Levy

AI-first builder. AI Agent Native Apps

[Levy @_levyathan_]

So people said GLM-5.2 from @Zai_org is good at design, I just prompted: "make a cool looking notch menu with cool animations"

I must say, it's not bad at all based on the basic prompt I gave it #zai #GLM5

[Levy @_levyathan_]

@jacobrodri_ Sometimes on Tuesday's

[Levy @_levyathan_]

@testingcatalog I thought that this is a joke again, but the cat is actually there 🤣😂

[Levy @_levyathan_]

@petergostev It's so good that the percentage fell of from the page

[Levy @_levyathan_]

@AlexanderKnigge

[Levy @_levyathan_]

@orcdev Sitting down with your son with a home made meal is a much better experience than going out

[Levy @_levyathan_]

@TheGameVerse Can I have a slice?

[Levy @_levyathan_]

@jjacky @OpenRouter Yeah I can imagine the costs, plus mash up a couple of AI models so an AI model can judge the best result.

Yeah seems promising

[Levy @_levyathan_]

@maverickecom 550 videos/day?
Yeah right

[Levy @_levyathan_]

@Mr_cryptodude Sad thing is with web design they use a youtube tutorial follow it post the end result and they post it here "oh look one shot" its BS plus the comments under the video explains everything

[Levy @_levyathan_]

@orcdev so it just read your mind ? "make my dream game"

wow..

[Levy @_levyathan_]

@AishwaryaDevv Tell it to make you a photo editor, post it on X saying "WOW I just killed Adobe Photoshop with Opus one shot prompt gave me this"
Gain traction, benefit from it, live with the consequences through your entire life

[Levy @_levyathan_]

@spoobsV1 yeah right, one prompt

[Levy @_levyathan_]

@Teknium I agree, and I'm glad someone finally said it. Since the Opus releases, Anthropic seems more focused on locking everything down than advancing what's actually useful. A lot of self-congratulation, not much ambition.

[Levy @_levyathan_]

@jacobrodri_ @progysteto a 1v1 video platform, you get paired with strangers and AI and users rate your appearance, its an Elo ranking system

[Aoden Teo @AodenTeoMT]

Today, we’re excited to introduce Miso One, the most emotive voice model in the world.

Miso One is an 8-billion-parameter text-to-speech model for highly expressive speech generation. It emotes like a human and responds faster than a human, with just 110 milliseconds of latency.

We’ve open-sourced the model weights, with API access coming soon.

Hear how Miso One sounds in the thread below.

[Levy @_levyathan_]

@Teknium built something like that way back: https://t.co/ZFU74BIwdj I can always upgrade if needed

[Levy @_levyathan_]

You can now claim your own handle on @cursor_ai.

Get yours now: https://t.co/2AnIHTwFl9

[Levy @_levyathan_]

@arena User experience is more important, not the "trust me bro" benchmarks