[exizt88]

> We are seeing improvements with each model release these days but it’s clear that the improvements are getting smaller and smaller.

This is obviously untrue, both with GPT-5.4, and Claude Fable as examples in the last 6 months.

[rubin55]

I would struggle to ascertain the day-to-day difference between GPT-5.4 and GPT-5.5 tbh. Also, imho, Fable is highly hyped, I don't think it is dramatically better than Opus 4.8. Maybe my tasks and interaction with AI is relatively simple (i.e., lots of Rust programming, Linux system engineering stuff).

[chiply314]

I haven't had enough time with fable, but I had to look back on how i worked with claude just 6 month ago to remind myself that it got a lot better.

Like i still used plan mode 6 month ago now I don't.

I would argue that with every model release we have a new learning phase.

[skerit]

> the improvements are getting smaller and smaller

The AI haters have been saying this for 2 years now.

[byzantinegene]

gpt 5.5 regularly wastes tokens on wrong commands, requires lots of handholding. I highly doubt there's substantial improvement