The 31st edition of Ruby AI News is here! This edition features app builders and skills that turn a single prompt into a full Rails prototype, benchmarks making the case for Ruby’s token efficiency, and much, much more🔻
Are scaling laws finally working for time series foundation models?
Today, @datadoghq is releasing Toto 2.0 weights in Apache 2.0 on @huggingface. It's a family of open-weights TSFMs from 4M to 2.5B parameters, where every size beats the last from a single hyperparameter config. First across the leading benchmarks: BOOM, GIFT-Eval, and TIME.
Most TSFM families ship multiple sizes that all perform roughly the same. This one doesn't.
Why it matters: scaling laws gave language and vision a predictable relationship between compute, data, parameters, and downstream performance. Time series hasn't had that curve until now. Once you have it, you can scale data and compute with confidence, and start asking which new capabilities emerge at the next order of magnitude.
2.5B open-source weights: https://t.co/prpcGoCw0U
4M open-source weights: https://t.co/5d6rw5NYL2
Blogpost: https://t.co/xKazgTMh1I
Fragments: 34th Thoughtworks Technology Radar, what happens when developers don't read the LLM's code, DirectFile and tech in large organizations
https://t.co/brBXaa3xOc
Latest prototype: `binding.irb(agent: true)` to let agents interact with your running Ruby program through a local socket.
No skill or MCP server required. Basic instructions prompted upon first socket connection.
@inazarova@garrytan How did the whole dev process go? Did you end up using any plugins/skills for specs, or was it quick back-and-forth chat with the agent?