This doesn't just happen - it takes relentless execution and hard work. Day in and day out, the entire bay area has seen and felt their unparalleled level of focus compounding into outlier progress. Beyond proud of Brendan (can/do) adarsh surya and the entire Mercor

(1/7) Inspired by DeepSeek's FlashMLA, we're releasing ThunderMLA—a fused megakernel optimized for variable-prompt decoding! ⚡️🐱ThunderMLA is up to 35% faster than FlashMLA and just 400 LoC. Blog: bit.ly/4kubAAK With Aaryan Singhal, Dan Fu, and @hazyresearch!

if you are a design engineer and free this weekend dm me - you won’t regret it

Paradigm is one of the most amazing AI tools I’ve seen. MS Excel where each cell is its own context-aware LLM

sometimes you just gotta go into the desert and code

fastest way to do investment research 🕵️‍♂️

It’s official – I'm excited to announce Superpower's $30 million series A from Forerunner and Day One, alongside celebrities Steve Aoki, actress Vanessa Hudgens, NBA star Giannis Antetokounmpo, and existing investors Cyan Bannister, Shaan Puri, Balaji, Arielle Zuckerberg and many

Introducing Phia — your new AI shopping companion. Phia instantly scans 40,000+ sites for deals on the best new & secondhand fashion. No more endless tabs or scrolling. Just smarter shopping, with fashion’s best prices in one click. Download the Phia shopping tool on the iOS

what decision would you make as the most ambitious version of yourself?

What we've shipped over the past few months... January - Pull live data with webhooks - Group sheet sharing - Template ecosystem - Paradigm home - Chat dataset integrations - Preview grid data February - Onboarding flow - Product docs - Retries & cancellation - Design &

design engineering can be a nebulous role... so, mathurah Anna Monaco & i are hosting a rapid fire ppt night — 3 min to share your path as a design eng! we're keeping this one small & cozy ‧₊˚ ⋅ 📍 SF ⏰ 8-10pm 📆 wed may 28, 2025

(1/5) We’ve never enjoyed watching people chop Llamas into tiny pieces. So, we’re excited to be releasing our Low-Latency-Llama Megakernel! We run the whole forward pass in single kernel. Megakernels are faster & more humane. Here’s how to treat your Llamas ethically: (Joint

was awesome to hear everyones presentations on design engineering and what it means to you thanks to floguo mathurah for co hosting⚡️

Today, we’re announcing Antimetal’s $20M Series A. Writing code is no longer the hard part. Maintaining it is. We’re automating everything that happens after you deploy. We have 60 spots available. Reply for early access.