Advice for OpenAI: restore limits, fix taste, ship a new model
The author urges OpenAI to bring back the 5-hour limits and 20x usage on Pro, says no OpenAI model is currently worth using over Opus 5.5, and argues Astra did not fix the taste problem.
If I were Tibo I would: - Bring back 5h limits on Pro, it's clear they don't have enough compute to keep them off - Make Pro 200 have 20x usage again - Release GPT-6.5 Sol/Astra or whatever, no OpenAI model is worth using over Opus 5.5 right now - Fix taste, I know it's not easy but we've been complaining about taste and UI forever and all we get is temporary solutions, they didn't fix it with Astra they just made everything pastel green so we wouldn't notice for a while
This post quotes another post: x.com/thsottiaux/status/2109152716954222849