Rhys Arkins

[199X @199x_now]

Tony Soprano discovers AI therapy

[Rhys Arkins @rarkins]

@trashpandaemoji

[Rhys Arkins @rarkins]

@steipete I'm curious - is there any way that we could break out token consumption by use? e.g. we know how much of our quota use is from coding vs the autoreview

[Aelfred The Great @aelfred_D]

Nowadays kids’ books have messages like “believe in yourself” and “you can do anything” but when I was a kid the messages were things like “a lady who has a candy house wants to cook and eat you” and “sometimes wolves can disguise themselves as old women”

[Rhys Arkins @rarkins]

@thsottiaux @dev_bordecraft @paraddox I'm on the 20x Pro and still not reset

[Rhys Arkins @rarkins]

@thsottiaux FYI it's 30+ minutes later and still not reset in my Pro account

[Rhys Arkins @rarkins]

@_philschmid @googlehealth But can you sync it to Apple Health?

[jo 🎣 @girlcohle]

are you ready to die ryan gosling

[Rhys Arkins @rarkins]

@MeepBird_SOL

[Rhys Arkins @rarkins]

@steipete I'm guessing you may not be using oracle as-is any longer due to the new and magical internal access you have. Are you still interested in PRs to maintain it? It seems broken since 5.4. Example PR: https://t.co/LkIaIPcAWh. My clanker has its own local fix too.

[Rhys Arkins @rarkins]

@jullerino https://t.co/nmjE4ATNHm

[Rhys Arkins @rarkins]

Rolling out GPT-5.3-Codex-Spark's independent limits after nearly blowing my GPT-5.4 limits for the week

[Rhys Arkins @rarkins]

@jullerino Same for me. Switched to codex 5.3 spark as it still has different limits

[Rhys Arkins @rarkins]

Rolling out GPT-5.3-Codex-Spark's independent limits after nearly blowing my GPT-5.4 limits for the week

[Rhys Arkins @rarkins]

@bytebot @ASalvadorini @antigravity @madebygoogle Aren’t CLI quotas unaffected by a Pro subscription? Ie same quota as free so no benefit of Pro?

[Rhys Arkins @rarkins]

@steipete Does this mean you're using worktrees now (implicitly enabled by the Codex app) instead of multiple git clones?

[Rhys Arkins @rarkins]

@MiniMax_AI I'm still confused. Now you need _Basic_ agent plan, and then can add on a coding plan for if you run out of requests on the agent plan?