Looped

I build open-source tools, games, and local-first apps. Unix enjoyer.

[Looped @NonLooped]

@forgebitz 1,000 different apps that solve this issue already exist yet I see a tweet like this daily

[Looped @NonLooped]

@bubbleboi focused on coding?? when was the last time you’ve seen the curriculum of a CS degree 😭

I can count the coding related courses with one hand compared to everything else

[Looped @NonLooped]

@yanb0nd @nikolay_dp @nikolay_dp my agent did flag an issue regarding this so I suggest you look into it

[Looped @NonLooped]

@yanb0nd @nikolay_dp that makes no sense, MonoCode just runs Claude Code under the hood, so it shouldn’t be using more of your limit than Claude Code itself

[Looped @NonLooped]

@byte_the_dev @nikolay_dp I kept Windows updates manual for the first release because the in-app Tauri updater had some issues and needed extra setup and testing, and I wanted to ship Windows support without delaying it

[Nikolay Petkov @nikolay_dp]

huge thanks to @NonLooped for making Windows support possible 🙏

[Looped @NonLooped]

@nikolay_dp Gotta make the Windows version look as beautiful as the macOS one 🙏🙏

[Looped @NonLooped]

@maria_rcks Orchestration has been two weeks away for AGESSSSS

[Looped @NonLooped]

@mitsuhiko @thdxr Even within codex they don't normalize all the actions so a few of these python and node tool calls are shown raw

[Looped @NonLooped]

@edinsoncode Sure, why not!

[Looped @NonLooped]

@luisnomad Same, 20x plan and it's burning like crazy with such small tasks. All the hype around it's efficiency doesn't seem to be applying for ANY of my use cases or workflows.

[Looped @NonLooped]

@reach_vb @ArielKwiat It goes off scope so hard, the same issue that we had with 5.6 but now it’s even worse. I’ll give it such a small task and it’ll go and refactor other stuff

[Looped @NonLooped]

@leodev I 100% agree with you on the reasoning mode point, I've used it on Medium on a Pro 20x plan and it consumed 7% of my weekly usage within minutes

[Looped @NonLooped]

GPT 6 Astra on Low/Light mode is such a blessing, 5.6 Sol's Low mode felt dumb but this is this different.

I've given it a ton of tasks and it completed them in 2 mins with great accuracy and even ran visual checks.

[Looped @NonLooped]

@theo Worse than Sol, not as bad as Fable 5.1

[Looped @NonLooped]

@nikolay_dp @al3rez The pros outweigh the cons for this type of application when comparing GPUI to Tauri, I don't recommend it here.

[Looped @NonLooped]

@bridgemindai You've "cancelled" your claude subscription like 20 times already 😭

[Looped @NonLooped]

@MarcosHernanz Based on my usage of Astra and the stats, this is definitely not the case. It seems like I was getting ~50-60% more usage out of Sol for the same amount of credits/usage. The gap in output tokens and steps is just not showing in my workflow at least

[Looped @NonLooped]

@lxopetrucci @fedesimio Now imagine astra intelligence with questions 🙃

[Looped @NonLooped]

@lxopetrucci @fedesimio Ofc, difference is they’re not as smart and their questions are often horrible and can be answered with common sense.