Excited to see everyone at Code with Claude! What would you like to hear from us at this year's sessions?
Claude: Our developer conference Code with Claude returns this spring, this time in San Francisco, London, and Tokyo.
Join us for a full day of workshops, demos, and 1:1 office hours with teams behind Claude.
Register to watch from anywhere or apply to attend: https://claude.com/code-with-claude
Daniel Owens
This is the difficult truth that many San Franciscans want to deny:
SF is not a modern city when it comes to infrastructure. Local leadership has systematically worked to create a bloated City Hall that focuses on identity politics and virtue signaling rather than modernization.
Bojan Tunguz: Every time I visit San Francisco I have the opposite of the “future shock”. I am absolutely floored by how backward and dilapidated most of the city is, especially the city that prides itself on being on the forefront of the future. There have been barely any significant
Jaclyn Konzelmann
13 years ago I was a founder in that room.
Today I got to experience it from the other side.
Nearly 200 companies on stage. Some had been building for years - deep tech, robotics, hardware that takes time to get right. Others started from an idea and built their whole company during the batch. One founder I talked to pivoted mid-batch and rebuilt from scratch.
All of them on the same stage. All of them moving.
The cost of starting has never been lower. The builder energy is very much alive. 🧡
Shubham Saboo
“OpenClaw is the iPhone of tokens” — Nvidia CEO on Lex Podcast
Sarah Wooders
Why does Anthropic have such hardass licenses for even their skills lol
Meanwhile most Codex skills are Apache 2.0
OK you caught me
Sinclair Ta: Garry Tan @garrytan just ran the sharpest VC marketing play I've seen in years.
And most people watching missed the architecture of it entirely.
Here's what happened:
Garry Tan — president of Y Combinator — built and open-sourced GStack: 15 Claude Code skills for founders
We have to accelerate
The Boring Marketer: what I'm seeing is that top 1% operators are collapsing into something beyond "vibe coders" it's more like can understand systems and architecture enough to be super dangerous with Claude Code. you apply your deep vertical knowledge to this and you become unstoppable.
Sherry Jiang
we're stoked to have @GeoffreyHuntley - the creator of ralph - coming to @aiDotEngineer singapore in may!
more cracked builders to announce soon...all coming to the red dot!
@swyx @agrimsingh @ivanleomk @aimuggle @unprofeshme
Gemini always getting wrecked in these screenshots
fai nur: the TSA app is now a top 4 app on the app store alongside chatgpt, claude and gemini
Kendrick Lau
YC is the YC for hardware. And ex-Tesla founders are here to get it done.
congrats on an inspiring W26 demo day!
@fortwearable
@gru_space
@VoxelEnergy
@ycombinator
This data from Lenny is super cool to see but I'd love to see a cut that's filtered to the top 50 companies (FAANG, AI-native companies, top SaaS, etc) or so that PMs really want to work for.
My guess is that the openings aren't growing as fast or could be declining.
Lenny Rachitsky: PM openings are at the highest levels since 2022
There are over 7,300 open PM roles at tech companies globally, and trending up. This is 75% above the low we saw in early 2023, and already up nearly 20% since the start of this year. Today we have the most open PM roles we’ve
Jon Xu
Congrats to all the YC W26 founders building generational companies. As we wrap up demo day, I’m so excited to live in the future that these companies are creating. 🚀🚀🚀
Y Combinator: happy YC Demo Day to all who celebrate
The awkward truth is that what counts as a good engineer just became a different thing in the last 4 months
I hope AI agents will help us all stop obsessing about:
- Specs
- Designs
- Issues
- Roadmaps
- Strategy docs
- Spreadsheets
and instead let us all obsess about improving the end user product and rapid iteration loops with users.
Product development is much more fun that way.
Linear: Issue tracking is dead.
We are building what comes next.
https://linear.app/next
Latent space is amazing, and "it's just prompts" is a lie
Also I'm giving it away for free, so if you want, I can pay you back what you paid me.
rohan: @garrytan i was actually really skeptical about gstack but i tried /office-hours and it was pretty good.
produced 2 actionable suggestions that i think generally are good ideas.
I've scaled app servers, I've scaled HAProxy, I've scaled Postgres replica dbs, I've scaled job servers
Never until today have I had to scale my local max_connections for Postgres but once you run 20 simultaneous workers doing full tests, now I have had to do even that.
Rohit Agarwal
Re this isn't a coincidence.
@swyx mapped it best -- code sits in a unique quadrant: high economic value, short timeline to get there.
three reasons coding wins first:
→ code is verifiable. commits, prs, tickets. you can measure everything
→ a swe-1 costs $130-200k/yr. the roi math is instant
→ it's the only ai category that's recursively self-improving - the people building it are also its first users
no other ai category has that flywheel.
Brycent
Type sh*t with GStack @garrytan 🐐
Brycent: YC is letting me bring a camera man to demo day! 🥹
@garrytan I need to interview you!
Yana Welinder
14 YC startups hit $1M ARR before demo day 🤯
That's before raising a dollar of VC.
AI is massively expanding who gets to build.
No more gatekeepers.
Garry Tan: YC Demo Day for W26 is in full swing
The craziest stat: 3X more companies in this batch reached $1M annualized revenue than W25
Also crazy: the fastest revenue growth rate of YC history at 14% week on week growth *on average* across the whole set of nearly 200 startups
For the polymath designer-coder, this is kind of like catnip https://app.principal-ade.com/
Building your own tools like GStack is an amazing experience. You come up with something you think might work and then two days later it replaces what you were doing before.
Instead of /plan-ceo-review and /plan-eng-review sequentially, I find myself using "/autoplan with adversarial review" every time (it is kind of like Claude's -auto-mode but for making choices that my prompts used to ask)
And after I finalize the plan and exit plan mode, CC codes it all, and then I do "/review with full claude and codex, full adversarial with claude and codex"
And then it just one-shots everything from there, no bugs, 100% test coverage
Conductor is still my favorite harness-harness
Charlie Holtz: Introducing our new, simplified sidebar!
You can now see GitHub status at a glance (merged, failing CI, conflicts)
Hover any workspace to see metadata and get easy access to next actions without leaving your chat:
This is 100% true, I like to read the reasoning traces too
kache: you still need to read and understand code
Oh snap GStack meetup!!
toki: i'll be in sf end of april
thinking of hosting a gstack meetup
@ycombinator, @gumloop, @browserbase any of you want to sponsor the venue?
hopefully @garrytan shows up.
who's in?
Shernaz Daver
When the Italian talks vibecoding with the Brits! @pirroh of @replit with the hosts of @etnshow
etn.: Michele Catasta (@pirroh) President of Replit (@Replit) says "AGI for vibe-coding is coming earlier than 2028":
"I would expect by the end of 2026, a lot of the core functionalities that models and agents have to have, in order to become very good vibe coding workhorses, will be
Lots of engineers think AI codegen is only good enough to do little bug fixes here and there
If you tell them you can ship 15k LOC of ai gen code to prod they think you have lost your mind.
It’s so so early. And also those engineers are living in 2025.
Many such cases
Adam Nutt: @garrytan My experience was very similar! I used my framework to dogfood my framework, then I added gstack to the framework and shipped 8 versions in 7 days.
Once you say “replace developers” it’s the wrong ideaspace
Cars didn’t replace walking, you still have to walk a little bit AND
It let people go to places you can’t go to on foot alone
Sarkar_connect: @garrytan This is contradictory. One hand it's being spread that AI will replace all developers, it's so good.
Other hand we need engineer to keep following codes generated by AI. ?
If ENGINEER need to check the codes generated by Ai , how can ai replace the engineer
He just like me
mal: hearing karpathy on this podcast talk about how
he is always anxious he isnt using coding agents to their full extent
brings me comfort cuz it seems like everyone is trying to figure it out
It is actually going to be 90x my 2013 output 👀
Steve Johnston: Spent the day at @ycombinator's W26 Demo Day in San Francisco, and a few things stood out from @garrytan's opening remarks:
• ~90% of the batch is AI
• Solo founders can now build faster than entire teams could just a few years ago
• Garry said he's built more software this
GitHub is the new Substack
Code is a form of self-expression
claire vo 🖤
Random @openclaw tips that are super simple but almost no one realizes
- your sessions poof at 4 am, overnight amnesia is built into the system
- you don’t need a monitor for your Mac mini turn on screen share & remote in from your laptop
- it’s SOUL is promoted to not bother you overnight or talk to much, it tries to get you to go to bed
- you probably have your session dmscope or cron target sessions set wrong which is causing your claw to act like it has a tbi
- it’s probably stashed an API key somewhere it shouldn’t
- you can give browser profiles their own color so you can tell when it’s working in its profile
- you can auth gog with read only perms
- read the docs on heartbeat vs cron
- give your telegram bots cute emojis
💥Susan Dyer Reynolds🗞️
Leaving drug addicts to commit slow suicide on the streets is not fair.
Mayor Matt Mahan: It’s not fair. It’s not compassionate. It’s not progressive. We need to do better.
The Marina Times ⛵️ 🗞️
50,000 people have died on our streets in the last decade — half from drug overdoses and suicide.
Mayor Matt Mahan: It’s not fair. It’s not compassionate. It’s not progressive. We need to do better.
What VCs do on company boards, an evolution:
2016: governance, guidance, support
2021: cheerleaders
2026: Anthropic and OpenAI sales representatives
Mike Solana
union leftists a greater impediment to progress than literally the vacuum of space
robyn: Why does it cost less to build a base on the moon than high speed rail from sf to la
Jawwwn
Palantir CTO @ssankar in 2024:
“For $10 billion, @elonmusk put 300 rockets in orbit.”
“For $11 billion the state of California has built 1600ft of elevated rail, with no rail.”
Marc Andreessen 🇺🇸: Everything costs less than the cost of high speed rail from SF to LA, which is infinite.
Katie Miller
Twelve years ago, San Francisco killed 8th-grade Algebra in the name of equity. Advanced math enrollment (including AP Calculus) dropped, and racial gaps expanded.
China graduates ~1.3 million engineers yearly vs. just 130,000 in the U.S.
Last night, the SF school board finally voted 4-3 to bring math back.
When DEI overrides academic standards, America falls behind.
https://www.nytimes.com/2026/03/24/us/san-francisco-public-schools-algebra.html?unlocked_article_code=1.V1A.V32R.pJ2hBlQYtmw9&smid=tw-share
If you believe in the efficient markets hypothesis, you need to explain why it took until now for this to happen
United Airlines: The entire row is alllllll yours.
Welcome to United Relax Row, three adjacent United Economy seats with adjustable leg rests that can each be raised or lowered to create a cozy lie-flat space for stretching out...
You'll also get a mattress pad, blanket and two pillows. If
Maybe the humans need a “do you know what the machine is doing?” harness
The “it makes unmaintainable code” argument is actually a bit more of human management failure than machine failure at this point
Garry Tan: @GaryHoberman Get better test coverage. Read the reasoning traces and see if you agree
If there is no recognition of your desired patterns in the code or the thinking trace then you are right it’s impossible to support.
If I could wave a magic wand to bring the future into the present faster, iPhone would become an open platform with open APIs and real consumer choice instead of a walled garden that is increasingly behind the times like the travesty that is Siri
Naval: AI coding agents can now deliver one-shot custom apps straight to your phone.
It’s the beginning of the end for the iPhone’s dominance.
Igor Kudryk
After reading it, this should be bigger news. Crazy stuff.
Why it's cool:
Hermes agent = self-improving memory & skills.
HyperAgents = self-improving behavior of the agent.
For example, it starts without a memory and after a few iterations discovers the need for particular memory system that works for you specifically.
This is the future.
Jenny Zhang: Introducing Hyperagents: an AI system that not only improves at solving tasks, but also improves how it improves itself.
The Darwin Gödel Machine (DGM) demonstrated that open-ended self-improvement is possible by iteratively generating and evaluating improved agents, yet it
Mario Zechner
I'm usually not one to write thought pieces without much technical depth. But here we go.
Slow the fuck down.
https://mariozechner.at/posts/2026-03-25-thoughts-on-slowing-the-fuck-down/
Wesley Yang
SF public schools finally ends its 12-year experiment in removing the option of taking algebra in the eighth grade in pursuit of racial equity.
Every retarded left-wing nostrum that has no possibility of working gets a decade+ of runway before we have enough studies to prove that the thing that could never work didn't work.
The key problem of public policy must be how to shut down this pipeline, which begins by mass defunding of corrupted academic programs and NGO's.
👀
Jared Duker Lichtman: In the latest paper of Terry Tao
Peter Yang
The biggest time-waster at most tech companies isn't meetings or Slack.
It's trying to predict what your leader wants.
In my latest post, I break down how a VP at Meta solved this by building an AI skill that reviews docs in his voice:
→ What the /exec-review skill does
→ How to build a leader profile in 6 steps
→ The full skill files so you can set it up today
📌 Read now: https://creatoreconomy.so/p/build-an-exec-review-ai-skill-to-stop-guessing-what-your-leader-wants
pedram.md
men in their 40s used to have cool midlife crisis but now they just have agentic workflows
Some observations from @HillValleyForum yesterday:
- It takes three to tango for hardtech/deeptech. Builders+Investors+PolicyMakers. All 3 were out in full force.
- The speaker lineup was top-tier. It's indicative of the great job that @zebulgar @CGarrett_15 @jacobhelberg have done but also to the fact that there is:
- Insane energy around hardtech+deeptech. Drones, Energy, Robotics, Atoms+Bits.
- This energy is infectious...but oddly not very Silicon Valley / Bay Area pilled. It's in a bunch of disparate places around the US. I think this is a good thing.
- The support from Washington / the political class is very real. This was awesome to see. Every single congressman+senator that I interacted with wanted to be part of building something amazing. Having folks like @sriramk in DC helps a lot.
- @shaunmmaguire captured this in his interview with TBPN yesterday but I think the interesting aspect of moving from bits to atoms+bits is that you also diversity your sources of capital. Lots of interesting conversations with investors who are far from traditional venture.
Super optimistic for 🇺🇸.
We have a lot to do but things are going in the right direction.
New @openclaw beta is out with better MS Teams integration, @OpenWebUI and more!
The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build products in this world, I invited Mike Krieger (@mikeyk) on @every’s AI & I podcast. Mike cofounded Instagram and is now a member of the technical staff at Anthropic, co-leading Anthropic Labs, their internal incubator for experimental products. He's been at the frontier of two transformative technology waves: mobile/social and now agent-native software.
We discussed:
- How to build a truly agent-native product. The best products today, like Claude Code, allow users to do things that their creators never intended. But that requires hard trade-offs between freedom and safety/reliability for frontier products, an issue that Mike's team is learning how to solve.
- What's different about building now versus building Instagram. At Instagram, it took months to hit dead ends and learn what to cut. Now, that cycle runs in hours.
- The trap of building too much, too fast with agents. You can go from idea to a nearly-shipped product in a day, but that process doesn’t give you the incremental feedback that used to tell you what not to build. The models are great at adding features, but can create a product that lacks coherence.
- How Anthropic Labs structures product teams. New product experiments are led by only two people, usually a product manager or designer paired with an engineer. Mike says bigger teams tend to be too slow because of coordination costs.
- Why you need to throw out your product and start over every three to six months. AI progress means most of your harness will be outdated quickly—the best teams build this into their product strategy.
And much more! You should watch this one.
Timestamps
Introduction:
What's gotten easier—and what hasn't—about building products in the age of AI:
Why vibe coding creates "indoor trees":
How rewrites have become a normal part of the development process:
What "agent native" product design means:
How Mike's labs team is structured and the cofounder model:
The best signal for a product bet is someone with "break through walls" conviction:
Navigating enterprise customers while keeping pace with rapid AI change:
OpenClaw, personal agents, and the product question defining 2026:
Austin Tedesco
"Why you need to throw out your product and start over every three to six months. AI progress means most of your harness will be outdated quickly—the best teams build this into their product strategy."
This is super frustrating to accept at first, but a much better way of operating when it becomes part of your flow, even with knowledge work optimizations.
Dan Shipper 📧: The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build
Mike Taylor
Damn now I have to listen to this before my claude code for product managers session tomorrow.
Dan Shipper 📧: The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build
The pep rally talk before demo day has become a bit of a tradition
Jai Bhatia: JUST GOT THE GARRY TAN TALK. YC W26 DEMO DAY LETS GOOOO
Brandon Gell
Mike Krieger x Every
Dan Shipper 📧: The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build
Sriram
I can wait on AGI, but can our agents get some money awareness... ASAP? 💸
Most agents are financially reckless by default. Add any complexity and you could one-shot your entire token budget. 🤯
Excited to co-host an evening in SF with @mastra_ai where we dig into the principles behind scaling radically efficient agents with some of the best minds in the arena.
On the agenda:
→ a workshop with @bookercodes 🛠️
→ a lightning talk ft. @swyx ⚡
→ a fireside chat with @NotionHQ's @sarahmsachs and our co-founder Rajaraman Santhanam 🔥
Followed by food, drinks, and conversations.
April 15 | Fort Mason | 5 PM | 100 spots
RSVP → http://luma.com/5nqllmul
Agarwal for Congress
It’s telling you’re most proud of “burnishing your credentials” instead of you know,
actually helping people in your district.
Ro Khanna: "Khanna couldn't ask for a more perfect opportunity to burnish his credentials than a fight with billionaires." https://www.sfchronicle.com/politics/article/ro-khanna-california-billionaire-22091811.php
Google Gemini
Longer tracks are here with Lyria 3 Pro in Gemini!
From experimenting with different styles to generating tracks with complex transitions, Lyria 3 Pro makes it easier to bring your full vision to life.
Rolling out today to Google AI Plus, Pro, and Ultra users.
Learn more 🧵
Tomorrow, the killer of Grandpa Vicha is expected to walk free. His pretrial detention already exceeded the max sentence of four years.
This is SF "justice."
https://gli.st/yamkccqy
Kelsey Piper
With return of algebra to San Francisco schools, Oakland Unified is now the only Bay Area school district to systematically deny all 8th grade students algebra instruction. Students have to take it externally if they want to take Geometry in 9th grade.
Katie Parrott
Really enjoyed this convo with @mikeyk an @danshipper
The idea that's staying with me: proof of thoughtfulness replacing proof of work. Did you make the decision, or did you accept the model's? Something every knowledge worker should think about.
https://www.youtube.com/watch?v=KRv9GpJYrUA
Sam Rose
I think this is the best post I've ever made.
ngrok: Quantization can make an LLM 4x smaller and 2x faster, with barely any quality loss. But what *is* it? @samwhoo crafted a beautiful interactive essay explaining it from first principles, aimed at coders, not mathematicians.
https://ngrok.com/blog/quantization
Adam Nelson
Stablecoins are emerging as the future of payments, but
imagine your bank account was public....
That’s basically how stablecoins work today.
@sidgandhi_xyz + @calummoore are fixing that with @payy_link , all the benefits of on chain payments with the benefits of institutional grade privacy.
Already live, global, real usage.
Proud to lead their $6M seed round alongside our friends at @robotventures and @6thManVentures
Evan Tana
A couple weeks ago @adityaag and I gave a talk at @southpkcommons that sparked a bunch of discussion in our community.
The core tension: The opportunity to build has never been greater. So is the risk of building the wrong thing.
The founders who win won't just build faster, they'll move smarter. Pick problems the Eye can't see.
Full piece out today 👇
Your work tools in Claude are now available on mobile.
Explore Figma designs, create Canva slides, check Amplitude dashboards, all from your phone.
Give it a try: http://claude.com/download
Andrew Curran
Periodic Labs update:
Rohan Pandey: Excited to finally announce what I've been up to since leaving OpenAI: autonomous science at @PeriodicLabs
Why? Because I Take the Bitter Lesson Seriously
To accelerate AI, we must enable it to hill-climb compute & energy through experimentally verifiable science 🧵
OpenAI Developers
Students: build something real in the Codex Creator Challenge, powered by @joinHandshake
Try new tools. Have fun. Break things. Repeat.
$10K in OpenAI API credits in prizes.
https://joinhandshake.com/students/codex-creator-challenge/?utm_source=openai&utm_medium=organicsocial&utm_campaign=FY26Q2-NWK-HS-Codex-Campaign
Auto mode is a step change improvement in the Claude Code UX, balancing autonomy and safety. Almost everyone on our team uses this as a daily driver.
Now available to Claude for Team users! `claude --enable-auto-mode` to turn on, then Shift + Tab to enter the mode
Claude: New in Claude Code: auto mode.
Instead of approving every file write and bash command, or skipping permissions entirely, auto mode lets Claude make permission decisions on your behalf.
Safeguards check each action before it runs.
People in tech are driven to create, and assume everyone else is too.
Most of my friends outside of tech don't want to create. They don't want to build their own personal software stack
And that's okay!
If you disagree, just don't be surprised when your product isn't adopted.
Ivan Burazin: OpenAI shutting down Sora is the most predictable outcome of misunderstanding consumer behavior.
Everyone grossly overestimates how creative people want to be.
99% of humans simply want to scroll and zone out instead of spending energy conceptualizing, editing, and creating
2036850081042211230
Bret Taylor
Today, Sierra is releasing Ghostwriter, our agent for building agents. With Ghostwriter, you can create an AI agent for your customer experience — one that can chat, pick up the phone, speak dozens of languages, take action on your systems of record, and be protected with industry-leading guardrails — simply by having a conversation. No clicking, no forms, no menus.
Codex and Claude Code have transformed how we build software, making it possible for software engineers to orchestrate and review the work rather than doing all the work themselves. We think the same transformation will happen for all software. Rather than every enterprise app having a web app for humans and an API for automation, every software platform’s UI will be an agent that can do the work on your behalf.
I recorded a demo of my building and optimizing an agent with Ghostwriter so you can see how powerful and easy it is to use. It’s completely changed the way our early adopters build agents, and it’s changed the way I think about the software industry. Let me know what you think, and, if you’re interested in trying it out at your business, please reach out directly.
配置一个好龙虾的本质是写好那几百上千字的identity.md soul.md agents.md等等文档
无论你是几个月满满迭代出来还是一口气半小时配置好效果都是一样的！
我针对我的工作流程配置了4个合伙人和一个助理。
ARC Prize
Announcing ARC-AGI-3
The only unsaturated agentic intelligence benchmark in the world
Humans score 100%, AI <1%
This human-AI gap demonstrates we do not yet have AGI
Most benchmarks test what models already know, ARC-AGI-3 tests how they learn
Matías N. Goldberg
Re What's next? A repurpose of QTangents, BFN (best fit normals), and octahedral mapping for LLMs?
François Chollet
ARC-AGI-3 is out now! We've designed the benchmark to evaluate agentic intelligence via interactive reasoning environments. Beating ARC-AGI-3 will be achieved when an AI system matches or exceeds human-level action efficiency on all environments, upon seeing them for the first time.
We've done extensive human testing that shows 100% of these environments are solvable by humans, upon first contact, with no prior training and no instructions.
Meanwhile, all frontier AI reasoning models do under 1% at this time.
2026-03-25 - Daily Briefing https://daily.reffwu.com/2026-03-25
In 5 to 7 years, UIs will be generated/streamed from the cloud, pixel-by-pixel. Phones/etc. will literally just be useless bricks with screens, speakers, and input.
That said, UIs won't be as dynamic as people expect. Imagine a new UI each time... that'd be so hard to use!
Sherry Jiang
we're stoked to have @dmsobol from @cerebras speaking at @aiDotEngineer in singapore!
daria designed mixture-of-experts (MoE) models, helping models like Qwen3 scale to trillions of parameters without scaling compute
@swyx @aimuggle @agrimsingh @unprofeshme @ivanleomk
Axios
Re Catch @replit CEO @amasad in conversation with Axios’ @danprimack at the #AxiosAISummit.
Tune in at 4:30pm ET
See full lineup & livestream info: http://trib.al/oj2VbGQ
Joseph Choi
talked to a YC founder who asked Sam Altman straight up "will OpenAI compete in my space / kill my startup"
the answer: behavioral health requires knowing if users are actually improving. OpenAI doesn't have that data, the wrapper startups do. v bullish for consumer AI
Joseph Choi
talked to a YC founder who asked Sam Altman straight up "will OpenAI compete in my space / kill my startup"
the answer: behavioral health requires knowing if users are actually improving. OpenAI doesn't have that data, the wrapper startups do. v bullish for consumer AI
Ali Rohde
http://x.com/i/article/2036572359474499584
Today's AI Builder Digest 🗞️
3 things worth your attention:
→ YC W26 broke records: 14% WoW growth and 3x more companies hitting $1M ARR than the last batch.
→ The "Pirate + Architect" model: Dan Shipper's proposal for the 2-person engineering team of the future.
→ Supply chain red alert: A 1-hour vulnerability in LiteLLM exposed millions to credential exfiltration.
The builders are moving fast.
Lee Edwards
The story here should actually be that 3 school board members voted against middle school algebra.
This is after a BoS resolution supporting middle school algebra, and support from the Mayor.
SF is still on the knife’s edge of clowns vs normal people. Don’t sleep on this.
Katie Miller: Twelve years ago, San Francisco killed 8th-grade Algebra in the name of equity. Advanced math enrollment (including AP Calculus) dropped, and racial gaps expanded.
China graduates ~1.3 million engineers yearly vs. just 130,000 in the U.S.
Last night, the SF school board finally
Victor
“The models today are good at adding features, they're not necessarily good at figuring out what to cut out of the product” — @mikeyk on 'AI & I' latest pod with @danshipper
A must watch if you build in AI consumer space.
Dan Shipper 📧: The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build
Real Estate Lawyer
Do you want to know why rent is so expensive in LA? Why the application process takes so long? Why landlords want so much info from you? Here is a recent story:
A family was referred to me for their eviction case. They were heading to a jury trial in one month and didn't have a lawyer, yet.
They did all the paperwork and filing themselves, in-house to save on legal fees. And surprisingly, they did a great job. They filed the paperwork with LAHD. Gave proper notice to the tenants.
I reviewed their paperwork and it was better quality than 90% of other eviction lawyers. I didn't see any viable way to dismiss the case on a technicality.
As long as they were properly represented in trial, the family was going to win the eviction. They did everything by the book. Followed all the local rules. Gave all the necessary notices.
The family told me the judge ordered the parties to mediate at the first court appearance. The family attended the mediation in court without a lawyer. The tenant was provided a lawyer by the city, for free. At the mediation the lawyer for the tenants offered this settlement:
1. 4 months to vacate the property
2. Cash to leave, paid upfront
3. Waiver of all owed rent
4. Sealed record
They rejected the offer, of course. Why would they accept this?
The family then asked me a great question. "What is our best case scenario with you in trial?"
Based on my review, I gave them my most realistic estimate of the best case scenario in trial:
1. Both parties announce ready at the next trial date (1 month away) and trial takes 3 days. We win the trial.
2. Sheriff locks out the tenant 75 days after the trial.
3. about 110 days to possession.
4. Gave them an estimate cost for fees/prep time.
5. No viable collection of back rent, tenants had no assets.
Obviously, this was the best case scenario. It could be worse. Trial can be delayed. While I was confident we are going to win, juries are unpredictable.
This is where we had a surreal moment of collective clarity. The settlement offer they rejected is basically their best case scenario if they win the trial.
This was not a coincidence. The attorney for the tenants asked for pretty much the same amount I quoted them for my fees.
The lawyer for the tenants knew the family had to hire a lawyer for a jury trial. The lawyer knows it takes the sheriff 2-3 months to lockout after a judgment. The lawyer knows it's hard (and expensive) to collect against tenants with no assets.
State and local government created a system in which cases take forever to litigate, eviction laws are extremely complex and technical, easy to dismiss cases, only one side has to pay a lawyer, and worst of all, possession enforcement takes 60-90 days instead of 5. And it's all getting worse.
The leverage for the tenants is systemic. It's by design. Why would the tenants make any other offer? The landlords are left with no real options but a shitty settlement.
There are no real choice. Even when you do everything right, you still lose.
Tenants don't pay rent during evictions. They had no viable way to win the trial. There were no habitability issues. The landlords posted all the notices. Never raised the rent. Didn't retaliate. The landlords did everything right. And the tenants still win.
The mother looked at me and asked "our base case in trial is the same as the shitty settlement offer? Are you telling me we should have taken the offer we rejected?"
I didn't know how to respond.
Alfred Lin
Everyone will adopt AI. Everyone will build more, ship faster, and deploy agents. That is table stakes. The founders who win will be the ones who make better choices, which requires judgment. (And if you're in the bottom left quadrant, get out quick.)
Alfred Lin: http://x.com/i/article/2036193814445203456
as agents make it easy to add features, design matters more, not less. the role is no longer just pushing pixels – it’s deciding what should exist, how it fits together, how humans stay in control, and how intelligence feels clear, trustworthy, and useful.
taste, craft, and judgment have always been the bottleneck. the game is not who ships fastest, but who makes the right thing for humans.
Lenny Rachitsky: I don’t know exactly what’s going on here, but it does feel AI-related. Unlike PM and eng, which started growing in 2024 (two years post-ChatGPT), design didn’t. If I had to venture a theory, I’d say that because AI is allowing engineers to move so quickly, there’s less
Mike Krieger
Always great to chat with @danshipper ! Love what he and the team at @every are cooking up, we could have talked for hours.
Dan Shipper 📧: The rules of professional product development are being rewritten in real time.
- PMs and designers can ship software as easily as engineers.
- Software is no longer just built for humans—it’s also built for agents as first-class citizens.
To better understand how we build
hkey
Re @justalexoki not yet
The Neuron
An AI editor goes viral… then crashes. 😬
@danshipper joins us LIVE to share the real story behind Proof, how Every builds and maintains multiple AI products with a small team, and why agent-native engineering is changing software.
Catch us here on Thursday!
#AI #ArtificialIntelligence #BuildInPublic #Founders
Olivia Moore
We are unprepared for how quickly the world is going to shift from marketing to people -> marketing to AI agents
If you were a technical co-founder or similar in a past life and you want to PM the future of software development come join us.
Nan Yu: @lennysan We are hiring for PM and Eng at @linear
https://linear.app/careers
PMs. We are especially interested if you were a technical co-founder or played a similar role before. We need someone with a high degree of agency who is comfortable mixing it up with top-tier design and eng.
Ivan Leo
Life update I’ve moved to San Francisco and joined @GoogleDeepMind.
Excited to work alongside @OfficialLoganK @thorwebdev @DynamicWebPaige @_philschmid @patloeber @ammaar @vadiamit @osanseviero @harrisonfjobe @goodside @alihcevik @matthewridenour
Gradium
Gradium TTS is faster than ElevenLabs Turbo and GPT-4o Mini, hitting 214ms P50 with multiplexing. This makes it a perfect choice for real-time interaction.
We show how to measure Time-to-First-Audio correctly and cut latency across the stack.
Blog post link in the comments.
Michele Catasta
The Buy vs Build dilemma is no more.
Always start by building.
etn.: President of @Replit Michele Catasta (@pirroh) explains why the decreasing cost of building tools has lead to several key shifts across start-ups:
"Before you had to make the choice of shelling out tens of thousands of dollars, spending several months creating the first
Axios
.@replit CEO @amasad on vibe coding:
"We started our company even 10 years ago before vibe coding, with the express mission of creating a billion new software developers. Now, with the rise of AI, what happened is AI turned out to be really good at coding." #AxiosAISummit
tomorrow's gonna be a big day @every!! 👀
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
Amazing progress in San Francisco
Garrett Langley: Huge results from @SFPD: From 2023 to 2025, major crimes fell 44%, larceny halved, car thefts fell 54%, and burglaries fell 33%. Step change improvement for the people of San Francisco in only a few years.
Noah Smith 🐇🇺🇸🇺🇦🇹🇼
Americans really, really do not like progressive ideology.
They liked liberal ideology. When liberalism gave way to progressivism, Americans started to walk away.
The Missing Data Depot: New YouGov report on American partisanship between 2007 and 2025 shows that with the one exception of college educated white men, every major demographic group in the US has become more Republican on balance over the last 20 years.
Every 📧
This week's Context Window: @Instagram's cofounder on building agent-native products, the skill that matters more than coding ability, and why you only need two engineers to ship
– 🎧 @danshipper sits down with @mikeyk on why you should throw out your product every three to six months
– @lauraentis joins @every as staff writer and reports on the persistence-over-technical-skills debate
– @tedescau, a self-described "tech doofus," built a data-pulling agent the whole team uses
– @naveennaidu_m on why vibe-coded apps break—and when to start from scratch
– @danshipper's pirate-architect framework: two roles, one breakout product
– 🗓️ Join us Friday for @every's Q2 Demo Day—live walkthrough of Plus One, our AI agent solution that lives in Slack
Melissa Chen
This is one of the most important threads that explains what has been eating at you over the last few years.
Ever since the great Precambrian Woke Explosion of 2020, it no longer mattered what the specific issue was whether it was climate change, BLM, trans rights, ANTIFA, Free Palestine, or feminism. All left-wing activist causes suddenly felt deeply interconnected and totalizing in very weird and strange ways.
If you are a sane person, you'd have asked:
> how come feminists didn't care about the Iranian women subjugated and slaughtered by the Mullahs for wanting more freedom?
> how come UN Women cared more about making sure trans women (men) have access to female prisons than Hamas terrorists raping Israeli women?
> how come Greta "How Dare You" Thunberg smoothly pivoted to campaigning against Israeli "apartheid" while never once mentioning China destroying marine ecosystems with overfishing?
> what the fuck EVEN is Queers for Palestine?
> why the hell was Microsoft doing land acknowledgments before Zoom town halls?
Everything from American corporate life to Hollywood movies to universities became vehicles for the omnicause: a single, all-encompassing moral framework where climate, race, gender, borders, and foreign conflicts are fused. The surface “issue” often became secondary to maintaining ideological purity across the entire bundle.
What was driving this?
I KNEW where it was coming from. Now, @DataRepublican and @AsraNomani have the receipts.
David Horowitz used to say that "the issue is never the issue. The issue is always the revolution."
The revolution here is the overthrow of the ideological architecture of the West and the remaking of the world order.
China's leadership explicitly wants to undermine American confidence of itself and of the US-led World Order because Beijing is a revisionist power that sees the system as designed to keep it subordinate.
And so it amplifies narratives that portray the US-led order as the "root of evil" - imperialist, morally bankrupt, white supremacist, exploitative, and in inevitable decline. The aim is cognitive warfare to erode domestic support for protecting this order which serves to facilitate China's seamless rise.
Neville Roy Singham has funneled at least $278 million through dark-money intermediaries, donor-advised funds, shell companies, and US nonprofits into a vast network of far-left activist groups, media outlets, and political education organizations to push this.
@DataRepublican directly quotes Singham at a 2024 CCP forum in Shanghai:
"If we want to have a new world order based on multilateralism that President Xi and CPC and China have proposed, we have to undo the ideological damage that has been done by the narrative of World War II."
Read that again.
Understand that Singham sees this WWII narrative as the last major cultural obstacle standing between the current order and the “multipolar” (China-led) future his network is working toward.
And now... notice all those around us who are parroting the same line.
It used to be just the crazy lefties and the NGOs like the UN. Or the globalists like the WEF.
Now you're seeing it on the right.
Beijing wins when a majority in the West treats US hegemony and the post-1945 system as invalid, so that it can replace it with one where it sets the rules and norms.
DataRepublican (small r): 🧵 THREAD: "House of Singham" : Neville Singham's mega-exposure.
This is part 1 out of 5 in a mega-project exposing Neville Singham and his money flows. Honored to have been friends with @AsraNomani throughout this.
I'm going to explain this article below - but you should
I'm generally quite low key when it comes to events. But when @andruyeung comes to town, you ball out..
Excited to cohost a VERY fancy founder dinner on April 7th in SF.
We already have some fantastic founders coming & would love to add a few more.
RSVP link in the next post!
Re Priority given to the founders with the funniest response.
Link to RSVP here: https://luma.com/juntodinnerapr7
They call it an overpaid CEO tax but actually it’s a tax on your groceries and retail
It’s bad policy wrapped up in anti-billionaire bullshit
Give us a break with this stuff please. We need normal people to speak out now. Help us 👇
Blueprint: 🚨Happening tonight!
If you care about SF's economic recovery, come out and speak out against Prop D.
https://www.eventbrite.com/e/1982900019336
Alfred Lin
.@garrytan is giving a masterclass in AI on X
- 6,000,000 LOC/year run rate
- 3 weeks (vs 2 years) to rebuild his old company Posterous
- managing 20 workers at a time
- 47,000+ stars for gstack on GitHub (https://github.com/garrytan/gstack)
- while being CEO of @ycombinator
Shifting my wakeup time from 4am to 3am so I can keep up with him
--
https://x.com/garrytan/status/2033306615966060734
https://x.com/garrytan/status/2019898808017252724
https://x.com/garrytan/status/2036626825125011659
