Unverified50% confidenceOpinionExact time
Noam Brown believes humanity has some time to solve the AI alignment problem, but not much.
1
Sources
50%
Confidence
Long-term
Relevance
9/24/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedA limitation of the framework that uses human time-on-task as a difficulty scale is its assumption of a monotonic mapping between the time humans spend on a task and the task's cognitive complexity, whereas certain mathematical breakthroughs hinge on moments of inspiration that cannot be measured in time.67% similarUnverifiedOnce AI chess strength crosses the human expert threshold, human experts can almost never beat AI again, and the transition window from 'always winning' to 'never winning' is extremely brief.63% similarUnverifiedRyan Lopopolo argues that the three truly scarce resources in AI-driven development are human time, human and model attention, and the model's context window.62% similarUnverifiedInvesting more inference time is a necessary but not sufficient condition for solving problems; thinking duration does not equal problem-solving quality.62% similarUnverifiedOpenAI research scientist Noam Brown believes that even physically air-gapping a computer may not necessarily prevent a misaligned AI from communicating externally.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/946507API
curl https://kongchang.com/api/v1/knowledge/claims/946507MCP
get_claim(id=946507)