Post #1986387
2026-04-23 14:05 UTC
@martin @ChadF @dave Theo Browne did a deep dive into the overlapping factors causing CC to perform worse.
- Tokenizer Changes in Opus 4.7
- Default 1 Million Token Context Window
- Thinking Content Redaction
- Increased API Thrashing
- Default Lower Effort Level
The only 2 things in your control are the default effort level and adding `CLAUDE_CODE_DISABLE_1M_CONTEXT=1` to your `.claude/settings.json`
https://www.youtube.com/watch?v=KFisvc-AMII
Replies (1)
-
@dave@podcastindex.social 2026-04-23 15:01
@nathan @martin @ChadF I think so much of it comes down to usage patterns and how each one of us interacts with it. I've basically seen zero difference in CC behavior over the entire few months I've been using it. It would be very interesting to see how other's work differs from mine and compare.