Jay Hack
@mathemagic1an
Building @codegen. Tweets about AI, computing and dev tools. Previously did startups, @palantir. Not a pseudonym.
Multi-agent flows in Linear are so powerful. You can take on 5-10x more ambitious tasks. I just had it build out our marketing site in 20 minutes. I will likely make minor tweaks in Cursor and ship. This was unfathomable a year ago
The new 10x engineer is a forward deployed AI engineer embedded by the humans at @codegen 1 day in person + >10 min human slack response (>1 second codegen response) changed the trajectory of our $5m ARR saas company Thank you @mathemagic1an and team!
Forward-deployed focus at @codegen is a superpower We will literally show up at your office and help you modernize your AI dev practices - just ask @MaxPGreenwald and @warmlyai Ping via DM if you're interested - accepting applications now 🚀
Forward-deployed focus at @codegen is a superpower We will literally show up at your office and help you modernize your AI dev practices - just ask @MaxPGreenwald and @warmlyai Ping via DM if you're interested - accepting applications now 🚀
We don't just sell software - we forward deploy to your office and turn your team into a feature-shipping machine ⚡️ When a founder at @warmlyai cried for help on LinkedIn, we parachuted in, set them up with code agents, and shipped 30+ features to production ✅ Learn more 👇
In defense of OpenAI, sounds like this was a coordination failure and they acted in good faith. Legitimately impressive what both teams have accomplished and I hope this doesn’t overshadow their results x.com/polynoamial/st…
Tough look for OpenAI They've pissed off the international math community by jumping the gun, meanwhile @GoogleDeepMind has an officially-confirmed result that will be available commercially months earlier
Tough look for OpenAI They've pissed off the international math community by jumping the gun, meanwhile @GoogleDeepMind has an officially-confirmed result that will be available commercially months earlier
Our IMO gold model is not just an "experimental reasoning" model. It is way more general purpose than anyone would have expected. This general deep think model is going to be shipped so stay tuned! 🔥
"No, and tell Claude what to do differently" This will be one of the most powerful flywheels in the code agents space. A clear signal to iterate on.
Stacked PRs are a force multiplier for human<>agent collaboration, and Graphite's MCP gives agents everything they need to start stacking.
Take on bigger tasks with your agents. GT MCP trains your agents to create smaller sequenced PRs, so you can easily follow what changes were made and get more done.
Kimi K2: "Dreaming of MCPs"
Next section is on Post Training. They first generate large amounts of synthetic tool use samples. To get the tools they require, they fetch 3000+ real MCPs from Github, categorize them and then evolve each of the categories to generate more synthetic tools in that category. This…
A better interface for agents Built with @codegen 👏
I built the LLM interface I wished existed. interface0 is for power users: cross-provider memory & chat, multi-model synthesis, sharing & team features, granular context engineering, template prompts, forking, prompt enhancing, & more Here's a demo—see 🧵 for the rest
Interestingly, both OpenAI and Gemini achieve the exact same score: 35/42 - and both teams solved problems 1-5 but did not solve 6, the most challenging problem
This is rapidly becoming the consensus The best agent implementation is “whatever Anthropic decides to RL on” Code agent products effectively wrap a UI and integrations on it. Doesn’t stop you from getting to $500mm ARR
I use CC from Cursor and I assumed most do as well (?). I end up with a mixed thing where Cursor is the UI layer for reading the code, manual edits, tab completion and chunk edits, and CC for larger changes, architecting, Q&A. Still rapidly evolving though...
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
