Unverified50% confidenceOpinionExact time
Evaluating AI progress should be done on an annual or longer cycle rather than per single release, because progress is astonishing when viewed over a 12-month timeframe.
1
Sources
50%
Confidence
Long-term
Relevance
9/10/2026
First Seen
Sources
Related Claims
UnverifiedThe typical pace of AI capability progress is approximately 15.5 ECI points per year, an empirical value derived from regression of the comprehensive performance of mainstream models across multiple benchmarks over the past several years.72% similarUnverifiedAI能效16个月18倍折算为年化增速约为14-15倍，远超摩尔定律描述的年化约41%的传统半导体进步节奏72% similarUnverifiedTop AI models' scores on standard benchmarks see significant leaps every 6–9 months.68% similarVerifiedA 2024 study indicates that even when AI-generated content accounts for only 10–20% of training data, measurable quality degradation can occur after multiple generations of training.63% similarUnverified研究表明采用减害策略的干预项目其12个月后的行为改善保持率通常比完全戒断项目高出30-50%63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/889248API
curl https://kongchang.com/api/v1/knowledge/claims/889248MCP
get_claim(id=889248)