In its amended IPO filing on Tuesday (Nov 10), Anthropic said it has determined it met the 'research acceleration threshold' in its Responsible Scaling Policy — the point where AI assistance is expected to roughly double the pace of AI capability gains. The company stressed the judgment rests on expectations about future models and the overall pace of progress, not on demonstrated capabilities of current systems, and said it is still assessing the data and will publish further reports. Independent evaluator METR estimated Anthropic's progress on defined capability tests has run at about 1.9x its earlier pace since October 2025, while cautioning its estimates are highly uncertain.
What interests me: meeting the threshold is meant to trigger stronger safeguards, not a halt — but Anthropic also warned faster development could outpace its ability to validate safety measures.…