Met some legends and friends today at Code with Claude.
Of all the AIs, Claude still feels the most like a trusted friend and I’m glad that now @AnthropicAI has the compute to scale.
http://vibecon.ai
Lenny x Replit
Lenny Rachitsky: Announcing the Lenny's Newsletter x @Replit Buildathon
Build something awesome with my podcast and newsletter data using Replit, and win fabulous prizes:
🔸 1:1 career coaching session with @amasad
🔸 $5,000 in free Replit credits
🔸 A free year of Lenny's Newsletter
This is
Daniel Lurie 丹尼爾·羅偉
The RESET Center has been open just over 24 hours, and already dozens of people using drugs in public have been brought off the streets. This is a new model that we are using in San Francisco to bring people on the streets into a health focused facility where they can have a chance to be connected to treatment.
This center is making it easier for officers to get back on the beat while giving people a real opportunity for treatment. We’re also hearing reports that staff have connected people with case managers who had been trying to reach them. This is an important new tool to help keep our streets safe and clean.
openclaw + cline velocity is underrated
pash: @swyx @vincent_koc https://github.com/openclaw/openclaw/pull/78234
fixed here - problem was you had guardian mode set in your config instead of default yolo, which exposed another bug that is fixed in this pr
Design, build, and operate compute with us at planet scale:
Uday Ruddarraju: There is a lot of news about compute being the bottleneck for AI. There is less visibility into the engineering it takes to make large-scale compute actually work reliably.
In my view, this is one of the most interesting computer science problems in the industry right now. It is
Peter Yang
Q: Can two AI agents schedule a meeting for their humans?
A: Yep 😅
Q: Can two AI agents schedule a meeting for their humans?
A: Yep 😅
Nice qmd vs gbrain benchmarking post
GBrain won 8.3x on this particular corpus
https://www.ai-heroes.co/en-gb/blog/gbrain-vs-qmd-benchmark-may-2026
lol there are 80 bids too
Ryan Cohen: I’m selling stuff on eBay to pay for eBay
https://www.ebay.com/usr/ryan_5050
Jatin Garg
Re the underrated story of 2026 is that the open-source agent ecosystem is leading on primitives. nous research with hermes for orchestration. gbrain for personal memory and eval substrates. these projects shipped working production systems before anthropic shipped a research preview of similar functionality. the closed labs have raw model capability. the open-source ecosystem has agent primitives. those are different layers. the open-source side has been further ahead on the second one for nearly a year now.
BrainMirror AI
AI Heroes ran gbrain against their own production memory stack on 150 real questions from their actual corpus. gbrain won 58 head to head matchups. Their own system won 7. The methodology, the surprises, and what the aggregate score hides 🧵
TL;DR:
→ 352 file corpus, 150 questions built from real operator sessions, not synthetic evals
→ gbrain won 58 questions vs qmd's 7 in apples to apples retrieval, an 8.3x win ratio
→ gbrain ran 41x faster: 608ms median vs 25,138ms for qmd native pipeline
→ qmd's LLM reranker actively reduced recall on this corpus
→ gbrain's graph extractor produced 0 typed links, every win came from hybrid retrieval alone
Marco and the AI Heroes team disclosed the conflict of interest directly: they run qmd in production and had a vested interest in it winning. The data did not cooperate.
They are pulling the reranker from production qmd regardless of whether they migrate to gbrain. That is the kind of benchmark worth reading.
@garrytan @openclaw
Ryan Petersen
The idea is to drive all the sane people who would oppose them out of the state so they can have supreme power in the world’s most op geography. Seems to be working.
Katie Porter: I’m Katie Porter and I approve this message.
Open source is eating hardware now too
https://www.digitalfoundry.net/news/2026/05/valve-releases-steam-controller-cad-files-under-creative-commons-license
Josh Reeves
Our mission at @GustoHQ is to grow the small business economy with technology and heart. We are deeply grateful to all the work small businesses do every day in communities everywhere, and we are committed to being their partner, taking hats off their head, delivering them peace of mind, and giving them superpowers, so they can progress their business.
@agarfinks from @FortuneMagazine has been following Gusto for a long time, and we recently shared with her some milestones in our journey of building Gusto (including passing $1b of revenue). These are still the early days of what’s possible. We are building Gusto for the long term, and when small businesses succeed, we succeed.
tae kim
Paul Tudor Jones on @cnbc “bought more AI stocks” “semiconductors” “It’s a crazy crazy time” brings up introduction of PC, Claude Code -> Microsoft 1981, Windows 95/internet. “beginning of productivity miracles that lasted 4-5 years” “we have a year or two to run” or “we continue to feel like 99” “October/November 1999” in terms of multiples. (either two years to run, another ramp to go)
Matt Dorsey
“Change the Party?” The Democratic Party?!
Call me crazy, but I think pro-terrorist, anti-American, antisemitic, anti-AAPI, misogynist, xenophobic, politically repellent Hasan Piker fanboys should go form their OWN party — and get the f*** out of ours! (1/3)
Giga
Introducing hallucination correction. We have reduced hallucination by 70%. Giga's hallucination rate is at ~1%. Better than the best frontier models.
Deploy AI your customers can trust.
GBrain will soon support multiple topologies
1/ local PGLite
2/ multiple thin-clients (e.g. Hermes Agent works with OpenClaw's hosted GBrain MCP server)
3/ multiple thin-clients WITH local PGLite for code repos (e.g. Claude Code worktrees with code indexing stay local while GBrain plan files are federated to your central GBrain in the cloud)
T Wolf 🌁
The fact that Progressives support billionaire Tom Steyer while watching him outspend his opponents by more than 20x using his own money, proves that they're stuck in a loop of hypocrisy. It's actually foul to see.
NYU Tandon
Research from Prof Julian Togelius found that despite AI's well-documented victories in chess, Go, and Atari games, humans still learn unfamiliar video games far faster than any AI model.
#NYUTandonMade
https://buff.ly/fbVkDrh
Erica Sandberg 舊金山的神奇女俠
Repulsed by Saikat Chakrabarti partnering with Hasan Piker? You're in the majority. I'm hearing it from every corner, all types of people. what are your thoughts?
thanks to @mattdorsey for speaking out. @saikatc
Matt Dorsey: “Change the Party?” The Democratic Party?!
Call me crazy, but I think pro-terrorist, anti-American, antisemitic, anti-AAPI, misogynist, xenophobic, politically repellent Hasan Piker fanboys should go form their OWN party — and get the f*** out of ours! (1/3)
Tom Reed
I don't think automation of AI R&D will rapidly lead to domain-general super-intelligence.
I think this will be true even if AIs can do *literally everything* a human AI researcher does today.
Even after the full automation of AI R&D, further capabilities progress will only happen through
(1) widespread deployment of AI throughout the economy, accompanied by data collection; and/or
(2) the wholesale recreation of much of the economy by AI labs.
Without access to the real-world signal provided by either of the above, I think that the only thing produced by automated AI researchers would be a "Goodhart Singularity".
If I'm right, this is obviously good news. I make the case for this in a new piece on my substack
Matt Van Horn
Introducing the Printing Press, a CLI-factory and a CLI-library. Built with @trevin. 🏭🖨📚
Most APIs suck for agents. Most MCPs suck for agents. Most official CLIs suck for agents. They waste tokens and time. @steipete started making his own because of this.
📚 A Library of agent-native CLIs you install today (Linear, ESPN, Flight GOAT (Google Flights + Kayak nonstop), Contact Goat (LinkedIn + Happenstance + Deepline more) +30+ more)
🏭 A factory that prints new ones for any service - just type /printing-press
CLIs are fast, local, SQLite-backed. Work in Claude Code, Codex, OpenClaw, Hermes.
🌐 https://printingpress.dev
Charlie Hills
Garry Tan just open-sourced his Claude Code setup.
It's a 6-person AI team in your terminal:
I installed it in 30 seconds.
Each project now runs through 6 specialists.
Each agent owns a phase of the build:
✦ CEO challenges every decision before code
✦ Eng Manager locks in architecture upfront
✦ Designer ships 4-6 variants, picks winner
✦ Release Manager creates the PR and deploys
✦ Doc Engineer writes the changelog after sprints
✦ QA Lead runs real browser tests and audits
The CEO agent alone changes how you build.
It asks "why does this need to exist?"
Before you write a single line of code.
People use Claude Code as a solo assistant.
gstack turns it into a team.
It's free and open source.
Repo: https://github.com/garrytan/gstack
Start with Claude Code
https://charliehills.substack.com/p/claude-code-beginner-advanced
Get more free AI guides here
http://charliehills.substack.com
Repost ♻️ to help someone in your network.
P.S. Which agent would you reach for first?
Clawvisor is going to be one of the most important parts of helping make the agent world especially OpenClaw/Hermes Agent secure and enterprise-grade.
We're in the Apple I moment for personal AI, but we're ABOUT to see the Apple II - the first moment when everyone can use it
Y Combinator: Clawvisor (@clawvisor) lets you give AI agents access to apps like Gmail and Slack without handing over your credentials or worrying they'll go rogue. You approve tasks once, Clawvisor enforce them.
Congrats on the launch, @ericlevine!
https://www.ycombinator.com/launches/QFP-clawvisor-the-authorization-layer-for-ai-agents
Journalism is journalism not political activism, but a lot of people in the profession seem to have forgotten
Garrett Langley: “Snitching ass startup.”
That’s what a reporter called us when we started @Flock_Safety. I framed it.
Hard work on important issues comes with scrutiny.
If it helps more people learn about the dynamics of crime, good.
If it helps deliver justice to victims, solve more than a
CLI made with Printing Press is a powerful manifestation of Fat Code when you need Fat Code + Fat Skills + Thin Harness
Matt Van Horn: Contact Goat. LinkedIn + Happenstance + Deepline, fused into one CLI.
"Do I know anyone connected to the new Apple CEO and what's his email?"
Happenstance found my warm connection through @CyrilLabidi . Deepline paid $0.10 to get his email address so I didn't have to bug Cyril.
Harj Taggar
Clawvisor was the unlock for me to connect my openclaw to more services and use it for actual work. So much room for it to grow.
Y Combinator: Clawvisor (@clawvisor) lets you give AI agents access to apps like Gmail and Slack without handing over your credentials or worrying they'll go rogue. You approve tasks once, Clawvisor enforce them.
Congrats on the launch, @ericlevine!
https://www.ycombinator.com/launches/QFP-clawvisor-the-authorization-layer-for-ai-agents
My OpenClaw/Hermes Agent setup just got way smarter
Matt Van Horn: Introducing the Printing Press, a CLI-factory and a CLI-library. Built with @trevin. 🏭🖨📚
Most APIs suck for agents. Most MCPs suck for agents. Most official CLIs suck for agents. They waste tokens and time. @steipete started making his own because of this.
📚 A Library of
NYU Tandon
On May 18th, Yann LeCun will be addressing the NYU Tandon Class of 2026 at the Barclay's Center.
#NYUTandonMade
https://buff.ly/D8LwlPY
Sarah Chieng
"Technical writing completely changed my life." - @trq212
In under 2 years, Thariq (@AnthropicAI) cracked the code on writing technical articles that consistently hit 1M+ views.
In this 20-min workshop, he breaks down:
→ his exact writing workflow
→ the tactics behind articles that go viral
→ how he leverages AI to write faster (without losing his voice)
→ why technical writing is the most underrated way to build mindshare
Technical writing is one of the most powerful (and completely free) ways to gain views, build authority, and teach the world what you know.
This is the 4th edition of [Technical] Write & Learn, a curated workshop series cohosted with @swyx and @KernelLabs_ai.
You can now just build amazing voice agents, with the GPT-Realtime-2 reasoning model in our API:
OpenAI: Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents.
Voice agents are now real-time collaborators that can listen, reason, and solve complex problems as conversations unfold.
Now available in the API
Ben Landau-Taylor
Reading another "Why don't rich guys build libraries like Carnegie anymore" article.
It's because that's illegal now. In most cities you can barely even build an apartment. Never mind the psychoanalysis, it's just against the law.
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
I am concerned that the Dems are becoming the party of "millionaires who resent billionaires".
"I made my millions fair and square, but you cheated and exploited the workers to make your billions, you capitalist pig!"
Marco Foster: AOC: “There’s a certain level of wealth and accumulation that is unearned. You can’t earn a billion dollars. You just can’t earn that. You can get market power, you can break rules, you can abuse labor laws, you can pay people less than what they’re worth, but you can’t earn
Rork
Introducing Rork AI Cloud
Rork can now use any of the 150+ models and one-shot almost any AI app, even Higgsfield.
Take full control over which model your app uses: a movie maker on Kling 3.0, a photo studio on GPT-Image 2, an AI tutor on GPT-5.5, or a voice assistant on ElevenLabs. No API keys needed.
Many such cases
Gowtam_Imagines: GStack Office hours Skill made Claude one of the best brainstorming partners. As a solo technical founder, this has been extremely helpful! Claude kept pushing me further and further on why a product needs to be built. Thank you @garrytan !
T Wolf 🌁
Becerra is not the answer. Porter and Steyer will tax the state into oblivion. If you're a Democrat, the only real choice you have is @MattMahanSJ. I'm not kidding.
POLITICO: ‘Him? Really?’: Becerra’s rise baffles former Biden colleagues http://dlvr.it/TSQQb3
California deserves better than a resource trap that destroys effective governance
We need effective spending on fixing real problems instead of whatever the state has been doing to date
Garry's List: California spending grew 70% since 2019. Revenue grew 60%. The state's own nonpartisan analyst said it: You can't tax your way out of a 10-point structural gap.
https://garryslist.org/posts/the-deficit-california-can-t-tax-away
They’re calling it the most viral petition in history, and it’s hosted on Replit.
(We have no opinion on Mbappe)
Polymarket Sports: 🚨The MBAPPE OUT petition has gotten over 10 million signatures in the last 24 hours.
It’s on track to be the most signed petition in history.
OpenAI
Codex now works directly in Chrome on macOS and Windows.
It’s even better at working with apps and sites in Chrome, and now works in parallel across tabs in the background without taking over your browser.
To get started, install the Chrome plugin in the Codex app.
have been excited for realtime voice-to-voice translation as an AI application since we started OpenAI. extremely cool to see it now available in the API for anyone to build with:
jason liu: 新しいリアルタイム翻訳モデルを発表できることをうれしく思います。ぜひ本日よりAPIでお試しください。
Patrick OShaughnessy
Why Brian still obsesses over recruiting:
"Sam Altman (@sama) told me you're going to spend 50% of your time on hiring. I never did. It was my death blow.
As a leader, you can choose if you want to spend time hiring or managing. The more time you spend on recruiting, the less time you get to spend on management.
The first and last call I make every day is the recruiting team, still. I probably spend two, three hours on it every day.
The 2000s, I didn't. I thought it was all about having a recruiting machine, managing people. The great thing is I don't manage as much anymore because the really good people are self managing.
People should think about their first employee being a recruiter, not an engineer. Because they are the ones that get you every other person.
A company is as good as its people. The difference between the good companies and the great companies are the people."
Patrick OShaughnessy: My guest today is Brian Chesky (@bchesky), founder and CEO of Airbnb and one of the great consumer founders of the last 20 years.
Paul Graham coined "founder mode" based on Brian's experience running Airbnb. This conversation is about what comes after it, what he calls AI
Sudhir Mantena
http://x.com/i/article/2052466094016008192
way cooler to help software developers pokemon-evolve into superheroes than to try to replace them
it is insane what one really good person can do now
Matt Pocock
Grill issue, incredible
Here's yet more AIE Europe content from me
swyx 🌉: @mattpocockuk @aiDotEngineer we just released a @latentspacepod chat about his talk and how he works as well
https://youtu.be/rlM_fAKxB3Q
Space launch was a clear case where there was a large difference in efficiency between what was possible and what was done in practice before SpaceX. A large part of that was due to everything being locked in to what (just barely) already worked, with huge risk aversion. WIth national prestige or a half billion dollar geosync satellite on the line, speculative engineering ideas that might result in a public debacle were not welcome.
When failure is not an option, success can stay very expensive. You need to experiment to improve, and that fundamentally means being comfortable with failure. If you know it is going to work, it isn’t an experiment.
I have long believed that nuclear power today is in precisely the same state as space launch two decades ago, but the even more pressing question now is if semiconductor fabrication might also be.
On the one hand, Moore’s Law has been a sequence of heroic miracles of technology at the wafer fabrication level, grinding out hundreds of compounding small improvements.
On the other hand, fabs are “too big to fail”, and there are elements of extreme conservatism at play. Intel’s “Copy exactly!” fab development exemplifies that mindset – instead of every new building being an opportunity to explore and optimize processes, it was deemed more valuable to just replicate.
While each individual machine may be straining against physical limits of technology, it is possible that the systems orchestrating them all together could be far from optimal.
The explore / exploit axis is fundamental to all decision making, but human risk avoidance probably biases away from optimal exploration.
Andrea Junker
Total Jobs Created by Party (1989-2026):
Democratic Presidents
50,600,000
Republican Presidents
1,469,000
The biggest scam the GOP has ever pulled off is convincing people that Republicans are good for the economy.
Sergio Duran
Just connected 7 AI clients to one brain.
Claude Code, http://claude.ai, OpenClaw, Hermes, Codex CLI, ChatGPT — all writing to the same GBrain knowledge base.
Every conversation, every platform, one persistent memory. Here's how 🧵
Everyone hates gstack until they try it and then they love it
Vexoa: Been cooking with @ycombinator's GStack, hated it but once i got used to it. My projects are shipping a lot faster, working on PRs takes half less time now.
Thin Harness / Fat Skills is actually a powerful mindset shift that helps you reduce your bugs in agentic flows
Marc Lindsay: Part 4:
The Engine That Produces:
@garrytan his posting on Thin Harness / Fat Skills has been a huge driver in how I create this.
@MiniMax_AI it's taken me a while.... but you can see here what I was talking about.
Check out my other part 1 through 3 for what has been built
Paul Graham
Sure you can earn a billion dollars. I've been teaching people how to do it for 20 years. The way you do it is to start a company that grows fast. You don't have to do anything bad to make a company grow fast. You just have to make something people want.
https://paulgraham.com/ace.html
Marco Foster: AOC: “There’s a certain level of wealth and accumulation that is unearned. You can’t earn a billion dollars. You just can’t earn that. You can get market power, you can break rules, you can abuse labor laws, you can pay people less than what they’re worth, but you can’t earn
Scott Kennedy ⠕
If you wonder what kind of things I care about or am excited about at Replit, just check out the blog and see what our team's been shipping.
Much more to come, we're not nearly done. Every builder deserves reliable and secure software🔒
Morgan J. Freeman
Trump Supporters Complain About Not Receiving Illustrious Gold Trump Phones After Paying $100 Deposits
Hundreds of thousands of Trump supporters paid $100 deposits for the Illustrious Gold Trump phone, also referred to as T1 or Trump Mobile, but have not received the devices months later. Posts claim Donald Trump Jr. and Eric Trump collected around $60 million from these preorders, with the website fine print stating no guarantees of production or refunds.
WTF did they expect?
Codex can now drive Chrome tabs in the background:
James Sun: Today, we are excited to introduce Codex for Chrome!
Now, Codex can drive its own Chrome tabs in the background to automate tasks while you use the browser simultaneously.
It does this by opening up tab groups for each task, cleaning up at the end, and handing back tabs for
Many such cases
Sid - e/acc: @karpathy and @garrytan skills helping me to get 10x productivity. Here is my retro, which I do at the end of every cycle to keep pushing boundaries.
x.com
