Alex Wettig retweeted
Can LMs discover & fix bugs if you don't tell them what went wrong? To proactively maintain a repo, agents need to find problems before users do.
SWE-sweep benchmarks this on 100 repos, 22 languages, 4k real bugs (numpy, php interpreter, lean kernel). Top models get <5%.
Alex Wettig retweeted
Chris Arnade's blog continues to be one of the very best: walkingtheworld.substack.com. (You can also follow him: @chris_arnade.)
Alex Wettig retweeted
We're making Git hosting more reliable, performant, and scalable.
This post traces 20 years of Git infrastructure and explains how that history led us to design and operate our Git storage, Origin, as if it were a database.
cursor.com/blog/git-at-any-s…
Alex Wettig retweeted
I'm so proud of the team. Building Cursor has been the highlight of my career.
10 months ago we shipped Composer 1, trained on 864 B200s. Today we're serving tens of billions of TPM on models trained on hundreds of thousands of chips. Onwards!
nitter.cf/cursor_ai/status/20882…
Cursor is now part of @SpaceX.
Today, we have officially closed our acquisition. We will join the @spacexai team to help make Grok the world's most useful AI and improve Grok Build, Grok Bot, Grok API, Cursor, and more.
SpaceX has built some of the most inspiring and impressive technology in the world, and we’re grateful for the opportunity to become part of such a special company. Onwards.
Alex Wettig retweeted
(1/N) 🧵 Chinchilla assumes you'll never run out of fresh data. That era is ending!
Compute keeps growing exponentially, but high-quality tokens don't. So what's the exchange rate between extra compute and fresh, high-quality data?
We propose Compute-Data (CD) scaling laws, which measure the exchange rate between extra compute and fresh data.
it's crazy that when i tell an agent to "run XYZ", i basically expect it to read 20 files every time just to figure out what the hell XYZ is. it sounds like engram will be big for model welfare
Alex Wettig retweeted
hi i know that guy
At the AI Industry Application Summit, Tencent's chief AI scientist @ShunyuYao12 on where agents are headed:
"I thought it'd change everything in 5–10 years. It came faster — and maybe I was still thinking too small." ⚡
Alex Wettig retweeted
Recently met @srush_nlp and he started giving me an impromptu lecture on how targeted on-policy self-distillation works.
I asked him if I could record it on my iPhone.
The basic idea is this: if the model made a mistake at some point in the rollout (for example, calling a tool that doesn't exist), we want to discourage this specific error, but we don't want to just learn from the final reward, because it's a very noisy signal spread out over the whole trajectory.
So we have another model read this trajectory and figure where the error was made. It simply inserts some hint tokens to the part of the trajectory right above where the mistake was made.
Now with these injected hint tokens, have the model run a forward pass. You're not having to regenerate a new rollout - aka no new decode required.
The hint causes the model to assign lower probabilities to the error tokens. You then trains the original model to match these new probabilities, teaching it to downweight that specific mistake.
Alex Wettig retweeted
Introducing Composer 2.5, our most powerful model yet.
It's more intelligent, better at sustained work on long-running tasks, and more reliable at following complex instructions.
For the next week, we’re doubling the included usage of the model.
Alex Wettig retweeted
Glass / Agents Window is the big story of Cursor 3 but there's lots of improvements we made to the standard IDE, its not going anywhere.
My favorite is using standard vscode tabs for our chats so you can easily pull them out side by side and use all the standard VSCode shortcuts to manage them. It makes managing parallel agents a lot easier.
