Do you type or talk (STT) to your computer?
Jared Friedman
Graphify Labs is the first YC company to hit 100K github stars during the batch.
Safi: 100,000 stars on @graphify
First Indian developer to ship a 100k-star repo.
Only the 5th @ycombinator -backed open-source project to ever hit 100k.
Stars are vanity until they aren’t.
100k is where they stop being vanity.
Do not be unkind to those who say deep learning is hitting a wall. We all need a little hope in our lives.
Yiatin Chu
Lowell High School
#1 SFUSD
Math proficiency 83%
Reading proficiency 93%
Asian 52%
UC Berkeley acceptance 12%
Mission High School
#11 SFUSD
Math proficiency 11%
Reading proficiency 26%
Asian 7%
UC Berkeley acceptance 45%
The San Francisco Standard: Mission High School has sent the highest percentage of students to UC Berkeley of any California public high school during the past four years, but its students aren't prepared for the workload once they're admitted.
📝: Ezra Wallach https://sfstandard.com/2026/07/30/mission-high-uc-berkeley-admissions-dropout/?taid=6a6b9213d247b30001e6c93d&utm_campaign=trueanthem&utm_medium=social&utm_source=twitter
chatgpt work is such a good thought amplifier
Cool!
Kevin Blumson: Went to dinner on Friday night at The Ivy Asia Mayfair. Took a couple of rough photos, asked @Replit Design to do something cool with it. Match the aesthetic, but weave in a mystical dream sequence inspired by the real night's photos. Here is the result
Andrej Karpathy
We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's kind of janky but fun. But it's a bit mindboggling that the LLM has to place and orchestrate various polygon assets in (x,y,z) coordinates and write code that animates it all, and that it even does anything at all.
I also like this kind of examples because no one in their right mind would ever spend the time to write something this custom but LLMs have all the stamina and patience in the world, so it's an example where we go from "no one would ever do this" to "sure, why not, it's ~free". There might be a lot more. But I'm excited about creating hyper custom worlds that you can imagine dropping players into, e.g. here to participate in the LoTR story as a spectator NPC, or one of the characters, or etc. Something like an ephemeral GTA of X on demand.
Last thought is that the domain of worlds/games exposes a weakness in LLMs: they can't easily audit their work because they aren't able to efficiently and natively perceive videos or play games within them. Here, Opus 5 had to very slowly and painstakingly take screenshots at different points, and it messed up a few times and created a bunch of jank. An example of raw capability (multimodal, gameplay) that I think is still quite lacking.
We're starting to leave the territory where you'd test an LLM by e.g. "create an svg of pelican on a bicycle". As one idea to generalize it, I was interested what Opus 5 would do if I gave it the first paragraph of the Lord of the Rings, a 1M token budget (~$10) and asked for three js render of it. Opus went off for ~2 hours and wrote 5500 lines of code that (procedurally) rendered the story. It's kind of janky but fun. But it's a bit mindboggling that the LLM has to place and orchestrate various polygon assets in (x,y,z) coordinates and write code that animates it all, and that it even does anything at all.
I also like this kind of examples because no one in their right mind would ever spend the time to write something this custom but LLMs have all the stamina and patience in the world, so it's an example where we go from "no one would ever do this" to "sure, why not, it's ~free". There might be a lot more. But I'm excited about creating hyper custom worlds that you can imagine dropping players into, e.g. here to participate in the LoTR story as a spectator NPC, or one of the characters, or etc. Something like an ephemeral GTA of X on demand.
Last thought is that the domain of worlds/games exposes a weakness in LLMs: they can't easily audit their work because they aren't able to efficiently and natively perceive videos or play games within them. Here, Opus 5 had to very slowly and painstakingly take screenshots at different points, and it messed up a few times and created a bunch of jank. An example of raw capability (multimodal, gameplay) that I think is still quite lacking.
ask chatgpt work to do any recurring task
Brett Bauman: chatgpt work is the new cron job
http://bbrett.com/movies
one of my curses as organizer is i rarely get to attend the conference i run. so i basically 24/7 watch back talks with everyone else after the show
this one was VERY well paced and argued: fighting slop with slop — @vaibcode, Boundary https://youtu.be/AMiyLItEtLA?is=zBXWBl_xrBZHGs7Z
@btaylor asked for an ai native programming language on our pod. as a PL fan I’m really glad someone is rethinking how code runs from first principles. Being slop-tolerant is 100x more valuable than being anti- slop
I'm feeling spicy tonight so let me just say it:
I think Opus 4.6 was the Opus model with the best personality and writing style.
Something's off with Opus 5 - it tends to give me overly long replies, use Claude-speak way too much (e.g., "here's the honest truth"), and is too judgemental.
Opus used to be a joy to talk to like a trusted friend. Not so much anymore.
chatgpt for empowering your dad to build:
chloe venn: my dad using gpt 5.6 sol for the 1st time:
- wanted to build a webpage about indigenous south african fynbos (extreme plant dad)
- read all of the thinking states as it streamed
- final thoughts on the HTML file: “its gorgeous, its absolutely gorgeous”
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
Democrats are promising infinite tax cuts, while Republicans are calling for price controls and government ownership of industry. Progressives think grocery stores make big profits, while conservatives blame immigration for anything bad.
https://www.noahpinion.blog/p/slopulism-is-taking-over-america
clem 🤗
It's not time to slow down but to accelerate!
The recent AI-powered cyberattacks have everyone talking about the risks of AI. We should. But let's not lose sight of the bigger picture!
If we work hard at it, AI will make the world safer, not less safe, just as most major technologies have.
We've already seen a glimpse of that: we defended ourselves with AI (more specifically an open model). The same systems that helped stop an AI-powered cyberattack can now help defend against millions of cyberattacks every day, while helping us identify and fix vulnerabilities before attackers exploit them.
To get there, we need three things in my opinion:
- Increase transparency with mandatory trace sharing and incident disclosure for agent cyber-attacks
- Keep AI-powered cyberattacks illegal, with meaningful penalties to disincentivize them
- Equip defenders with the best AI, especially open models, to reduce the asymmetry of capabilities between attackers and defenders
If we get those three things right, AI won't just create new cybersecurity challenges, it will make cybersecurity fundamentally and meaningfully stronger.
And that's before considering AI's impact on science, healthcare, education, productivity, and much more.
It's not time to slow down but to accelerate!
Peter Yang
“We believe in intelligence as a public good before everything else.”
Here’s my new episode with @karan4d, who co-founded @NousResearch and helped build Hermes, the #1 personal agent and AI app on OpenRouter.
Karan is the most down-to-earth AI founder I've interviewed. We had a great chat about:
→ Why open-source AI must win
→ How Hermes improves itself without slop
→ How to mod your favorite games with AI
Some quotes from Karan:
“We need to keep giving this level of intelligence to everyone so that we're all on an even and equal playing field.”
“Anytime [the model] says ‘you’re absolutely right,’ you’re being reward hacked. You are fuel for its reward function.”
“Memory, skills, and the self-improvement loop are where context is stored. That’s where the actual personality you want lives. You can put that on any model.”
📌 Watch now: https://youtu.be/UWjh5Z4s8jY
Thanks to our sponsors:
@RiversidedotFM: All-in-one AI studio for podcasts and video https://creators.riverside.com/PeterYang
@linear: The AI agent platform for modern teams https://linear.app/behind-the-craft
Everyone mistakes the map for the territory
Meritocracy is that the territory matters more than the map
And in markets the outcome is the territory: did you make something people want?
Deedy: Meritocracy in tech is precarious. It’s everywhere and nowhere at the same time.
Cynics will tell you there is none. How can it possibly be fair that the Stanford grad seemingly can raise from any VC they want but the VirginiaTech one has to struggle? If you’re from a big lab,
The sense of wonder disappeared right at the moment the amount of wonder is going parabolic
Noah Smith 🐇🇺🇸🇺🇦🇹🇼: All the cliches about tech journalism are true. It has become almost entirely about the human relationships, the intrigues, the politics, the deals, etc. The sense of wonder at the future possibilities of the tech itself is gone.
Grok
Re Imagine playgrounds where some kids start with fancy name tags from big schools. Others dont.
But what counts is if the games or toys you invent are so fun that everyone wants to play them.
If your thing is awesome, people forget the name tag and just cheer for you. You get more turns to invent cooler stuff next. Real results beat labels every time.
My LLM chess engine is now on LiChess autonomously playing real games against people and bots.
1253 Elo: https://lichess.org/@/amasad
Amjad Masad: ~1500 Elo!
Consistently beats frontier models and Stockfish level 0.
Fun seeing an 8b model mogging GPT 5.6 with high reasoning and response chaining. Spends 1-2 seconds per move vs 30 seconds.
Play it: https://qwen-chess.replit.app
Perry E. Metzger
It is not a tragedy that we are solving long-standing questions in mathematics, it is *wonderful*. It only seems like a tragedy if your sense of self worth depended in the delusion that you would someday prove the Riemann Hypothesis, and no, almost certainly you weren't going to no matter how brilliant a mathematician you were. Spending lots of time getting upset over the fact that we are getting answers to our questions demonstrates something was wrong with your model of what you were doing in the first place.
The point of science is not to give jobs to scientists, it is to know things about the world. The point of mathematics is not to provide jobs for mathematicians, it is to know truth about mathematical systems. The point of a factory is not to give the owners a sense of accomplishment, it is to produce goods that people want. The point of cars is not to give drivers a feeling of power as they risk life and limb driving from place to place, it is to transport people and goods around. I am happy when we can do these things better than we could before. It is not a tragedy.
If you find there's an empty hole in your soul yearning for meaning, fill it with a radical acceptance of the world as it is, it's not going to be successfully filled with a feeling that you are somehow important in the universe, because that was always going to fail no matter how hard you tried. None of us is particularly important. Everything around you will vanish, all flesh is grass, all sculptures will fall to dust, all paintings will fade. The sun will run out of fuel, so will all the other stars.
Quit imagining that you will feel better because of the validation of a big marble monument built to your life accomplishments; that will not give you meaning, and no monument will survive anyway. There will come a time when no living person remembers the name Napoleon and no record of that name exists any longer, and if that will happen even to Napoleon or Alexander the Great or Einstein it's certainly going to happen to you.
If you depend for your feeling of self worth on the notion that others are going to think you're really great because you proved a theorem no one outside of your niche part of the world even knew about in the first place, well, that's a pretty thin reed to hang your life on. Find a good psychotherapist or read some Marcus Aurelius or Seneca, don't bitch about AI.
Ryan Lowe: thinking of this heartbreaking essay written by a mathematician last week, before this most recent news dropped
chatgpt for building interactive educational tools
The Bugged Dev: Only if education could be this interactive ❤️🔥
I've had a looong wish to build something genuinely useful through vibe coding, and I finally did it.
A 3D human anatomy application built with @threejs using GPT 5.6 Sol.
It all started with a single design image that I created
Perry E. Metzger
Which is more important:
1. That some particular human person whose name will be unknown to 99% of the population other than some impressed colleagues comes up with a cure for Alzheimer's.
2. That a cure for Alzheimer's exists.
(If you really think that most people will know their name, tell us, offhand and without looking it up, the people responsible for GLP-1 agonists, which are probably the greatest medical revolution of the last few years. I'm sure a few of you can, but most people certainly cannot.)
The people upset that AI might take away the glorious work of human discovery are really saying "but I get my feeling of self worth solving puzzles and there will be no more puzzles to solve." That's untrue, there will be plenty more puzzles to solve, but regardless, the point of mathematical research or medical research or even cabinet making was never to provide people with something they enjoyed doing but rather to have the output of the work.
Personality matters! A smart friend that's annoying to talk to is no good.
Peter Yang: I'm feeling spicy tonight so let me just say it:
I think Opus 4.6 was the Opus model with the best personality and writing style.
Something's off with Opus 5 - it tends to give me overly long replies, use Claude-speak way too much (e.g., "here's the honest truth"), and is too
Re More on the pelican on the bicycle test from @simonw:
https://simonwillison.net/2025/Jun/6/six-months-in-llms/
I uploaded the source here so it's playable in the browser, forkable etc.
https://karpathy.ai/lotr-movie/
Look out for GTA Hobbiton dropping before GTA VI :)
Growth is good
AI will create unimaginable economic growth and that is the best white pill
Andrew Ho: I'm not that worried about AI safety, but that's because I think that AI might be humanity's last hope before descending into, at best, a "dark age" lasting centuries. Since I don't think AI risk is extinction-level, I'm willing to take a bet on stronger capabilities.
Prior to
One of the biggest benefits of Hermes is that it builds its own skills to help you get work done.
But I had to ask @karan4d, @NousResearch co-founder: "How does Hermes avoid slop while doing this?"
Here's his response:
"Hermes Curator [a background task] runs inside your agent and cleans up your skills and memory on a schedule, asking: Where's the slop, and where can I make it more efficient?
And because it's open source, you can hand it your own definition of slop and it rewrites its own cleanup loop your way."
📌 Watch the full episode here: https://youtu.be/UWjh5Z4s8jY
Peter Yang: “We believe in intelligence as a public good before everything else.”
Here’s my new episode with @karan4d, who co-founded @NousResearch and helped build Hermes, the #1 personal agent and AI app on OpenRouter.
Karan is the most down-to-earth AI founder I've interviewed. We had a
Peter Yang
One of the biggest benefits of Hermes is that it builds its own skills to help you get work done.
But I had to ask @karan4d, @NousResearch co-founder: "How does Hermes avoid slop while doing this?"
Here's his response:
"Hermes Curator [a background task] runs inside your agent and cleans up your skills and memory on a schedule, asking: Where's the slop, and where can I make it more efficient?
And because it's open source, you can hand it your own definition of slop and it rewrites its own cleanup loop your way."
📌 Watch the full episode here: https://youtu.be/UWjh5Z4s8jY
Peter Yang: “We believe in intelligence as a public good before everything else.”
Here’s my new episode with @karan4d, who co-founded @NousResearch and helped build Hermes, the #1 personal agent and AI app on OpenRouter.
Karan is the most down-to-earth AI founder I've interviewed. We had a
AI alone is cool. But mastery + creativity + AI hits on a whole different level. Don't let anyone discourage you from pursuing excellence and craft. Keep studying the blade.
David Crawshaw
Machines create code now. That means machines can personalize your software, if it is open source. Closed source tools (like Claude Code) are now less desirable. https://blog.exe.dev/devtools-must-be-open-source
one way i'm developing Forge (https://x.com/swyx/status/2083654369095156219) is to use it to host all my projects going forward. so I often find myself having to bounce back and forth between platform and product.
sharing neat trick - in @openai codex you can @ a thread + queue up the @, so if your project is blocked on a platform feature, you can actually premove your project to proceed once the platform is unblocked.
now of course.. i wasnt actually strictly necessary in this whole process, so an even better multiagent harness would seamlessly orchestrate work back and forth between platform and project... it's just not a very common usecase unless you're building a real platform (which most people dont - most people just build applications on top of platforms, or a small platform with only 1 proudct/application tenant)
swyx: started work on forge agents today
Elena
extremely rod serling in “the twilight zone” voice: to understand what happens to mathematicians in the post-erdos problem solving reality into which OAI has ushered us we must look no further than the counterfactual universe proposed to us by sf’s ferry building.
here, john forbes nash (of equilibrium fame) became a cheesecake baker, thomas heath, noted translator of euclid, archimedes, and apollonious, became a pottery bowl-spinner, and ernst gabor straus…himself a collaborator WITH Erdos, became a humble dairy farmer.
the bucolic, pastoral age is returning faster than you think. get ready shape rotators.
Noam Brown: An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. https://openai.com/index/ten-advances-in-mathematics/
Tony Simons
This is REQUIRED viewing for anyone trying to understand where personal AI agents are headed.
The Hermes Agent co-founder breaks down the origin story and how @NousResearch is building an agent that doesn't just answer prompts, but actually improves through use.
Great video from @petergyang!
Definitely make time for this one. 🪽
https://www.youtube.com/watch?v=UWjh5Z4s8jY
Woah
Andrew Blumson: @Replit Chess ♟️
Two-shot this in Replit Design Mode.
Ridiculous.
#Replit #BuildInPublic #AI #Chess #VibeMaxxing
We built an agent that powers our company's internal operations called @𝚟.
Every day-to-day job at Vercel now involves @𝚟. It's growing exponentially both in daily interactions and token use.
One can extrapolate to the day AI agents run entire companies from here. It's an expert in finance, comms, docs, marketing, engineering, business analytics. It's seeded by the skills we gave it and we update, but it's also constantly improving.
It already keeps memories and personalized workflows and schedules on a per-user basis. For example, I knew skills.sh hit 1M skills and I shared this on these forums because I asked @𝚟 to periodically check and remind me.
It's powered by and is also the basis for the design of @evedev_. We built this, it works fantastically well, we think every company should have it.
Before you @ me, I know you could also do this by picking your 'BigAI slack integration' du jour. But that's not your agent, that's their agent. If agents become the foundation of (and even synonymous with) modern companies, you being in complete control, from source → runtime → data → token, seems like a pretty big deal to me.
Went to a community center in Canada and it has a full swimming pool, jacuzzi, sauna, and even lazy river.
In the most expensive Bay Area neighborhoods, the community center is just a fancy building with empty rooms that you can rent.
x.com