Venture capitalist at @theoryvc Student of Startups Backer of 9 unicorns Author of https://t.co/IWw3R3RVLm Subscribe https://t.co/iDgoLXaF98
Skill distillation : big models teaching local models to work better. Latest example : Opus optimizing a skill for Qwen 3.6 35B A3B. Thinking on : 518s, 15,663 tokens, 7/10. Thinking off : 133s, 2,219 tokens, 10/10. https://t.co/hMVcwKj2Bv https://t.co/okIbXB51JO
View original →The better the harness, the better the business. Berkeley published a study this week showing harnesses, the systems that control AI agents, set the price of an answer. The right harness cuts the cost of the same result by 71% without a loss of accuracy. https://t.co/mxf1MusbZ2
View original →Are we seeing an AI pacing rotation? 🧵 While mega-cap infrastructure digested capex today (ORCL -4.3%, NVDA -2.8%), software surged : • Median SaaS stock up +4.5% (avg +5.2%) • 12 companies gained >= 10% in a single day • 92% of the index closed green (80 of 87) https://t.co/slzbEVwiA3
View original →AI consumption is not a smooth curve anyone can extrapolate : it arrives in three waves, each an order of magnitude larger than the last. The second one has already broken. The third is on the horizon. https://t.co/hv3rEQZMet
View original →Meta launched two things yesterday : a state-of-the-art model & a new pricing system for foundation models. Muse Spark propels US open source models to the frontier. Meanwhile, the pricing system resets the industry’s economics. https://t.co/OQZhU7hqMh
View original →From −94% gross margins to 0m of revenue per megawatt: how AI models learned to print money. https://t.co/KEG5ePxeP8
View original →Does the token supply chain also experience the bullwhip effect? Here's my explanation of this classic phenomenon and what it means for AI data centers. https://t.co/cHI2qViF74
View original →OpenAI's new Luna pricing now even below DeepSeek on intelligence/dollar. https://t.co/ps8r8dYnBi
View original →Every model you’ve ever used froze the day its training ended. The answers are the same even if you have used it every day. What if a model kept learning as you use it? https://t.co/YZAHIHmmfI
View original →State of the art models are two-thirds smarter than they were last November. The frenetic pace of improvement is sustained, two new models every three days. But 84% of tokens on OpenRouter aren’t state of the art. https://t.co/qdLjBgZDIp
View original →
