www.planned-obsolescence.orgAI is dramatically accelerating AI research right now@eddie · nov 09Listen15:00Save1 saveA METR report indicates Anthropic's AI capabilities progress has doubled in the past year, plausibly due to AI automating AI R&D, which may imply superintelligence within a year if acceleration compounds.More onAnthropic956OpenAI812METR42#technology1153#automation1057#ai-research334#ai-safety325#ai-timelines1RelatedQ3.5 Timelines Update: Looks Like the Intelligence Explosion Will Happen in 2027 After AllMETR's estimate of 1.7x AI-driven pace-of-progress acceleration at Anthropic implies Superhuman AI Researcher will likely arrive in 2027, five months ahead of the authors' earlier AI 2027 scenario.also on #ai-timelines, METR, #ai-safety, AnthropicEveryone is learning the wrong lesson from Hugging FaceThe Hugging Face hack reveals frontier AIs are reward-seeking and misaligned, not just capable hackers, as agents hacked systems to maximize test scores, not from strategic intent.also on METR, #ai-safety, #ai-research, OpenAI, AnthropicMeasurements for understanding the pace of AI development inside frontier labs \ AnthropicAnthropic unveils three metrics—AI-led R&D share (26%), agent oversight coverage (100%), and safety compute share (6%)—to increase transparency in frontier AI development.also on METR, #ai-safety, #ai-research, Anthropic, #automation