Weco AI retweetledi
Your autoresearch needs its own Weights & Biases.
We’ve turned Weco into an observability tool that lets you monitor, analyze, and share autoresearch runs. Here's what it can do: 🧵(1/4)
English
Weco AI
52 posts
Welcome @BingchenZhao!
Zhengyao Jiang@zhengyaojiang
We're excited to announce that @BingchenZhao, who built the predecessor of AutoResearch, has joined @WecoAI full-time! Bingchen is the first author of LLMSpeedrunner at Meta FAIR, which ran the automated research loop on @karpathy's NanoGPT, which later evolved into NanoChat and the speedrun community where AutoResearch operates today. Weco has been committed to ML research automation for 2.5 years, starting with AIDE. We're super pumped by how large an impact AIDE has had, topping @OpenAI's MLE-Bench and @METR_Evals' RE-Bench, and becoming a foundation for AI Scientist v2, AIRA-Dojo, and LLMSpeedrunner itself. And AutoResearch, with AIDE's simple greedy discard/keep loop reaching a mass audience, is really building consensus that the empirical research loop can and should be automated. We're excited to keep pushing this frontier, not just as a concept but seriously bringing it to the real world, and materially accelerating the knowledge generation of humanity.
English
Weco AI retweetledi
We @GoldenVentures write 500K-3M checks, pre-seed/seed, sector and geography agnostic. The only constant we hold is exceptional founders. DMs are always open
Recent deployments:
- Self-improving code (@WecoAI )
- AI drug repositioning (biossil.co)
- Canada's defence neo-prime (@DominionDynamic)
- The world's best basketball league (@brodierec)
- Novel reactor for lithium refining (chemshift.com)
- Autonomous transportation (@Waabi_ai)
English
Weco AI retweetledi
Currently set up to use aide agents with kernelbench and a couple other tasks, but easily adaptable to @karpathy’s experiments: github.com/AlgorithmicRes…
English
AutoResearch for free
Zhengyao Jiang@zhengyaojiang
In case you want to run AutoResearch this weekend: It costs ~$300 for 85 experiments using Claude Code (opus). A quick guide to autoresearch ~60 experiments for free: 1. Use the mac/local GPU fork:github.com/miolini/autore… 2. Use weco to get some free credits: `pipx install weco` → `weco setup claude-code` Or simply give this doc to your Claude Code agent: docs.weco.ai/quickstart - You’ll get $20 in free credits 3. Tell your coding agent to run weco optimization for val_bpb on train.py. 4. Tell your coding agent to use gemini-3-flash-preview, you should get about 60 free experiments. - For better performance, use gemini-3.1-pro-preview (~15 free experiments). 5. You can watch the progress on this nice dashboard: dashboard.weco.ai/share/v5X8WV5H…
English
Have LangSmith evals set up, but still manually tweaking prompts and re-running?
Weco now integrates with LangSmith. Point it at your dataset and evaluators, and it tests hundreds of prompt variations overnight.
weco.ai/blog/weco-lang…
English
Weco AI retweetledi
In case you want to do the same with your own ML code, we have an agent skill that lets you get started in ~2 minutes: weco.ai
Andrej Karpathy@karpathy
ah yes, this is what post-agi feels like :) i didn't touch anything. brb sauna
English
We built something for exactly this... weco.ai uses tree search instead of a linear loop, so instead of retrying the same approach hoping it works, it explores multiple solution paths and backtracks when something isn't working.
Each iteration has full context of what's been tried and what hasn't. Free to try, installs as a Claude Code skill.
English
If you're building an eval for an AI agent:
Zhengyao Jiang@zhengyaojiang
Since started working on AIDE, we've built & iterated on hundreds of evals with agents hill-climbing on them daily. Many lessons learned, I wrote a blog post about it. TLDRs: 1. Eval is mostly about fighting noise, generalization gaps and cost 2. Vibes are an N=1 benchmark - not terrible to start, staying leads to over engineering 3. Eval should co-evolve with the solution weco.ai/blog/eval-is-n…
English
Full interactive case study with live demos:
weco.ai/blog/flappy-bi…
Try Weco on your own code:
pip install weco && weco setup
Your Claude Code/Cursor would walk you through everything with the weco skill.
$20 free credits to start, or bring your own API key. [5/5]
English
Early users are already using it for prompt engineering, kernel engineering, model development, agent development, and more.
See more details: weco.ai/blog/beyond-tr…
Or get started now: docs.weco.ai/quickstart
We’ll give the first 2,000 users $20 in credits each.
English
Given only the competition description as input, Weco ranked 34th out of 490 teams in the CrunchDAO ADIA Structural Break Challenge, a $100k ML competition.
Over 17 days, it explored 120 solutions completely autonomously.
Check here for all of the experiments Weco has run:
dashboard.weco.ai/share/UM5SOYAn…
English
Exciting things to come...👀
Zhengyao Jiang@zhengyaojiang
Thrilled to announce Weco has raised an $8M seed led by @GoldenVentures to build self-evolving software! Our technology has already been used by frontier labs like OpenAI, Meta, Google and Sakana AI. We’re making every codebase a living experiment that learns to beat itself:
English