degenerating degenerate 🔒
Middleaged tech looking for likeminded souls
- Posts
- 965
- Followers
- 13
- Following
- 3
-
"We also saw a recurring behavior that we described internally as “speculation.” When the agent lacked sufficient context, it filled in the gaps with assumptions that appeared reasonable but were not verified."
https://1password.com/blog/what-we-learned-using-ai-agents-to-refactor-a-monolith
Right... it seems to me to be related to choosing to do the minimum asked for. If it claims the problem was a broken very unlikely error path, it can just stone cold tell you it fixed the symptom by fixing the error path, without showing any relationship.