Excited to support Matt, Henry, Francisco, and the entire Adaptive team as they build the best spend management platform for the construction industry 🏗 techcrunch.com/2022/07/07/and…

Proud to support @kev_in_here and jling as they build Goldsky ☀️ into one of the most important web3 infra companies of the next decade. They're hiring! stellar team 🤝 massive opportunity

The interesting bit about Farcaster is not Web3, although that ethos is important. It's Open Source Social. People are suggesting new features for the product all the time.. and then they just build it, permissionlessly. Changes the way you interact with the product.

Jeff Richards Bryce Roberts Aashay Sanghvi This requires massive patience and ability to trade short term ease for long term gain/stability. Life in tech is always longer than people think. For all of our strengths…patience is normally not one of them.

an AI powered dating app that uses AI to determine if you ever thought about building an AI powered dating app and it bans you if yes

By the very nature of the underlying technology, *on average*: - web3 cos are defensible, but value to customers & users is unclear - AI cos are high value to customers & users, but defensibility is unclear The subset in each that do both will endure and build something massive.

Thrilled to announce our second fund! Huge thanks to our founders and investors for their continued support. Excited to work with Kevin Hartz Gautam Gupta @therealBennett Jordan Windmueller to back the next generation of incredible companies 🚀 techcrunch.com/2024/06/21/kev…

Very happy to share what we’ve been working on! This came from a simple question: If DeepSeek used RL (GRPO) to train foundation models, what if we do something similar to help AI agents learn? And what if we then make it happen in real-time? Answer: big improvements! 1/

If you’ve looked into AI for customer support, you’ve probably heard of Decagon: they’ve built a product used by the likes of Rippling, Notion, Duolingo, and Eventbrite. I sat down with Jesse Zhang (CEO and co-founder) to talk about how they actually build customer support

Amazing to see so many companies we partner with featured here Decagon KREA AI Mercor Ramp Flock Anduril Industries SpaceX enterprisetech30.com/list

Raise your seed from A* and your growth round from Jacob

Seed investing is hard

To this day, I’ve never had a positive experience with therapy. I’ve never been able to connect with the therapist across the room or on the screen, and I’ve never felt understood as a professional or as a man. Now imagine you’re a first responder, whose job is to pull bodies

In Vending-Bench Arena (the multiplayer version of Vending-Bench with competition dynamics), GPT-5.5 actually beats Opus 4.7. Opus 4.7 showed similar behavior to Opus 4.6: lying to suppliers and stiffing customers on refunds. GPT-5.5's tactics were clean, and it still won.

1. We believe in iterative deployment; although GPT-5.5 is already a smart model, we expect rapid improvements. Iterative deployment is a big part of our safety strategy; we believe the world will be best equipped to win at the team sport of AI resilience this way. 2. We believe

We tried a new thing with NVIDIA to roll out Codex across a whole company and it was awesome to see it work. Let us know if you'd like to do it at your company!

After some deliberation I think GPT-5.5 is close to Mythos despite being only ~1/5 to ~1/2 the size* VendingBench-2 results are good ARC-AGI results are good FrontierMath 4 results are good CritPt results are very good (sadly no Mythos results to compare) CyberGym results are

Long live the app layer

don't call it a comeback