opus 5.5 helped me build the best pdf parser i cant imagine i can make it myself, will be included in spokio for mac v1.13 soon next week 🚀😍 spokio.pro

no matter how many apps u have, 1 star reviews always get into your head 😌

i'm sick of the trend many fake posting their accounts terminated by apple 🫤

when learning to code not important anymore i will learn design 🤓

i guess waiting from 15 to 20 days for new app review is the new normal 🥲

omg i hope wont receive one of these 🤯

be extremely careful when using AI these days, the moment u trigger a safeguard like security or abuse you'll be flagged and will get banned anytime, and if they map the ban to your name on credit card too then hard to sub another account. should use disposible accounts for non-standard work like reading logs or experiments.

never trust these lower tier model and let it touch your codebase no matter how high the benchmark, there's a reason it's cheap because it's very lazy in reasoning and dont perform enough checks. for weak models like luna and terra they need very detailed prompts to complete a task, so they need a lot of hand holding if u want to use in real projects, the moment you're lazy and give them a short prompt they will slip and mess your codebase with thin implementation.

the more u work with 6.1 sol the more u realized why openai giving out so many resets recently, because the model is stupid at heart, so bad in understanding instructions let alone implementing, and always making bad assumption and never bother to ask for clarification.

hey @ChatGPT stop downgrade my chosen reasoning effort automatically in work mode, it's annoying to keep switching back and sometimes i didnt notice low effort messed up my codebase 😳

HBO is furious right now lol, imagine new series so flop and not even as good as AI generated female malfoid 🤣

never trust low effort models like luna or terra on any tasks touching your codebase, they are extremely lazy models, they dont check or verify unless un specically ask, most of implememtation will be thin and shallow or just happy path, dont even trust a simple task when they done, must review and verify result if u have to use them.

med tracker is the new habit tracker, seem like all vibe coders building 🤷

it's so cheap for a frontier ai company to give out resets every 2 days like that to cover the inability to make a better model 🤨

i feel so annoyed everytime AI says "fair point" when i point out its bs

the best way to waste time in 2026 is having unlimited frontier AI tokens and build a lot of internal tools and useless apps, eventually u get tired of asking agents to build even tho they can absolutely build anything you ask, now you're the absolute bottleneck of your future.

stupid 6.1 sol work mode just saying working with no feedback for 5 minutes before spit out a work, yikes

when the world already put AI's focus on security, there will be new vulnerabilities and attacks happen everyday. be care with what u store online both personal and user data. 🫤

the most annoying thing in the world right now is talking to a slow model 😠

i cant stand stupid and slow gpt models anymore 😳 cant wait for the sub expires 😑