Y Combinator CEO Garry Tan is urging US open-weight AI labs to actively 'distill' top-tier frontier models, igniting fresh debate over copyright hypocrisies and venture capital motivations. The call comes as proprietary labs try to guard their proprietary capabilities while smaller, open-source alternatives attempt to close the performance gap.
For early-stage founders and £500-ticket backers, this isn't high-minded tech politics—it is a practical question of runway survival. Relying exclusively on expensive proprietary APIs like OpenAI or Anthropic quickly consumes pre-seed capital, forcing founders to raise larger, more dilutive rounds just to keep their product online.
Tan is undeniably talking up YC's portfolio economics, but selfish incentives often yield public goods. If distilled open-weight models can deliver near-frontier performance at a fraction of the cost, early-stage startups will get significantly more leverage per pound raised, stretching angel capital further before liquidity.