YC Partners are cooking for open source
We are going to have 1000x more hyper usable open source. This is the golden age of personal software, infinitely customizable, and it will be way better than corpo software
Ankit Gupta: 0.7.0 is now out, featuring:
- Several major performance upgrades. Turns out coding agents are really good at perf optimization. The app should be buttery smooth now.
- Security updates from the community (thank you!)
- in-email search
https://github.com/ankitvgupta/mail-app
For the haters who say you can’t speed up with agentic coding I will just prove you wrong
And if you still don’t believe, Have fun coding at 1x speed✌️
The rest of us are going to go fast
We are fighting to stop this
Join @garryslist if you want to help us
Rob Henderson: "San Fran has become both the richest city in history and a cautionary tale..the very people a city most needs to thrive,the highly educated and culturally refined, are most likely to vote for the sort of politics which drives away families and businesses" https://www.edwest.co.uk/p/the-city-of-luxury-beliefs
Chengpeng
This isn’t an edge case. From anonymized U.S. ChatGPT data, we are seeing:
• ~2M weekly messages on health insurance
• ~600K weekly messages from people living in “hospital deserts” (30 min drive to nearest hospital)
• 7 out of 10 msgs happen outside clinic hours
Simon Smith: I’ve been critical of OpenAI lately, but for the past three weeks my family has been dealing with a health issue with my dad, and a ChatGPT shared project with live document syncing has been essential to organizing and understanding everything happening.
Me, my four siblings, my
Codex app server makes it easy to build your own agentic apps:
am.will: The Codex app server was such a brilliant stroke of foresight that really doesn't get enough love
Not only are you allowed to use your chatgpt account with any harness, but you can build your own apps directly on top of theirs.
They just make building on and with codex such a
Codex app server makes it easy to build your own agentic apps:
am.will: The Codex app server was such a brilliant stroke of foresight that really doesn't get enough love
Not only are you allowed to use your chatgpt account with any harness, but you can build your own apps directly on top of theirs.
They just make building on and with codex such a
14 security bug fixes just landed for GStack, half of which were community PR's.
Sneak peek: Your OpenClaw is about to be able to control a browser on your computer using GStack Browser. 👀
Rohan Paul
Incredible possibilities for on-device small models.
Here @adrgrondin is running Google’s Gemma 4 E2B on iPhone 17 Pro.
~40tk/s with MLX optimized for Apple Silicon
SOTA coding & math on mobile with 128K context. Fully offline with thinking mode.
First time putting live shrimp in sweet alcohol and then eating it
Peter Yang
"We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team built the beautiful Codex app
→ How they ship without traditional specs and roadmaps
Some quotes from Alex and Romain:
"The fewer people you need in a room to do anything, the more pure every decision is."
"Our designers write more code now than was written by an engineer six months ago."
"I'm much less likely to read someone's resume than their ideas and what they've built."
Thanks to our sponsors:
@meetgranola: The best AI meeting notes app I've ever used http://granola.ai/peter
@linear: The AI agent platform for modern teams http://linear.app/behind-the-craft
📌 Watch now: https://youtu.be/9qXc-THAvc0
Re Also available on:
Spotify: https://open.spotify.com/episode/4CT2R6lK43jNnsroxyV8P8?si=z0ic0SD2Tmqd7qTQMK3G8A
Apple: https://podcasts.apple.com/us/podcast/behind-the-craft/id1736359687?ign-itscg=30200&ign-itsct=podtail_podcasts
Newsletter: https://creatoreconomy.so/p/how-openai-codex-team-builds-with-codex-alex-romain
ericosiu
Every org needs a world intelligence to operate at a 10x level. It might seem daunting, but it’s worth the investment.
We’re already starting to see a multiplier at my company.
Here’s how to get started:
ericosiu: http://x.com/i/article/2040532479988355072
Jing Wang
最近在对比使用三大harness skill：
Superpower: https://github.com/obra/superpowers
Compound Engineering: https://github.com/EveryInc/compound-engineering-plugin
Gstack: https://github.com/garrytan/gstack
每个人有不同的喜好和偏向啊，我的感受如下：
首先，这三个skill本身是在解决同样一个问题，AI Agent在执行任务，尤其是写代码的时候容易跳过规划，质量不稳定，没有一个固定的workflow来限制它。但这三个skill又有不同的思维模型和侧重点：
Superpower：强调流程纪律，每一轮会从spec -> plan -> 拆解任务 -> 执行任务 -> 测试和review -> 提交。优点是流程控制的很好，缺点是小任务时间按这个走很长，需要授权的次数多，消耗token。
CE：强调知识复利，重点是给 AI 装记忆，错过的东西下次不再犯。/ce:compound 会把历史错误给沉淀下来。另外/ce:ideate 可以针对性提出改善建议，并会对优先级排序，筛选掉不重要的improvements。但是我实际的感受是，这些改善点优点不痛不痒。
gstack：最大的亮点是YC的知识和评判标准，对于产品和idea的打磨，让AI对你进行灵魂拷问。更新的版本也支持知识沉淀，未来使用。
我的感觉是，如果是大工程用gstack来进行产品和idea打磨，superpower来执行。
#skill #claudecode
Shanu Mathew
Wild “Our designers write more code now than was written by an engineer six months ago”
Peter Yang: "We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team
Guilty as charged
David Tran: @garrytan Shipping at 2:35a on a Sunday! Love it
Re Do you want this for yourself?
Tell your Claude Code to "Install GStack" and it'll just work
Start with /office-hours and then /autoplan on your next plan mode, and watch how much better your software ends up being
https://github.com/garrytan/gstack
This seems messed up actually - when do the boundaries stop moving?
Anthropic only allows subscriptions with a real human pressing enter? You're going to have to verify with FaceID?
Peter Steinberger 🦞: Anthropic now blocks first-party harness use too 👀
claude -p --append-system-prompt 'A personal assistant running inside OpenClaw.' 'is clawd here?'
→ 400 Third-party apps now draw from your extra usage, not your plan limits.
So yeah: bring your own coin 🪙🦞
Created a flag on the moon based with hyper-realistic physics with @v0: http://v0-moon-flag.vercel.app.
The flag is simulated as a grid of particles connected by springs (cloth simulation). It's influenced by gravity, wind, air resistance.
The initial approach v0 one-shotted was quite simplistic, but giving it references to more sophisticated approaches in C++ and Rust quickly improved it. It was mind-blowing it to see it deeply understand the physics and port new ideas across languages.
It simulated a real sun for lighting, and procedurally generated moon-like terrain with craters. I found a sick free texture on Reddit that I uploaded and it layered on top.
Another epic moment was watching it improve performance. It seamlessly moved the physics simulation to a Web Worker. (Kinda cool to see given we just improved support for them in Turbopack 😉)
Created a flag on the moon based with hyper-realistic physics with @v0: http://v0-moon-flag.vercel.app.
The flag is simulated as a grid of particles connected by springs (cloth simulation). It's influenced by gravity, wind, air resistance.
The initial approach v0 one-shotted was quite simplistic, but giving it references to more sophisticated approaches in C++ and Rust quickly improved it. It was mind-blowing it to see it deeply understand the physics and port new ideas across languages.
It simulated a real sun for lighting, and procedurally generated moon-like terrain with craters. I found a sick free texture on Reddit that I uploaded and it layered on top.
Another epic moment was watching it improve performance. It seamlessly moved the physics simulation to a Web Worker. (Kinda cool to see given we just improved support for them in Turbopack 😉)
Alex Finn
If you used a Claude subscription with OpenClaw, read this:
Unfortunately all other AI models out there absolutely suck with OpenClaw compared to Opus
It's just a fact and anyone denying this is delusional
So here is my new recommended OpenClaw setup:
Pay for the Opus API and use it as your orchestrator
Then use other models as the execution layer
If you do this correctly, yes your costs will go up, but not by as much as you think
I use my ChatGPT subscription as the coding execution. GPT 5.4 is excellent at coding. When The Opus orchestrator gives a coding task to the ChatGPT subagent, it always performs really well
If you are on the Pro plan, you should have enough usage to have ChatGPT be the execution layer for every task. But if youre on the $20 a month plan, youre going to need other subscriptions to handle other tasks
GLM 5.1 and Qwen are excellent. I'd get a cheap sub through them and have them handle all other tasks given to them from the orchestrator
The best setup tho if you have the hardware is Opus API for orchestrator, ChatGPT for coding, then local Gemma 4 and local Qwen handling everything else.
Right now have Gemma running on my DGX Spark and Qwen 3.5 on my Mac Studio. They handle all other execution from my Opus API orchestrator
Unfortunately all options above will cost more than the $200 a month subscription. It just is what it is. But if you optimize correctly it wont cost much more, and you'll still get frontier performance.
OpenClaw is the most powerful piece of software ever released. $200 a month ($2,400 a year) was a steal for a digital employee. Honestly anything under $50,000 a year is a no brainer if you run a serious business.
The situation isn't great but you also need to face reality: Claude Opus 4.6 is the best model for OpenClaw. If you use any other model, your productivity will suffer
Business is a battlefield and I refuse to fall behind, so despite me not being happy with the Anthropic decision the setup above is what I'm going with
Virtue signaling might get me brownie points on the internet, but it won't increase my productivity
Many such cases
Juan David Gomez 🚀: @garrytan I just want to thank you for #gstack. I no ZERO about coding and with the skills you developed and kindly shared I have been able to create some apps for myself. It's an amazing feeling. For a couple or years I always dreamed about coding. Today that's a reality!
We are all getting faster and better at making things together
Om: Is this what coding really look like @garrytan ?
btw huge thanks to Gstack, helped a lot
Justin Gray
here is @embirico helping is think through the future of teams:
"all these roles are blurring together. a designer can do more engineering. an engineer can do more design. a PM can do more building." "labels are losing meaning"
"if you were a good PM but not that good at engineering, maybe you should become an engineering manager wth a coding agent"
"it comes down to interest and agency. do what you're most interested in"
"every problem needs a human accountable for that problem area. but that doesn't have to be a PM"
Peter Yang: "We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team
Starting in September, Boston will become the first major city in the US with an AI literacy course for public high schools 🏫
Not sure about you but…
I had to take a typing class.
And learn how to research on the internet.
And how to save files to a floppy disk.
This feels like an evolution of that tech literacy, but with higher stakes.
According to Boston’s mayor, the curriculum is “really grounded in ethics and grounded in understanding how to maintain and develop creativity, leadership” and is meant to “enhance the learning that's happening, not replace or substitute for it.”
It is not a graduation requirement, and they don’t yet know what the exact course structure will be or what grades will take it.
My main worry (and it’s a big one) is that leaders and teachers will be too slow to update the course every year citywide.
Because while I believe some training is better than none, I worry about teaching outdated AI practices or conventions and the negative effect that could have on a person’s technical decisions.
Starting in September, Boston will become the first major city in the US with an AI literacy course for public high schools 🏫
Not sure about you but…
I had to take a typing class.
And learn how to research on the internet.
And how to save files to a floppy disk.
This feels like an evolution of that tech literacy, but with higher stakes.
According to Boston’s mayor, the curriculum is “really grounded in ethics and grounded in understanding how to maintain and develop creativity, leadership” and is meant to “enhance the learning that's happening, not replace or substitute for it.”
It is not a graduation requirement, and they don’t yet know what the exact course structure will be or what grades will take it.
My main worry (and it’s a big one) is that leaders and teachers will be too slow to update the course every year citywide.
Because while I believe some training is better than none, I worry about teaching outdated AI practices or conventions and the negative effect that could have on a person’s technical decisions.
Happy Easter. He is risen.
Mike Solana
“california high-speed rail is creating good-paying jobs” very cool but is it creating high-speed rail
CA High-Speed Rail 🚄💨: California High-Speed Rail is creating good-paying jobs for hardworking members of the trades.
🛠️16,000+ jobs created
🏗️1,600+ daily-workers dispatched
These jobs mean steady, mortgage-paying careers for thousands of Californian families. #BuildHSR
I am absolutely convinced this is the most awesome moment in history to have high agency
*You can just do things*
Kpaxs:
What are my current valuation predictions for OpenAI and Anthropic?
OpenAI was valued at $28B in April 2023. Its latest funding round puts it at $852B.
Anthropic was valued at $4B in early 2023. Today, it’s worth $380B.
In my opinion - and this is not investment advice - these companies could easily reach $2T in valuation in the next two years.
I tweeted in November 2023 that we would see companies start to raise tens of billions of dollars. At the time, I believe the largest round that we had seen was $10B, with $1B raise announcements still hitting top headlines. People - including some prominent investors - told me I was crazy.
Over a year ago, I said companies would raise $100B at a $1T valuation, which is nearly what we just saw with OpenAI. Again, people said I was crazy.
When you start to think of these companies as GDP movers with massive opportunity in healthcare and life science and finance (and not fun tools to create party invites, though I delight in doing that too), it’s easier to map out potential impact.
Who knows if I’m right. And who knows what the timeline will bring. We could absolutely see a tech dip in the coming months. But in the fullness of time, as Amazon folks always used to say, I believe AI providers have much more value that can be created compared to what we see today, and I believe the valuations haven’t hit their maximum.
Allie K. Miller: We will see companies raising $100B at $1T valuation within a few years.
Databricks already raised a $10B round end of 2024.
I suspect more companies will follow (OpenAI, Anthropic, SSI) this year and next.
(The below tweet is from over a year ago.)
What are my current valuation predictions for OpenAI and Anthropic?
OpenAI was valued at $28B in April 2023. Its latest funding round puts it at $852B.
Anthropic was valued at $4B in early 2023. Today, it’s worth $380B.
In my opinion - and this is not investment advice - these companies could easily reach $2T in valuation in the next two years.
I tweeted in November 2023 that we would see companies start to raise tens of billions of dollars. At the time, I believe the largest round that we had seen was $10B, with $1B raise announcements still hitting top headlines. People - including some prominent investors - told me I was crazy.
Over a year ago, I said companies would raise $100B at a $1T valuation, which is nearly what we just saw with OpenAI. Again, people said I was crazy.
When you start to think of these companies as GDP movers with massive opportunity in healthcare and life science and finance (and not fun tools to create party invites, though I delight in doing that too), it’s easier to map out potential impact.
Who knows if I’m right. And who knows what the timeline will bring. We could absolutely see a tech dip in the coming months. But in the fullness of time, as Amazon folks always used to say, I believe AI providers have much more value that can be created compared to what we see today, and I believe the valuations haven’t hit their maximum.
Allie K. Miller: We will see companies raising $100B at $1T valuation within a few years.
Databricks already raised a $10B round end of 2024.
I suspect more companies will follow (OpenAI, Anthropic, SSI) this year and next.
(The below tweet is from over a year ago.)
I think the Codex app is the best AI coding app in the market. But it almost didn't get built.
From the Codex team:
"We knew we needed to get to a point where it feels really natural to be talking to multiple agents.
But [the app] was actually quite contentious. The IDE extension was super popular. Should we just focus on that? What about CLI?
There was no specific spec for the app. We had a bunch of prototypes that engineers had built during hack week.
The decision to build the app was basically it's the best beginner-friendly interface to manage multiple agents at the same time."
📌 Watch my full episode with the Codex team: https://youtu.be/9qXc-THAvc0t
Peter Yang: "We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team
Peter Yang
I think the Codex app is the best AI coding app in the market. But it almost didn't get built.
From the Codex team:
"We knew we needed to get to a point where it feels really natural to be talking to multiple agents.
But [the app] was actually quite contentious. The IDE extension was super popular. Should we just focus on that? What about CLI?
There was no specific spec for the app. We had a bunch of prototypes that engineers had built during hack week.
The decision to build the app was basically it's the best beginner-friendly interface to manage multiple agents at the same time."
📌 Watch my full episode with the Codex team: https://youtu.be/9qXc-THAvc0t
Peter Yang: "We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team
Simon Willison
Billing different based on text contained in the system prompt is a really bad look
Peter Steinberger 🦞: Anthropic now blocks first-party harness use too 👀
claude -p --append-system-prompt 'A personal assistant running inside OpenClaw.' 'is clawd here?'
→ 400 Third-party apps now draw from your extra usage, not your plan limits.
So yeah: bring your own coin 🪙🦞
Many such cases
We're building software factories everywhere
Kathryn Wu: Built this over a weekend to solve my own pain, iterating with @garrytan’s Gstack 🙌(It helped a lot with thinking through features + design)
An executive assistant in your email - just tag bot@littlekat.co to auto-schedule and add to your calendar.
Now free for founders.
Amjad Masad
After 4 years on the App Store, Apple suddenly blocked our updates but our app continues to organically grow and rose to #1 again!
After 4 years on the App Store, Apple suddenly blocked our updates but our app continues to organically grow and rose to #1 again!
The @vercel team is always listening and engaging with customers on @x. We shape the product as a result of these conversations.
Vercel gives us the velocity to ship, X gives us a nonstop stream of user feedback. Thanks @simhskal & @riyvir!
River Marchand: I'm so impressed by the @vercel team. Last weekend I told @rauchg that more clarity around billing would give me peace of mind and they've already shipped an update to address it.
Creative tools should help you focus on creating so Vercel is quickly becoming a really important
The @vercel team is always listening and engaging with customers on @x. We shape the product as a result of these conversations.
Vercel gives us the velocity to ship, X gives us a nonstop stream of user feedback. Thanks @simhskal & @riyvir!
River Marchand: I'm so impressed by the @vercel team. Last weekend I told @rauchg that more clarity around billing would give me peace of mind and they've already shipped an update to address it.
Creative tools should help you focus on creating so Vercel is quickly becoming a really important
Kpaxs
High-agency people seem to have this weird immunity to embarrassment.
Getting rejected? Not embarrassing, that’s just data collection.
Looking naive? Not embarrassing, that’s just information asymmetry you’re fixing.
Breaking minor social rules? Not embarrassing, most rules are just Schelling points anyway.
What would be embarrassing to them is not trying. That’s the thing they can’t live with.
Kpaxs: High-agency people genuinely believe that reality is negotiable in a "there are always more levers to pull" way. It's about having this bone-deep conviction that if you keep poking at something from different angles, eventually something will give.
Peter Steinberger 🦞
Re @MartinFromTon @openclaw More people need to spend time with AI, to understand the transformative potential of this technology. We need this for the next century. Just doing my part.
I'm a knowledge base MONSTER in Claude Code right now.
Introducing: Claudeopedia.
1) I took @karpathy's 'llm-wiki' idea doc (90% of this, so the biggest credit goes to @karpathy) and
2) Combined it with the /last30days skill (HT @mvanhorn) and
3) Added a /wiki skill with screenshot and download arguments to transfer raw inputs faster and
4) Built an interactive visualization to search my knowledge base (with date ranges to compare knowledge over time!)
5) Set up a "question your assumptions" cron job that runs my recent writing/client emails against the wikis
All happening in Obsidian for now. All of this was done this weekend, including testing. Will keep adding more features.
For now, the main topic I'm building out is (surprise, surprise) enterprise AI.
I'm drooling.
I need a Claude-branded bib.
Probably the best way to catch up to what is really going on is listen to the newest @steipete @lexfridman interview
It’s a perfect encapsulation of the extreme edge of what is possible now in codegen
Just in time smart personal software is here, just not evenly distributed!
Legalize self driving cars
The Road to Autonomy®: Why can't @Waymo drop-off curbside and pickup at the @flySFO rideshare garbage? Politics.
Golden age of open source is here
ℏεsam: bro created an AI job search system for Claude Code that scored 700+ job applications and actually got him a job.
AND IT'S NOW OPEN-SOURCE.
It scans multiple company career pages, rewrites your CV per job, and even fills application forms. The repo has:
> 14 skill modes
Rob Henderson
"all of this dysfunction is bizarre because San Francisco is the richest city in human history...The extremes of wealth and poor here are awesome, and I can see why so many are now communists, except that it’s the rich who are communists." https://www.edwest.co.uk/p/the-city-of-luxury-beliefs
Happy Easter!
مقال رائع عن عقلية ريادة الأعمال
نادي ريادة الأعمال: http://x.com/i/article/2040473188568649728
if you're seeing this, make sure you turn thinking on
claws are actually pretty great in gpt-5.4, but they seem really stupid unless this gets turned on
Jeff Escalante: Transition to gpt in openclaw not going so great this far 😂
Morgan
Shipping without traditional specs or roadmaps, no longer the way of the future.
Happening, now.
Peter Yang: "We write very few specs on the Codex team. We're talking 10 bullet points and that's it."
Here's my new episode with @embirico and @romainhuet where they gave me an inside look at how OpenAI's Codex team operates:
→ Live demo: Building in seconds with Spark
→ How the team
New @X API + @Replit is lots of fun.
@tannerlbraden built a @NASAArtemis II mission visualization w/ live @X feed & stats.
Amjad Masad
New @X API + @Replit is lots of fun.
@tannerlbraden built a @NASAArtemis II mission visualization w/ live @X feed & stats.
nyx 👻
Sharing my review for @conductor_build (cc: @charlieholtz ) 🧵
Next up: @getsome_air , @agentastic and more.
More observations from Shanghai:
1. A full-time, live-in nanny costs only $1,500/month and a personal chef costs $7/hour. There's alot of support for professional working couples here.
2. Didi (Chinese Uber) rides are $3-5 for most trips and you can order delivery for anything for a few bucks. Things are super convenient.
3. Speaking of cars, every Didi I've been in has been a Chinese EV. Feels like China has adopted EVs much faster than the US. Tesla has <5% market share here.
4. The best food is inside the high-end malls, which are everywhere. Service is outstanding at most places and you don't have to tip.
5. Now the tradeoffs - there are ALOT of people. Traffic is everywhere and motorbikes have no qualms about riding on the sidewalks. Have to be on the lookout for my kids.
6. I haven't seen a single blue sky day since I've been here. The air does feel a bit cleaner now thanks to the EVs.
Overall, if you make anywhere close to US tech salary here you can live very well.
Peter Yang: Some initial observations about Shanghai after not being back for 10 years:
1. The city is incredibly modern - more so than New York and even Tokyo. It's funny riding modern subways and trains here and reading about how California has to shut down the BART/Caltrain due to budget
Nous Research
Introducing the Manim skill for Hermes Agent.
Manim is an engine for creating precise programmatic animations for mathematical and technical explainers, made famous by the @3blue1brown channel.
