Amazon open-sources TorchNeuron, so PyTorch runs natively on its Trainium chips
Developers can use Amazon’s AI chips without rewriting much code.
Tip
How close are non-Nvidia AI chips to everyday developers?
Inception Labs’ diffusion language model Mercury 3 tops 1,000 tokens per second
It writes text with the “diffusion” method used for images, fast and with a longer context.
Tip
Why are some AI answers fast and others slow? How a diffusion model differs from the ChatGPT you use
Enterprise AI startup Town raises a 1B valuation
A unicorn only months after launch.
Tip
Why enterprise AI gets to $1B so quickly
MiniMax releases M3.1 Flash, with weights coming Oct 23
A lightweight model from a Chinese company, downloadable this week.
Tip
What “open weights” means and what individuals can do with them
AI chipmaker Tenstorrent raises 6.2B valuation, double last year
Investors include Fidelity and ARK.
Tip
Where the AI chip money goes, and who besides Nvidia is getting it
⚠️ Compiled by an AI assistant from public AI news roundups; original articles are not individually verified. For reference only.