alina 🌸🦴

thus 5-6, not 8. haiku interference takes about 1.5s

what

how the fuck does claude -p have ~5-6 seconds startup time. isn't bun supposed to have the fastest startup?

i don't like gpt models and i also don't really want to pay that much

rocm is annoying af to get to work

the only thing that kinda worked for me was using the repo.radeon.com wheels they have linked in their docs (although tbf that was in wsl and they have a separate doc page on wsl that also says you to patch something lol)

the sad thing about that article is that its plausible enough that i wouldn't be very surprised if it was real

ok i do have spare api credits apparently

it used to work in claude code and claude web too
i dont have spare credits to check the "real api" but i doubt it would be different there tbh

ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86 is no longer a thing i think???

the checkbox workaround for ios safari is so dumb. i think its worth noting this website is best experienced on an iphone

i do wish we had a proper haptics api and not a semi supported vibration api tho 😭

usually a mitm proxy of sorts? pretty sure I saw something like that used for gradle at least

github.com/polygon/nix-...
github.com/chayleaf/mit...

though to be fair it's my fault for staying up at 5am if anything

im hitting the 100$ plan limits and im STRUGGLING

access to claude is a human right

she's a 10 but she listens to 2000x mobile phones' ringtones instead of spotify

wtf 😭😭😭

submitted my two weeks :3

spaced repetition is goated

claude.ai chat search so bad i had to vibecode a script that scrapes all chats into an sqlite db and does fts bruh

ohhhh

is this the beginning of the end...

kinda checks out with their recent compute struggle tho...
i really hope its temporary and they'll figure something out, i can't see myself paying even more than what i already am

it feels like they reduced the 100$ claude max limits significantly.

i never hit above 50-60% previously, even when working in parallel on multiple projects at once
now it's almost at 75% after like 2 hours of very moderate work...

i need 15k tps opus, ts would be so epic

i mean like, it fetched the reference source code three times in a row, but it wasn't enough apparently and it went to use grep.app mcp, failed bc of an overly complex query, fetched the same file a few more times, dispatched a subagent to fetch the fucking file.
like lol.

got around trying sonnet4.6 in a real task and uhh

its dumb. like really really dumb. it just. gets lost or something. and slow.

working with it feels like an overconfident and overly eager junior dev. and its *really* bad at following rules from CLAUDEmd.

bweh

i have more claude screen time than tiktok

lol

>The government is reportedly making the same demand to OpenAI, Google, and xAI. An anonymous Trump administration official told Axios that one of those companies has agreed, while the other two have supposedly shown some flexibility.

damn i wonder who /s

sysprompt more like sisprompt

opus4.6 truly is the new 4o