Anthropic’s AI development has doubled in speed despite slowdown calls
Independent monitoring finds the San Francisco AI firm is improving model capabilities about twice as fast as a year ago, even as its CEO urges a pause.
An analysis by the nonprofit Model Evaluation and Threat Research (METR) shows Anthropic is accelerating its AI development at roughly double the rate of the previous year. METR reviewed internal data and interviewed staff over an 11‑week period from mid‑August to late October, gaining unprecedented access to the company’s research process.
The faster progress is attributed largely to the use of AI itself to improve its models, rather than increased computing power. METR says Anthropic’s Claude chatbot is already helping create newer, more advanced versions of the system. The firm is preparing a Wall Street debut later this month, the largest IPO for an AI company to date.
Why it matters to your business
Founders should note that AI capabilities can improve rapidly even when leaders publicly call for slower development. Faster model upgrades could raise the bar for product features, cost structures and talent competition.
The report also highlights the growing role of AI‑in‑AI, which may accelerate innovation cycles and increase safety concerns. Companies that rely on AI need robust oversight and risk‑assessment processes to avoid unexpected consequences.
What to do on Monday
- Review your own AI development roadmap against the pace reported at Anthropic; identify any gaps in capability or safety review.
- If you use external AI models, add a checkpoint this week to assess model performance and alignment with your risk standards.
- Consider establishing an independent audit or consulting partner to monitor AI progress and safety, following the METR example.
Sources
Reported by fastcompany.com.