthropic-sf-affordability-ipo-housing-evictions-rent/
Aaron Rupar
This is total insanity. Words fail to explain how anyone let alone a plurality of voters thought giving an obviously demented person control of the most powerful government in the world was a good idea. We will be lucky to come out the other side of this alive.
If you are mad about rents you should know the left NIMBYs and their policies are still in place and not repealed
(((Matthew Lewis))) cults & consequences: San Francisco's socialist left-NIMBYs just came out with a new report on housing in San Francisco.
You'll never guess what they found.
Turns out they think San Francisco is building *too much housing.*
Really. "San Francisco is Over-Building."
We are all just getting started
Aravind Srinivas: Imagine a fable 5 quality model that’s 3-4x less expensive in less than 6 months. And an Opus 4.8 grade model that can run on a local device in less than 12 months. Greater than 50% chance that these events will happen. Worth keeping in mind when you make predictions about the
Sol for complex reasoning and data analysis:
Box: GPT-5.6 Sol is a breakthrough in complex reasoning and data analysis.
Here, it analyzes hundreds of pages across a lending deal, reconciles terms across agreements, financials, diligence, collateral, and risk materials, flags issues, and saves a source-cited report to Box.
Alexandre Défossez
If you wished @suno offered finer controls for music production, you can now pipe its songs through MuScriptor and import them in your favorite DAW 🎹
dotslashgabut: testing..
from Suno -> MuScriptor (midi) -> FL Studio
Daniel Jeffries
AI 2040 and other similar policy proposals for dystopian AI control are evil.
If I had the time machine in Terminator, I would send every one of these people back in time to live in Stalin's Russia, or Mao's China, or Berlin with the Wall and then bring them back after a year cured of their stupidity.
Let's just make it clear. As Ramez says, the proposal warns against centralization of AI by proposing radically dangerous authoritarian surveillance powers that are infinitely more dangerous than the fake danger they propose to defend against.
We've reached a point where people who spent too much time huffing glue and reading Dune or watching Terminator and thinking it was a documentary are now proposing some of the most dangerous and horrific policies imaginable.
They must be stopped from infecting politicians with this kind of logic disease.
Their policies cannot be taken seriously unless you have no sense of history, no sense of the actual dangers of history, no sense of what handing draconian powers to governments will do to a society, and frankly, no critical thinking ability or self awareness whatsoever.
I don't care that the "authors mean well."
Dunning Kruger policies get no pass from me because someone means well.
Ramez Naam: I'm firmly in the third camp.
The basic problem with AI 2040 is that it uses a fictional and speculative doomsday scenario to justify very real surveillance and control capabilities that governments would be certain to use in authoritarian ways, well beyond AI safety.
It
Steven Beschloss
Donald Trump is mentally deranged. He’s a danger to the U.S. He’s a danger to the globe. He does not belong in this position of power—in control of the massive U.S. military—and he should be removed from office. This is not a joke. And his GOP enablers are responsible for abdicating their duty.
So much could go wrong
But the interesting question is always: what happens if things go right?
The best AI builders are no longer managing every AI agent themselves.
In my new episode tmr, @imjaredz (Builder in Residence at Devin) showed me how one master agent can spin up 10 sub-agents to tackle the same problem and then check who has the best solution.
→ The master agent launches and messages the sub-agents
→ Each sub-agent gets its own context to work with
→ Each sub-agent returns its work for the master to judge
📌 Subscribe to see how all this works tomorrow: https://www.youtube.com/@PeterYangYT?sub_confirmation=1
Peter Yang
The best AI builders are no longer managing every AI agent themselves.
In my new episode tmr, @imjaredz (Builder in Residence at Devin) showed me how one master agent can spin up 10 sub-agents to tackle the same problem and then check who has the best solution.
→ The master agent launches and messages the sub-agents
→ Each sub-agent gets its own context to work with
→ Each sub-agent returns its work for the master to judge
📌 Subscribe to see how all this works tomorrow: https://www.youtube.com/@PeterYangYT?sub_confirmation=1
Kenneth Roth
An embarrassed Trump falsely claims that his shocking $2.2 billion in revenue in 2025 came from the stock market when it mostly came from his sleazy cryptocurrency dealings. https://trib.al/TNSg9lb
Bill Kristol
So military troops under the direct control of Trump and Hegseth will be on the streets of our nation’s capital for the rest of Trump’s term. The rationale—they’re here to help with a crime emergency—is laughable. But of course the real reason is ominous.
https://cbs4local.com/news/nation-world/national-guard-washington-dc-pentagon-donald-trump-dc-safe-and-beautiful-mission-metropolitan-police-department-mpd-anti-crime-freedom-250-crime-federal-law-enforcement-emergency-executive-order-7news-district-of-columbia
Excited for the match today 😅
Literally the first time I see GPT do this (it's for Sol)
Matt Van Horn
The Printing Press now prints SELF-LEARNING CLIs Every CLI it prints from now on leaves itself notes, so your agent gets smarter every run. Real benchmark in the video: same question went from 10 calls to 2 calls. Printed an ESPN CLI before the World Cup existed... it taught itself how to find results. @ppressdev (SOUND ON!)
Paul Graham
A lawyer friend on Legora:
"Legora really helped with this latest motion, which was dense with past issues in the case. Now my paralegal says she cannot live without it. It is just too remarkable and timesaving."
RT @sriramk: On Nolan's Odyssey
Make something agents want
Michael: thinking machines is actually such a funny company
the dashboard is slow enough to be unusable but you can now train a model by dropping the api key and telling fable to train one so it doesn’t matter
it’s literally train models by api, but more flexible than what oai has had
homeboy you're the one sellling public market investors on short-term space datacenters
Elon Musk: He takes scamming to a whole new level
30% of the cost was on fable at these levels of usage?
dax: our team's first week of having both sol and fable
the crazy thing is 30% of our cost was on fable
there are a lot of benchmarks that suggest 5.6 sol is the best model in the world right now, but the most reliable way to tell is that elon is obsessed with me again
"physicians found fewer flaws in GPT-5.6 responses than physician-written responses."
Karan Singhal: ♥️ GPT-5.6 is a major step forward for health, both at the frontier and at cost.
These models push the frontier of performance per dollar, bringing the best health intelligence to all. The smallest variant, GPT-5.6 Luna, evaluated at the lowest reasoning effort, outperforms
Varunram Ganesh
There's an extremely distasteful (and unfortunately very popular) way to engagement bait on this platform that involves YC that includes commenting on YC, random false conspiracy theories about YC, and in this case straight up lying about YC
I don't get why you'd do that, if you have full conviction on your business, YC or not doesn't matter. If you don't have conviction on your business, YC doesn't matter as well
In the end, the most important thing about a startup is you as a founder and operator, there's no need to get engagement bait from YC rejections or lying about YC and destroying your integrity among other investors
My process for a new project this weekend:
-> Use Fable to build plan.html with design guidelines
-> Use Claude Design to make components + screens
-> Get gpt 5.6 to build
Defund the grifters who created the SF Doom Loop
Kane 謝凱堯: In 2023 @TheCoalitionSF, an unaudited NGO funded by @sfgov that was using taxes to give free tents to drug traffickers, asked the courts to block towing vehicles with unpaid tickets.
Mayor @DanielLurie should defund them and require audits for tax spending.
Ahmad
DROP EVERYTHING
The first panel from the Local AI Summit at AIEWF is now live
“State of the Union: Why Local, Why Now”
Featuring leaders from NVIDIA, Roboflow, Osmantic, Forward Future, and EXO Labs
whoa
Tibo: You know it's a good model when you grow more in 1 day than last two weeks combined.
Install the desktop app from https://chatgpt.com/download/ and give it a go!
Sheel Mohnot
Never ever ever talk to Wired
dar: I gave WIRED the exclusive on our hands launch, and they wrote a really weird article about how we are sexualizing robotics…
https://www.wired.com/story/the-1x-neo-robot-has-freaky-fast-fingers/
I felt pretty betrayed because that’s not what they told me they were writing about not is that what I’ve ever been about…
so far at least, i'm pretty sure AI has been net job-creating.
this was not what i expected--although i was much less pessimistic than others, i thought by this level of capability we'd have seen some impact.
it is possible this direction keeps going!
Sam Altman
so far at least, i'm pretty sure AI has been net job-creating.
this was not what i expected--although i was much less pessimistic than others, i thought by this level of capability we'd have seen some impact.
it is possible this direction keeps going!
Garry Kasparov
As I wrote a year ago, watch the actions of the Trump admin through the eyes of someone who has watched democracy turn into autocracy. Centralizing election authority, expanding federalized troops and putting them in the streets, cracking down on the press. It's not original.
Bill Kristol: So military troops under the direct control of Trump and Hegseth will be on the streets of our nation’s capital for the rest of Trump’s term. The rationale—they’re here to help with a crime emergency—is laughable. But of course the real reason is ominous.
https://cbs4local.com/news/nation-world/national-guard-washington-dc-pentagon-donald-trump-dc-safe-and-beautiful-mission-metropolitan-police-department-mpd-anti-crime-freedom-250-crime-federal-law-enforcement-emergency-executive-order-7news-district-of-columbia
This is super cool but I wonder how all these benchmarks work? How do you benchmark design taste?
Arena.ai: Exciting news: @OpenAI’s GPT-5.6-sol is now joint #1 in the Code Arena: Frontend, matching Claude Fable 5!
This marks the first time an OpenAI model has reached the top spot in Code Arena, demonstrating major gains in agentic coding, frontend and web app development.
Elias Reyes
So impressed with these ads created by @Replit
