创始人 Hill-Smith：单位美元换来的 AI 能力提升将持续递减
Hill-Smith 的观点解释了 OpenAI 和 Anthropic 为什么烧钱：单位美元的能力收益在递减。他还判断"fast takeoff"基本不可能，等于否定了单家实验室碾压全行业的剧本。关心模型竞争格局和选型成本的人可以看看。
创业公司创始人 Hill-Smith 提出，AI 能力随投入增长呈现边际递减，每一美元能买到的能力提升会持续缩小。他用这一点解释 OpenAI 和 Anthropic 为何需要巨额资本：要保持竞争力，必须不断加大算力与训练投入。他同时认为"fast takeoff"（一家开发者的模型进步快到对手无法追赶）"pretty unlikely"，即很难出现单一玩家甩开所有竞争者的局面。
这一判断直接影响对头部实验室的融资逻辑的理解：如果能力提升的性价比在下降，规模化的资本投入就从"加速器"变成"维持牌桌资格的门票"。对开发者与企业买家而言，这意味着短期内更可能出现多家模型能力接近、以价格和服务竞争的格局，而非一家独大。
AI capability gains per dollar will keep shrinking, according to startup founder Hill-Smith, which is why OpenAI and Anthropic need so much capital.
He called a "fast takeoff," in which one developer's models improve so quickly that rivals can't catch up, "pretty unlikely."
Full story: https://t.co/MGhG9A45ZD
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include OpenAI, Anthropic, funding, industry trends.
Source attribution: curated from The Information (@theinformation). Cross-source corroboration: 11 independent sources have covered this development (including Andrew Curran, IT之家, Gary Marcus), confirming industry-wide signal strength.
