cageface

Do you have any citations for this? Not that I doubt you but I’d like to read more about it.

The new concise output style they just released definitely helps with Opus 5 in general.

I’ve been getting good results from using Sol for planning and Luna for everything else. At current rates it barely scratches my weekly quota.

Makes me wonder if the current Luna prices are sustainable.

Just my own apps. Getting good performance out of SwiftUI on Mac is quite difficult. Loads of footguns.

I’ve had good experiences with Flutter too. It’s actually much easier to build a performant macOS app with flutter than with SwiftUI.

This is true but if you want to monetize it then there are similar review hurdles you're going to have to clear.

Verified this locally myself. Thanks for the concrete test. I guess it's time to give Claude another try.

Codex does better if you ask it to take screenshots and critique its own UI work and iterate. It rarely one-shots something I like but it can get there in steps.

In the US. In many other countries it's borderline useless.

LLMs tend to do better with component libraries like shadcn or daisy than writing raw CSS in my experience.

The pro plan is useless. You need at least the 5x max plan to get any real work done.

That said I find the GPT plans much better value.

If you think Android tooling is subpar wait until you try iOS.

I've had exactly the opposite experience. Getting great results using GPT for hours every day since 5.3. You need to put the effort level on at least high though.

Every time I hand off a task to Opus to see if it's gotten better I'm disappointed. At least 4.7 seems to have realized I have skill files again though.

On at least high effort level I find GPT 5.4 easily beats Opus 4.6 in code generation and debugging issues.

Unless you need to work with a repo that uses submodules or lfs.