Developer questions the hype around Fable for coding
A developer posted on Reddit's questioning why so many people are hyping up the for coding, since first-shot results are said to be near-perfect for others. In this person's own experience, Fable performed about the same as opus 4.8, occasionally a bit better, but not dramatically so.
Fable would over-complicate problems and run down unproductive paths, only snapping back on track when explicitly told to stop guessing and be certain. Even after getting on track, it would eventually repeat the same unproductive pattern again.
The poster finds it strange that other users describe strong emotional reactions, like anxiety over running out or losing access to Fable. They're currently modding a video game using VS Code and an embedded terminal with Claude, and haven't noticed much quality difference compared to using a regular chat interface before.
Key points
- Poster saw little meaningful difference between Fable and opus 4.8 in their own coding tasks
- Fable sometimes over-complicated problems or went down unproductive paths
- Explicitly instructing it to stop guessing and be certain was what got it back on track
- Other users' strong emotional reactions to seem exaggerated to this poster