Kim-Mai Cutler
Guess who fixed the loophole?
Marc Andreessen 🇺🇸: How bad is it? Until just last year, in California, “proving theft from a vehicle required proving that all the doors were locked”. https://www.luecklaw.com/california-closes-locked-door-loophole-for-theft-from-a-vehicle/
gbrain.io
Gbrain is always getting better.
Garry Tan: Early review of GBrain - I am working on it day-to-day, dropped more than 20 PRs in the last week, so it is a young codebase growing quickly and changing. This is great feedback.
GBrain getting better every day
https://vectorize.io/articles/gbrain-review
GPT-5.5 Pro for fact checking:
Ethan Mollick: GPT-5.5 Pro is a very solid fact checker. I can throw entire chapters at it and it will hunt down every key reference accurately. The only real annoyance is that it loves nuance, so returns a lot of “the general idea is right, but you are not taking into account tiny detail X”
How do you build a great brand? Build a great product
Elon Musk: Product >> Brand
I'm having the time of my life frankly
Harj Taggar: Building agents has the same emotional arc as programming. I start every project thinking it should be easy to get what I want, then end up deep in retrieval quality, context engineering, and cross-modal eval loops before anything actually works.
When all your friends are building on the edge
Tom Blomfield: Went to a beach house for Memorial Day and everyone is inside on Claude Code 🤦♂️
High agency high taste is the unlock these days
Nicolas Dessaigne: So many people are stuck doing boring work because they don't have the right background or aren't in the right community.
But now? Anyone can use ChatGPT for free. And it doesn't cost that much to access the superpowers of Claude Code and Codex. These tools giving people
Aaron Epstein
Only 3 things founders should focus on during the YC batch:
1) build your product
2) talk to users
3) exercise, or whatever you need to do to stay sane doing the first 2 things
Ines Boutemadja: YC partners: these are the 3 highest leverage months of your life, focus on the right things
Me on Wednesday nights: 5 6 7 8
This is exactly how I build my agents and I’ve built 4 back to back now: both personal AI and company brains
It’s the same methods over and over. Do it. Skillify it. Add to cron. Check if it is resolvable. Evals and integration tests. Repeat.
Vox: >started treating my Openclaw/Hermes like patients.
>the bugs that kept coming back all had real medical names. take the one everyone calls "AI hallucination." medically it's confabulation.
>now i scan for the organ that failed. gbrain handles memory. OpenClaw approvals handle
Basile Terver
📢 Accepted to TMLR, with reproducibility certification 🏅
v2 of our JEPA-WM study (arXiv:2512.24497) is out, with new data-scaling experiments, a Lipschitz analysis of multistep rollout training, and extended discussions.
Recap + what's new 👇
w/ @JimmyTYYang1, Jean Ponce, @AdrienBardes, @ylecun
Basile Terver: My first PhD paper is out! 🎓
"What Drives Success in Physical Planning with Joint-Embedding Predictive World Models?"
tl:dr: JEPA-WMs for robotics: learn dynamics on top of visual encoders, optimize actions towards goal 👇
w/ @JimmyTYYang1, Jean Ponce, @AdrienBardes, @ylecun
Ankit Gupta
v0.12.0 is out! Some notable features include:
- user requests like blocklists, QoL improvements with multiple accounts
- for developers: a major testing overhaul, featuring an LLM-as-judge for AI features and an agentic verification loop to automatically test features on every PR (quite fun to build)
coming next:
- big performance jump
- open source models
Ankit Gupta: Fun update: I got tired of disliking every email client I’ve ever used and built my own. It’s called Exo (for exoskeleton). It’s Claude Code for my inbox. It manages my inbox for me, and it’s open source. Link to repo + some notable features in thread!
Eric Topol
The ramp up of cancer immunotherapy is remarkable. Now we're seeing vaccines achieve some cures or remissions in the most refractory cancers: pancreatic, melanoma, glioblastoma, renal, triple-negative breast cancer.
✓ out the new Ground Truths (link in profile)
Yann LeCun
Re Major difference in my mind:
- an engineer, given a problem, invents and tries multiple solutions and stops when the solution is good enough. The goal is product innovation and shipping.
- a scientist asks new questions, proposes various new solutions, compares them (sometimes with old ones), and writes about it. The methodology must be sound or else peers will sneer. The goal is scientific breakthroughs and technological progress.
Both can be called "researchers". Many people can do both: these are activities, not identities.
Importantly, most product innovations are built on scientific breakthroughs and technological innovations that happened 2, 5, 10, or 20 years earlier.
Sometimes you have to build smaller first to be able to build bigger later. Viable fusion energy is so important in the age of intelligence. Energy becomes almost the only bottleneck to abundance in that world of superintelligence.
https://www.geekwire.com/2026/helion-makes-big-bet-on-tiny-merge-fusion-testbed-to-meet-aggressive-microsoft-timeline/
Harj Taggar
I don’t think local models will ever be the default but every device will need one as an emergency generator to stop people freaking out when they lose connectivity and have to make a decision.
Matt Van Horn
Me: I wonder what the top trending repos on GitHub are today.
Me: Of course, gstack by @garrytan
Me: Omg, why is my face there as the 3rd "built by" contributor to GStack.
Thanks for letting me contribute @garrytan !
Forrest Liu
Bravery = sacrificing everything to come here for the chance to build something people want.
Cowardice = making your fortune here, trying out politics in DC, failing, then coming back here to try and buy an election by supporting a ballot measure that would stifle other founders
Garry Tan: We should fight the people who want to stop startups from staying in San Francisco. We want more startups and more prosperity.
I'm fighting back through Garry's List Action, I'm in for $50K, I need 9 more people to match me. Are you in?
https://secure.actblue.com/donate/stopsfwealthtax
Don’t get “this is only one” syndrome about your first customers
Tom Blomfield: Counterintuitively, I see some startups obsessing *too much* over their first customer.
Big enterprise customer ghosting you? Maybe they’re just not a good fit. Go get 10 more.
Top-of-funnel solves most problems.
Tom Blomfield
A founder I spoke to last week spent 3 years learning everything about their industry.
Now they want to pivot to “AI for Dentists”.
The grass only looks greener because you’ve not trampled through the shit yet.
👀
OwlyPost: @garrytan
Ankit Gupta
deepseek just permanently priced their frontier model at 1/30th of american labs
anyone know what the hardware story here is? is this huawei chips driving lower costs or model optimizations or lower margins?
arman
it's crazy how @greptile has had such a noticeable improvement in the last few months. i've never seen an agent at that scale improve drastically so fast
Replit ⠕
Make your next Replit purchase more rewarding with Amex Offers®. Eligible American Express® Card Members can enroll and earn a $50 statement credit after they enroll and spend a minimum of $100 in one or more qualifying purchases on their enrolled Card at http://replit.com by 8/6/2026. Limit of 3 statement credits (total of $150).
Learn more and enroll at http://go.amex/3XF78L. Enrollment required and may be limited. Terms Apply.
I haven't written a personal blog post in over 5 years so if you see posts that claim to be written by me, they're not. I'll update if this ever changes. Maybe it should.
Kathryn Wu
I think one reason YC likes logical engineers is not just because they can code.
A lot of them are unusually clear communicators.
Coding trains you to think in strict logical sequences:
input → output, cause → effect, constraint → solution.
You can hear it immediately in good founders.
Not necessarily charismatic, but coherent.
People underestimate how much startup momentum comes from simply being easy to understand.
Qwen2.5‑7B Instruct apparently is GPT-3.5-turbo level
Harj Taggar: I don’t think local models will ever be the default but every device will need one as an emergency generator to stop people freaking out when they lose connectivity and have to make a decision.
Memorial Day is a reminder that freedom isn't the default state of the world. People fought and died for it.
I am thankful America is where you can build, create, and push the frontier forward.
Thank you to the men and women who served and gave their lives. 🇺🇸
Look guys, LinkedIn is many things, but it is a place where you can find real customers
George Jefferson: At YC we got the advice of posting on LinkedIn
At first ngl I thought it was bullshit because I’ve never liked or engaged with the platform before
But holy shit, Ive tried to post x5 a week and its mostly been slop like this, but it’s pure gold for driving inbound & sales
Very grateful to @david__booth for inviting me to Tahoe.
Got to meet @nateliason in person, see how @clairevo teaches her kids, and just meet a great group of folks.
Got a nice kayaking trip in too with my daughter.
What used to feel like procrastination (building systems instead of the MVP) is now a prerequisite to ship effectively with AI agents.
My number 1 lesson from @ryancarson:
"We used to say just do the bare minimum to get the MVP out. Don't spend time on systems. It's literally reversed now.
You have to spend a lot of time setting up your documentation. Build all that into a cron job with a skill file, and suddenly you're doing the work of 10 people."
📌 I asked Ryan how he ships 10 PRs a day, here's his answer: https://www.youtube.com/watch?v=IDqdVZwAwjw&t=1713s
Peter Yang: "We used to say build the MVP. Now you should build the system that builds the MVP first."
Here's my new episode with @ryancarson where he shared how he runs his startup solo with AI agents:
✅ OpenClaw as his AI chief of staff to triage emails, book meetings, and do sales
Peter Yang
What used to feel like procrastination (building systems instead of the MVP) is now a prerequisite to ship effectively with AI agents.
My number 1 lesson from @ryancarson:
"We used to say just do the bare minimum to get the MVP out. Don't spend time on systems. It's literally reversed now.
You have to spend a lot of time setting up your documentation. Build all that into a cron job with a skill file, and suddenly you're doing the work of 10 people."
📌 I asked Ryan how he ships 10 PRs a day, here's his answer: https://www.youtube.com/watch?v=IDqdVZwAwjw&t=1713s
Peter Yang: "We used to say build the MVP. Now you should build the system that builds the MVP first."
Here's my new episode with @ryancarson where he shared how he runs his startup solo with AI agents:
✅ OpenClaw as his AI chief of staff to triage emails, book meetings, and do sales
Codex for finding space on your laptop:
BOOTOSHI 👑: i had codex audit my entire macbook to see how much space we can save and it's found 500 GB to save, AWESOME
prompt was: "do a FULL read only analysis on my Macbook to help me optimize storage"
note: why tf is there a codex-tui.log file that is 116gb ??????? WHAT ????
Nicolas Dessaigne
Today is the last day to apply to @ycombinator Summer 2026.
Applying to YC changed the trajectory of Algolia, and my life. We were two French founders with a search API and a dream. YC gave us the network, the clarity, and the urgency to build something real.
If you’re on the fence: just apply. The worst that can happen is you don't get in. The best? Everything changes.
👉 http://ycombinator.com/apply
