Unverified50% confidenceFactExact time
The industry has recently observed that benchmark improvements from scaling have noticeably diminished, and high-quality training text data is approaching the upper limit of what is obtainable from the web.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/11/2026
First Seen
Valid until: 12/10/2026
Sources
Related Entities
Related Claims
UnverifiedSince 2024, evidence from multiple directions has suggested that relying solely on pre-training scaling may be encountering diminishing returns, as the available amount of high-quality text is approaching the ceiling of what exists on the internet.78% similarUnverifiedExperiments show that after only a few generations of self-training, the variance of model outputs drops sharply and the text tends toward homogeneity74% similarUnverified数据规模的增长并不等同于有效学习信号的增长，在模型能力提升到一定阈值后简单堆砌样本的收益会急剧下降73% similarUnverifiedThe retraining involved a clear trade-off: slight decreases in some very high scores on cleaner validation distributions, in exchange for improved real-world performance.71% similarVerifiedThe marginal returns of simply increasing model parameters and training data are diminishing, a phenomenon known as 'Scaling Law slowdown.'70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/901924API
curl https://kongchang.com/api/v1/knowledge/claims/901924MCP
get_claim(id=901924)