r3bix

i do fun stuff

[r3bix @r3bix_]

@anonzro @OpenAI Don’t use advanced voice mode, the new one is called live and it can search web/ run other agents to do stuff and get results

[r3bix @r3bix_]

@Dmnn_ed @BhMaq888 very good, one weirdo less

[r3bix @r3bix_]

@ramasarat Inspiration at most lol

[r3bix @r3bix_]

@un3valuated @ouzdor I'm using claude code with gpt 5.6 models via claude code proxy and limit is going down much slower (maybe bc it burns less tokens also if i check tokscale)

[r3bix @r3bix_]

@un3valuated @ouzdor That's what I experienced with 5.6 in codex harness it's being so much inefficient when compared to opencode / claude code. I had been working on a single thread for longer time and every 2 messages (short UI tasks) it had to compact context lol

[r3bix @r3bix_]

@ouzdor codex harness has problems, when you switch to opencode / claude code with gpt 5.6 proxy it's gonna be dropping usage much slower unless you rail 5.6 sol max / extra high all the time lol

[r3bix @r3bix_]

@kyudo_krypto https://t.co/X9bRLfN8hK

Jakby ktoś szukał bo apka naprawdę fajna

[r3bix @r3bix_]

@kyudo_krypto Masz 3 mieś za darmo i potem jakaś duża zniżka ze to groszowo wychodzi za mieś jak zrobisz na maila .edu (mi nie zaliczyli na https://t.co/2rYFjFXabt i musiałem legitymację studencką wysłać na maila)

[r3bix @r3bix_]

@hadesdevs @0xSero Is computer use so token heavy? I found it useful for unity game dev, because codex can play around with unity and even test the game (its package delivery management type so click buttons to play) and it's doing great job to test with computer use.

[r3bix @r3bix_]

@thsottiaux @Kappaemme1926 It’s a bit confusing, in my opinion. If I needed to reset it urgently and saw that the first reset didn’t work, I would immediately use a second reset and end up wasting one.

[r3bix @r3bix_]

@thsottiaux @Kappaemme1926 Okay that is some nasty UI bug because after a few minutes it applied and i could see 100% weekly in my macos app, but on windows I still have 2% low usage warning...

[r3bix @r3bix_]

@XAppUser @thsottiaux What you're making with GPT-5.4 that it's better then newer models for it?

[r3bix @r3bix_]

@thsottiaux I have 4 banked but would be nice to get one more xd

[r3bix @r3bix_]

@gdb I use Work to iterate on my mobile app designs and create concept art / assets for gamified app I'm working on. I like that ChatGPT Work allows me to progressively iterate over the design and it has good knowledge about my project. Chrome connector allows Chat to look up designs.

[r3bix @r3bix_]

@sylogista @Mordrago @AtrakcyjnyAdam Ta ściągnij sobie kilkuterabajtowy model i odpal na swoim chromebooku

[r3bix @r3bix_]

@thsottiaux I like that even pro 5x plan gives me more usage than competitors and that 5h limit got removed. GPT 5.6 Sol and Luna are great models. They're fast, smart, and make building so much smoother and Codex makes it seamless to generate great UI mockups with the best GPT Image model!

[r3bix @r3bix_]

@milan63529480 @maks6361 its probably shitty app that's all

[r3bix @r3bix_]

@ProfitAII no i na chuj tak szponcic jebane cebule jak masz projekt z ktorego zarabiasz to kupujesz claude max i sie nie pierdolisz z 12 kontami google kazde ma max 3 prompty w duzym codebase, a do nauki to polecam za 5$ opencode go, dostep do roznych modeli w tym glm 5.1 i duze limity

[r3bix @r3bix_]

@saeid64 @Surendar__05 @Copilot @claudeai Opus 4.7 is in 39$ sub not the 10$