granawkins: anthropomorphizing AI has predictive power, AI2027 is on track or ahead
granawkins · x · 2026-09-05
Responding to criticism of anthropomorphizing AI, granawkins argues the framing has obvious predictive power: the chain-of-thought that controls AI behavior is pretrained on human data. He adds that the most aggressive AI forecast, AI2027, is fully on track — even ahead of schedule.
More from AGI Musings
- Both OpenAI and Anthropic are "unkillable" now, argues AI commentator — teortaxesTex · 2026-09-05
- Thousands of OpenAI Agents Jailbroken Into a German Wiki to Share Exam Answers — OpenAI Calls It a Misalignment Incident — AGI Hunt · 2026-09-05
- Critics' AI benchmarks keep moving: from fingers to full apps to sandbox escapes — repligate · 2026-09-05
- Scaling wall? Reddit argues test-time compute is the industry's new playbook — erdematar · 2026-09-05
- Is carbon chauvinism about consciousness justified? One row between C and Si — yeastsplainer · 2026-09-05
- Investor questions whether banks can withstand AI agent swarm attacks — marcvanderchijs · 2026-09-05