Jimmy Wu

@leodev @Replit @fence__post @simplygandan @daradoescode great 🎉🎉🎉

@Replit @fence__post @JimmyWuDev @simplygandan @daradoescode Free opus 🥳🥳🥳

Ready to have some fun building this weekend?

@leodev @unkeydev Make it like this so people will copy 100%

Top tier UX from @unkeydev

@leodev started my account back then too, met so much great people doing awesome things within the half year of this journey. never quitting.

started this account at 13, now i'm 14 met so much cool people, shipped products, learned way more than school ever taught me back then i was just another kid who played lots of video games. now i'm actually doing things with my life currently focused on: - scaling my existing projects - going deeper into AI/dev - building more oss projects not stopping 🫡

@leodev congrats man

Totally forgot But I reached 450 followers! Thank you to everyone! 🙏

@saltjsx @convex congrats

we got into @convex for startups

@nicdunz commands working??

I built a Telegram remote for Codex on your Mac. Text the bot. Your Mac runs the real Codex CLI locally and replies back. No hosted relay. No new agent platform.

@LexnLin actually, I think both are too great for me to pick one

which image gen do you like more? :)

@LexnLin what are you building

Reached my 5h limit in codex in 3.5h 😭 I'm gonna miss you codex for the next 90min (pro 20x plan)

@leodev @WisprFlow yeah, I’ve tried before. but i didn’t by because of the misunderstanding of the price getting it soon…

@JimmyWuDev @WisprFlow You should totally get it, if you use AI a bunch this will save so much time.

While you looksmax I talkmax:

@leodev @WisprFlow I checked the pricing again. I thought the plan was USD$200/mo... 🤣🤣 (NTD$200 is like USD$6) It might actually worth it if it's this price.

@JimmyWuDev @WisprFlow Yea, it's super worth it.

I'm going to make a course called "Vibe Code 101" It would be 100% FREE Teaching new people in the vibe code community how to start Anyone interested in contributing?

I like GPT 5.5 Not a lot better from 5.4 On my experience, it's not 2x expensive. I feel like it's just the same as 5.4 Also it feels faster. Maybe they really achieve token efficiency.

I hate Anthropic as a company. The most important part of a company is trust. I don’t think a company intentionally making moves and rules unclear is trustworthy. They still got the chance since the problem isn’t about the models.

@AliGrids @jmtrivedi love this actually. this really makes users feel the premium vibe since start

users might forget why they opened the app, beautiful trap for users @jmtrivedi

@leodev @joncphillips I think so, older than me as well

@joncphillips Is it older than me?

If you remember this, you are a real OG.

@kevinkern 5.5 is coming

So... gpt-5.4 isn't the latest frontier model anymore.

@bridgemindai apparently claude models are really expensive and anthropic isn't using their compute well

A Claude Pro user just got rate limited on Opus 4.7 after ONE prompt. One request. Limit hit. 10 hour wait for the next answer. $20/month to send one message every 10 hours. Who else is experiencing this?

still crazy looking back got a 200k+ view post, that's only the 9th post on my acc also just created this acc in less than 2 weeks...

@leodev yeah, can’t wait for 5.5

@JimmyWuDev I'd still advise you to wait a few days and get off it. I'm like 70% sure that tomorrow openai is launching gpt 5.5 and it will probably be as good as opus on UI or better

Crazy part is: I was gonna get a Claude $20 plan to use it for UI design And I'm pretty sure it's way higher than 2% users... Anthropic as a company is losing trust and has a bad way of communicating any changes Anthropic models are actually pretty good, both ethically and intelligently I really hope they stop doing stupid stuff like this