METR report notes Anthropic model improvement rate has doubled, raising concerns among security researchers
METR has released an 11-week report on Anthropic, stating that its AI model improvement rate is about twice that of a year ago. This means the progress achieved in one year now equals roughly two years' worth of advancement from before. Researchers believe this acceleration stems primarily from AI itself rather than simply increasing computational power.
Return to original text
A short introduction is retained here; the full text from the source website is not reposted. For specific wording and subsequent updates, please refer to the original article.
Open AIHOT original