开启时会模糊预览图，关闭后正常显示
The question with PaCoRe is not «can an 8B model beat GPT-5 on some math benchmarks» (of course it can) but «how much upside it provides to actually frontier models». If the same scaling law worked for V3.2, it'd be an absolutely insane data machine.
Gemini dreams…
Use Claude Code status command to see which models you use mostly. I took a break from Claude Code for couple weeks (busy with my other love Kiro, Cursor, Windsurf). I checked today which model I used the most when it started tracking stats. Turns out it is Kimi-K2-Thinking @Kimi_Moonshot @crystalsssup It is a good model! I will publish this stats at least once per month going forward.
Artificial Intelligence @amazon, @awscloud Reinforcement Learning, OSS AI, General Purpose Agents, Autonomy All views personal!
2026 will see AI go deeper into the Fortune 500 from experimentation into multi agent products - with it, we’ll see a greater need for a shared context layer and new roles emerge to coordinate agents
partner @a16z // saas + b2b fintech // strong opinions on 🍕
People are still in such a bubble. As I’ve been arguing forever, DeepSeek is an educational project that describes a path to the frontier in model performance and eventually AGI. You should just collate their papers into a textbook.
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
Holiday PSA: if you’re a company sending a gift to clients this season, you will be beloved if you send a nice generic regiftable candle that can be given to an in-law or aunt last minute. Logo on the card! Generic gift in the box! Thank you for your attention to this matter.
Co-Founder, American Dynamism. General Partner @a16z. Catholic. Mother. American. 🇺🇸 🚀💪
RT @peterwildeford: I'm a proud member of the "people who think AI is going to be really important, not like internet important but like re…
Applying the security mindset to everything @PalisadeAI