METR: Anthropic Models Are Advancing at Roughly Twice the Pace of a Year Ago
TechFlow reports that on November 7, according to The Verge, AI safety research institute METR stated that following an 11-week investigation into Anthropic, the company's model development pace is approximately double what it was a year ago, with the progress achieved over the past year equivalent to about two years previously.METR believes that as of October, the majority of the R&D acceleration is driven by AI, and its contribution is growing. Currently, this acceleration primarily comes from AI assisting human researchers rather than operating autonomously. Employees using Anthropic's strongest internal models self-reported in an informal survey that their work efficiency averaged approximately 12 times that of periods without AI assistance.The report indicates that Anthropic extensively redacted the publicly released version. Co-founder Sam McCandlish stated that the relevant information involves commercially sensitive content and could also help other companies accelerate their R
TechFlow reports that on November 7, according to The Verge, AI safety research institute METR stated that following an 11-week investigation into Anthropic, the company's model development pace is approximately double what it was a year ago, with the progress achieved over the past year equivalent to about two years previously.
METR believes that as of October, the majority of the R&D acceleration is driven by AI, and its contribution is growing. Currently, this acceleration primarily comes from AI assisting human researchers rather than operating autonomously. Employees using Anthropic's strongest internal models self-reported in an informal survey that their work efficiency averaged approximately 12 times that of periods without AI assistance.
The report indicates that Anthropic extensively redacted the publicly released version. Co-founder Sam McCandlish stated that the relevant information involves commercially sensitive content and could also help other companies accelerate their R&D.
