smuv

ai can do pretty much anything these days. most people still have no idea how to turn that into money.

[smuv @smuvnerd]

@smoothynerd building is easy now. getting people to care is the hard part

[smuv @smuvnerd]

@nexor404 does this mean jobs in economic security at banks and in government could soon become obsolete? pretty interesting direction for ai to be heading in

[smuv @smuvnerd]

HOLY SH*T. THIS IS WHAT HAPPENS WHEN AI STARTS TURNING IDEAS INTO ACTUAL PRODUCTS.

13 completely different projects. One Claude model.

and some of these demos are genuinely hard to believe.

Nick Saraev just showed off 13 things you can build with Claude Fable 5.2, and we're getting dangerously close to the point where having an idea is the easy part.

just LOOK at what he pulled together:

— a playable 3D game with combat, movement, and actual gameplay.

— a desktop-style environment with windows, notes, a terminal, and even a mini-game.

— interactive 3D websites with insane lighting, glass effects, and smooth animations.

— a mechanical ocean diorama with moving gears and animated details.

— a digital weather station that looks like a real physical device.

— interactive nature scenes with detailed 3D environments.

— a virtual pop-up book with animated pages and illustrations.

and that's not even the whole video.

a few years ago, you'd need serious coding and design skills to pull off even ONE of these.

now imagine what a solo builder could create with these tools and a free weekend.

of course, turning a cool demo into a polished product still takes work.

but the gap between having an idea and actually building it is shrinking FAST.

and that's a pretty big deal if you've been sitting on ideas because you thought you didn't have the skills.

WATCH THIS VIDEO. seriously.

one of these 13 projects might be all the inspiration you need to finally start building something of your own.

the biggest mistake right now might be watching everyone else build while you're still waiting for the perfect moment.

follow for more alpha and put it to work. the tools are already here.

[broke boy @0xfuckpoverty]

Claude Fable 5.2 is the strongest AI model you can use today, and most people waste it:

Too scared of the price, or burning limits on easy jobs.

Here's the setup that gives you Fable's brain on a Sonnet bill, plus 10 copy-paste prompts: https://t.co/vUyYC12vPG

[smuv @smuvnerd]

@smoothynerd damn, think I could pull this off at my furniture factory in Germany?

[smuv @smuvnerd]

@kursormaxx crazy result man

[smuv @smuvnerd]

@0xZenad the scary part is how long it takes to even realize it's fake

[smuv @smuvnerd]

@Wombaflow cutting 20 hours of work down to 2 is actually wild, imagine what this looks like a year from now

[smuv @smuvnerd]

@kumo0x that shark looks so real

[smuv @smuvnerd]

@slash1sol @n8n_io the industry's growth is absolutely insane

[smuv @smuvnerd]

@kumo0x dope, she did it

[smuv @smuvnerd]

HOLY SH*T. The CEO of Anthropic just made one of the wildest predictions I've heard about AI.

Dario Amodei believes cancer, Alzheimer's, heart disease, and even mental illness could become things of the past in LESS THAN 10 YEARS.

Yes, you read that right. Before 2036.

And he's not talking about slightly better treatments or faster diagnoses.

He's talking about AI helping humanity solve medical problems we've been fighting for generations.

His exact words:

"If we do this right, we can create an age of miracles."

Think about that for a second.

Diseases that have taken millions of lives. Decades of research. Billions spent trying to find answers.

And now the guy leading one of the world's biggest AI labs believes we could make more progress in the next decade than most people would dare imagine.

Of course, it's a prediction, not a guarantee. Clinical trials, safety, and biology aren't things you can just speedrun with a better model.

But if he's even HALF right, the 2030s could look completely different from the world we're living in today.

We're spending so much time arguing about which AI writes better code that we might be missing the much bigger picture.

AI might end up changing how long we live, not just how we work.

And I genuinely don't think most people are ready for how fast the next few years could move.

Follow along. I'll keep digging into the AI breakthroughs worth paying attention to. You really don't wanna sleep through this decade.

[NIK @ns123abc]

Anthropic CEO predicts cancer, Alzheimer's, heart disease, mental illness will be cured in less than 10 years

ACCELERATE

[smuv @smuvnerd]

here's the game

https://t.co/gwT6F0Cfyx

[smuv @smuvnerd]

WHAT THE F*CK. SOMEONE BUILT THIS GAME IN JUST 24 HOURS.

And you can actually play it.

It's called Sumiro, and @xikhar built the whole thing in a single day using AI.

Just look at the footage. The environments, the lighting, the character animations, the way the world comes together. This looks like something an indie studio spent months working on.

And there's actual gameplay here. Platforming, parkour, puzzles, enemies, and multiple levels. Not just a pretty 3D world you can walk around in.

Here's how he pulled it off:

He started with a video generated in Flux 3 Max, then used it as the visual reference for the entire game.

GPT Image helped design the characters and levels. Tripo turned the character into a 3D model. Blender and Three.js brought the world into the browser.

Meanwhile, Claude Code with Opus 5.5 handled the coding, physics, animations, and a bunch of the building and debugging through parallel agents.

He kept playing the prototype, finding problems, and sending the agents back to fix them.

All of that in ONE DAY.

Sure, it's still rough in places. Even the creator admits the music needs work.

But if this is what one person can put together in 24 hours today, imagine what we're gonna see a year from now.

And btw, the gameplay footage was recorded by Claude itself. That's an AI playing the game, not the developer.

You can play Sumiro yourself. I dropped the link in the comment below.

[Shikhar @xikhar]

I made a full original game in 24 hours using AI

Stack: Opus 5.5 + Flux 3 Max + GPT Image + Tripo + Blender + Three.js

Anyone can make games now. Full workflow thread and prompt below:

(1/7)