Matthias Amon

This is why Opus feels so good to work with. I just asked for an additional Astra review, and it kept the agreed work on track. OpenAI models too often treat my latest message as the entire task and forget everything we were already doing. Opus treats it like a follow-up!

OpenAI today

🌶️ 6.1 Sol is only useful when Opus 5.5 prompts it

Opus 5.5 >>> Sol 6.1

Spent now 2 hours setting up Sol 6.1 properly in Claude, so Opus 5.5 can spawn it. It's an amazing combo 😍 Anybody interested in that skill?

And I thought this would be easy 😅 These are the games I have the best memories with 🥺

Atleast give us a normal reset too.. because of this post I just wasted all my usage in fast mode 🥲

Union Alpha, was a fun ride so far, but now getting a lot of 429 "Too Many Requests"

Union Alpha first impression: Running it via Open Router in Codex Desktop, and it's pretty good actually. - Code Quality 3.5/5 - UI 2.5/5 - Speed 2/5 - Understanding what I want 4/5 - Chatting 4/5 Its pretty nice to talk too, compared to Sol, Astra, .. Best use case, porting something that is well defined, UI is really lacklustering if you build more from scratch.

So hows Astra for you after the fixes? Any improvements so far?

Astra's code is the ugliest working code I have seen so far from any LLM

I am back again at Sol, can't stand Astra. Maybe it's skill issues, but this model is too costly and kind of hit and miss output wise 🤷‍♂️ Hows your experience so far?

Astra first impression: very slow 🐌 even on medium effort

AGI is here, but I can't load the blog post? 🫠

I am now 20% on my codex sub.. when do I use the banked reset now? Is there a general reset coming soon..? What you guys doing?

Holy shit, how good is "Ox Alpha" 🤯

Wtf is going on with Codex Limits? I do like 3x 5.6 Sol medium threads in parallel today.. never has been that bad.. already at 63% 🥲

Day 2 without GPT tokens @thsottiaux