#开源生态

This is how open source should work. A closed lab publishes research, and within days the community turns it into training environments anyone can use. Early and rough, as Clem says, but the direction is lovely. #OpenSource #AI

Berlin-based search engine Ecosia, used by government services, is dropping its French AI partner, Mistral, its CEO told POLITICO. The company will now bet on open-source models, including those from China. https://vist.ly/5mh5y

认识 Mistral Large 4，又名 Le Chonk。 • 1T 参数，原生多模态。49B 激活。 在综合基准测试上，它是来自美国或欧洲的最佳开放权重模型。 • 在关键工作负载上达到最先进水平，包括网络防御、制造业和金融，并且在视觉定位方面超越了闭源前沿模型。 • 端到端在欧洲打造，并可通过我们自己的 Mistral Cloud 基础设施从欧洲部署。 • 今天即可通过 API 向所有人开放。正与网络安全合作伙伴私下合作。 开放权重版本将于十月底发布。

One step closer to 4-8x faster Ethereum finality! It took some time and lots of tokens, but we now have a formally verified proposal for a decoupled consensus protocol in I* (a future Ethereum upgrade)! Not yet a full spec (up next), but it includes all the key consensus-relevant details to become one. Since Ethereum aspires to be live without most of the stake online, the protocol involves many more components than a normal BFT protocol, and its correctness involves much more than standard safety and liveness. Those nuanced properties are now verified! What's more, I came away convinced that all protocol design will involve AI-assisted Formal Verification in the future, both for correctness and iteration speed. The work wasn't limited to just: Design the protocol -> Formally verify it Instead, the loop became more like: Design -> Formal Model -> Find exactly what breaks and why -> Redesign it. For a fairly complicated protocol like this one, I think having the Lean model be part of the design loop played a big role in accelerating the process. A future with agents paired with formal models is a superpower for Ethereum development, because they can then use those models to find exactly where an argument breaks down, formalize counterexamples, test proposed fixes, iterate on the protocol. Many details that would slip under the radar when asking agents (and indeed, humans) can now be specified exactly and checked by the Lean kernel. This then forces agents to be more precise and lets them make verifiable progress on their own. It's been incredible to see this play out, seeing agents find gaps and propose protocol changes to fix them. In other words, autoresearch can speed up protocol design, formal verification is here to stay, and Ethereum Finality will get faster.

🚢 Marin 535B-A23B started training this week! As usual, the whole process is open. Voyage plan: pretraining (80%) + midtraining (20%) on 18.75T tokens on 11 x GB200 NVL72 for ~3 months (2.7e24 FLOPs). Post-training will follow. Before kicking off the run, we trained a 4-rung scaling ladder from 1.6B-A61M (48B tokens) to 27.7B-A1.2B (926B tokens) to debug issues, and to make a forecast of our hero run. This is by far our biggest run, so definitely expecting the unexpected.

Moonshot AI's latest open-weight model, Kimi K3, is now available on Databricks through Unity AI Gateway. @Kimi_Moonshot Run Kimi K3 where your data already lives - governed, secure, and ready for custom AI apps and agents built with the data in your Lakehouse. Unity AI Gateway lets you deploy Kimi K3 with enterprise-grade access controls. Govern every call, monitor performance, and scale securely across your AI apps. Test Kimi K3 alongside other frontier models without changing your application code. The open-weight frontier just arrived on Databricks. Try it today. https://www.databricks.com/blog/kimi-k3-moonshot-ai-now-available-databricks-through-unity-ai-gateway