David Sacks: Open AI Models & Harvey's Legal AI Innovation

David Sacks says restricting open artificial intelligence models will not stop China from building advanced AI. He backs Harvey's new Tenet model, which is designed for legal work. Here's the thing: Sacks, a former White House AI and crypto czar, defends open-source AI. He highlights Harvey's Tenet model as an example of American innovation using open systems. Harvey used the open-source Kimi K3 as a base and trained it on legal data. This resulted in top-tier performance on legal benchmarks at a fraction of the cost of proprietary models. Sacks argues that restrictions would only cripple startups like Harvey. These companies rely on open systems to build affordable, high-performance tools. He also notes that some closed AI labs would welcome such restrictions as they would eliminate competition. What's interesting is that Sacks had previously called Kimi K3's benchmark rise "concerning" in July. Kimi K3 is a 2.8 trillion-parameter open-weight model. Harvey's Tenet model, built on Kimi K3, showed an 82% improvement on the LAB legal benchmark. It also runs at under one-fourth the cost of leading foundation models. The bottom line: The debate around open-source AI models continues, with implications for innovation and competition in the AI landscape.