Anthropic and OpenAI share a rare view: AI isn't "hitting the wall," it's just getting too smart

Recently, the controversy over whether the Scaling Law of AI models has hit a "wall" has attracted a lot of attention. In this debate, OpenAI and Anthropic, as competitors, are in rare agreement that AI is not "hitting the wall". The core of this debate focuses on: with the continuous expansion of the model scale, its performance improvement will encounter an insurmountable ceiling. In response, OpenAI CEO Sam Altman has made it clear on social media that "there is no wall! （There is no wall." Michael Gerstenhaber, Anthropic's vice president and technical director, also weighed in. He pointed out that large language models and other forms of generative AI are steadily improving in terms of "self-correction" and will be able to perform more tasks and promote the development of "AI intelligences". (Tencent Technology)