Post

Why does it seem like every new model release starts out feeling like a big improvement, but slowly gets worse over the course of 2 or 3 months to the point they are almost unusable? Are my expectations just changing, or is model degradation a real phenomenon?

8/4/26, 2:11 pm0 reposts1 like

Stephan Schlecht

@sw7com.bsky.social

8 Apr

Good question. In fact, one often gets that impression. Still, it would be interesting to know whether this claim is largely unfounded and based solely on subjective impressions or isolated reports, or whether there are independent long-term comparisons?

Stephan Schlecht

@sw7com.bsky.social

9 Apr

www.pcgamer.com/software/ai/...

pcgamer.comClaude 'cannot be trusted to perform complex engineering', says AMD's AI head honchoThis is reportedly since changes made in February.

Alex

@invoke.dev

9 Apr

My impression is that the models get lazy and apply less reasoning to save on compute time. I found this study that suggests the performance varies by infrastructure load throughout the day or week. arxiv.org/html/2602.15...

arxiv.orgDaily and Weekly Periodicity in Large Language Model Performance and Its Implications for ResearchContent selection saved. Describe the issue below: