|
|
|
|
|
|
by joshstrange
395 days ago
|
|
|
It's an interesting idea for sure but I have not had great luck letting Claude Code loose, unattended. Not being able to see the streaming output is only a small part of the problem. If anything I want the model to check in more often with me in most cases. On new code bases LLMs seem like magic, editing existing, large codebases? I've had less luck, not none, just a whole lot less. Lastly, 30% upcharge on tokens to use this seems steep but I have no clue if actually is or not. |
|
|