
Igor Varga
193 posts



At the core of our mission is working through how to ensure increasingly powerful AI benefits everyone. We believe that, at some point in the future, AI acceleration for frontier model development may be so high that the world will need to pace the rate of AI advancement. We hope to contribute to work led by the U.S. government, alongside other labs and the open-source community, to develop the tools and mechanisms that could make that possible. pacingthefrontier.com





AI security advances when the industry builds in the open, together. We're introducing the Open Secure AI Alliance with industry leaders to develop new techniques and tools to safeguard software and agents. By sharing models, tooling and research in the open, we can broaden the community of defenders. Learn more about the founding members’ contributions: nvda.ws/4pD8Fc5



@DominikTornow We will still have the acceptance phase of whatever comes out of agentic loops. I can already envision human board sessions, during which they are pressing the platinum "Accept" button and declare productive day of work after.




There’s been a lot of speculation about where we stand on open-weights models. We’ve outlined our views in full here: anthropic.com/news/position-…





Here is how I see the Kimi K3 license works in real life: - Clause 2, the Model as a Service clause, is intended for all the cloud platforms essentially reselling K3 to their customers. This includes hyperscalers (AWS, GCP), to neoclouds (Nebius, CoreWeave), to inference platforms (Fireworks, Together, baseten). The $20 million trailing 12 months revenue threshold is comically low. Everyone meets it and will have to negotiate a commercial, revenue-sharing contract with Kimi. All the ones that have offered Zero 0 support today already did. - Clause 3, what I called the Branding clause, is less about revenue generation, more about aura farming. For example, if say AT&T (who does a lot of open source AI work btw) wants to use Kimi 3 to power its customer service agent, if if they makes tweaks and fine-tuned, it will have to show "powered by Kimi 3" or something along that line to deploy this service. AT&T is obviously big enough to pass the user base and revenue threshold. Customer service is also a commercial in nature. But AT&T does not need to pay Moonshot a dime.



Releasing the model weights and technical report of Kimi K3. Kimi K3 is our most capable model: a 2.8T MoE model with native visual understanding and a 1M-token context window. New model architecture: 2.5x the intelligence per unit of compute, not just more params. Alongside Kimi K3, we're opening up more of the stack behind it — high-performance attention kernels, MoE communication library, and infrastructure for running agent environments at scale. Model weights: huggingface.co/moonshotai/Kim… Tech report: github.com/MoonshotAI/Kim… Tech blog: kimi.com/blog/kimi-k3










