Comment by layoric
2 days ago
Very timely, just this morning:
Me: "Review the following <file> and work through the implementation"
Opus 5: "Called tool <blah>, Called tool <blah>..." - for a few minutes.
Opus 5: "I've implemented X, do you want me to commit changes?"
Me: "None of those changes are on the file system"
Opus 5: "You're right, all the tool calls were fabricated."
Should have said yes, told it deploy to prod, then given it a real browser to test its changes. For science.