AI Workflow Lab (@AIWorkflowX) header
AI Workflow Lab (@AIWorkflowX) profile picture

Practical AI for work, content & online business. Testing tools, workflows and ideas. Sharing what works - and what doesn't

Czech Republic
Joined October 2026
Claude built a mechanical watch from a parts list, and the seconds hand only moves because the gears push it. No battery, no chip, 97 parts. The barrel has 80 teeth, the center wheel turns once an hour, the third wheel has 75 teeth and the fourth turns once a minute. 80/10 x 75/10 = 60, so the small seconds is right by arithmetic, not by keyframes. The balance swings 5 times a second and the escape wheel steps 12 degrees on every beat. That's 18,000 beats an hour, and every tick in the sound lands on one of those steps.
2
1
3
66
AI Workflow Lab retweeted
N3on took the plumber mission in GTA 6 and we all know those pipes weren't the only thing getting fixed Be honest, has a "friend" ever called you over at night to "check her pipes"?
5
5
8
221
AI Workflow Lab retweeted
GTA 6 lets you own a dog You can praise it. You can scold it. You can even get a car covered in paw prints. Rockstar thought of everything
9
6
18
480
AI Workflow Lab retweeted
A giraffe's neck has 7 vertebrae. That's the same number you have. I built a little wildlife viewer where you can see what moves each animal. There are 5 of them: a tiger, a lion, a plains zebra, a giraffe and a bald eagle, with 4 clips each. The big cats walk, run and roar with the jaw dropping open. The zebra puts its head down to graze and the giraffe stretches up for leaves. The eagle calls from an old log, then takes off and soars. A few times I flip on wireframe or skeleton view, so you can see the mesh and the bones under the skin. Which animal should I rig next?
5
2
8
307
AI Workflow Lab retweeted
Xbox One owners tomorrow: 8.4 GB down, 210 GB to go, 837 days left. two jet engines bolted on for cooling and it's still on fire. small detail, GTA 6 isn't even coming to Xbox One. PS5 and Series X|S only. what are you playing it on?
8
2
12
546
AI Workflow Lab retweeted
Ready to give your entire 2027 to $GTA 6? My AI agent already cleared my calendar Now it’s asking for its own controller
GTA 6 is HERE! You can already download it on your PlayStation 5 Be honest… are you really ready to dedicate your entire 2027 to this game?
4
1
22
601
Same seagull, same code, three render budgets, GTA The bottom one is 691 triangles with flat shading and no shadows. The middle one is 100,176 triangles with one smooth shader and hard shadows The top one is 158,904 triangles plus 12 layers of feathers, soft shadows, and depth of field on the skyline. Going from 2013 to 2026 only added about 59K triangles. Most of the jump is light, feathers and the water
6
1
8
304
AI Workflow Lab retweeted
$NVDA tonight is the one I'm watching. it's basically the AI trade's report card. ~$109B expected on a $108B guide, margins near 74%, and the first quarter without Jensen. if data center growth even looks like it's slowing, every AI name feels it by morning. beat or miss?
Packed calendar and it starts tonight. Today: $NVDA reports after the close, first quarter without Jensen Huang. Street wants ~$109B on a $108B guide, margins near 74%. Tomorrow: FOMC minutes from the October hold. Cook and Hammack already wanted a hike, i wanna see how many more did. Thursday: Trump, Xi and Putin meet in Shenzhen. First three-way sit-down between them, agenda still not public. Nov 25: October PCE. Dec 9: Fed decision. Kalshi has a hike at 96%. Dec 10: November CPI, a day after the Fed already decided. Jan 3: new Congress, Democrats run both chambers. BTC already swept 62K on the way here. Which one moves it most?
4
1
6
275
I turned this calendar into a domino run, one tile per day from Nov 17 to Jan 3, so 48 tiles in total NVDA, the Fed minutes and the summit knock over the first 3. After that it’s 5 quiet days until PCE, then 13 more until the Fed The pair that got me is Dec 9 and Dec 10. The Fed decides at 2:00 PM ET, and November CPI comes out at 8:30 AM ET the next morning, 18.5 hours later. So the last rate decision of the year gets made without November’s inflation number The BTC coin at the end of the run is still standing after all 48 fall
Packed calendar and it starts tonight. Today: $NVDA reports after the close, first quarter without Jensen Huang. Street wants ~$109B on a $108B guide, margins near 74%. Tomorrow: FOMC minutes from the October hold. Cook and Hammack already wanted a hike, i wanna see how many more did. Thursday: Trump, Xi and Putin meet in Shenzhen. First three-way sit-down between them, agenda still not public. Nov 25: October PCE. Dec 9: Fed decision. Kalshi has a hike at 96%. Dec 10: November CPI, a day after the Fed already decided. Jan 3: new Congress, Democrats run both chambers. BTC already swept 62K on the way here. Which one moves it most?
4
1
5
126
AI Workflow Lab retweeted
Tonight’s $NVDA report is basically the invoice for every AI agent launched this week. $109B expected. Let’s see who’s actually paying Fed sets the price of money. Nvidia sets the price of compute. BTC has to trade both by Friday. Everyone’s watching the Fed. I’m watching one thing in Nvidia’s report: how much of that $109B comes from customers who have revenue of their own.
Packed calendar and it starts tonight. Today: $NVDA reports after the close, first quarter without Jensen Huang. Street wants ~$109B on a $108B guide, margins near 74%. Tomorrow: FOMC minutes from the October hold. Cook and Hammack already wanted a hike, i wanna see how many more did. Thursday: Trump, Xi and Putin meet in Shenzhen. First three-way sit-down between them, agenda still not public. Nov 25: October PCE. Dec 9: Fed decision. Kalshi has a hike at 96%. Dec 10: November CPI, a day after the Fed already decided. Jan 3: new Congress, Democrats run both chambers. BTC already swept 62K on the way here. Which one moves it most?
15
8
1
30
4,686
AI Workflow Lab retweeted
And here we go…GPT-6.2 Astra and GPT-6.2 Sol are yours to make more games with. Oh, and they’re also super efficient and really good at making slides, building websites, and tackling the work you were doing before you started making games. Rolling out today to ChatGPT Work and Codex for Plus, Pro, Enterprise, and Edu users. openai.com/index/introducing…
🤖 Made with AI
519
1,142
439
12,709
844,952
T-RX4 is a 6.8 m robot dinosaur I built with the prompts from this article Every number on screen comes out of one function of the gait phase. I checked one by hand: at 1.3 m/s with a 1.72 m stride it does 0.76 cycles a second, 2 steps each, so 91 steps a minute, and the panel says 91 LEGS pulls the shells off the knee actuators. POWER opens the 48-cell battery rack and HEAD drops the jaw. TAIL shows the 11-segment counterweight swinging against each step. At the end the slider pushes it to 10.1 km/h and the stride stretches to 2.54 m Before the 1,500-frame render, the test stills caught 4 bugs. A shadow was cut off at the edge of the light, the head sat behind a panel, one label was hidden under another panel, and a title was printed over its own subtitle. The soundtrack thumps on each of the 38 footfalls the model computed
3
8
346
AI Workflow Lab retweeted
Holy f*ck, Opus 5.5/Fable 5.2 can now remake almost any animation you see with these 15 repos FFmpeg. Lemo-Opuscar. HyperFrames. Remotion. Blender MCP. 99 film styles, 120,000 particles and a Blender that takes orders, in one list Everything below is free: FFmpeg → takes any video apart, shot by shot github.com/FFmpeg/FFmpeg Lemo-Opuscar → 99 film styles your agent can shoot in code github.com/lemomo-ai/lemo-op… EffectCraft → an open-source motion graphics studio github.com/storytold/effectc… PixiJS → real pixel art, rendered on the GPU github.com/pixijs/pixijs canvas-sketch → generative art that loops forever github.com/mattdesl/canvas-s… HyperFrames → your agent's HTML, played as a video github.com/heygen-com/hyperf… Remotion → videos written in React github.com/remotion-dev/remo… Remotion Skills → the rules that make agent videos move right github.com/remotion-dev/skil… Blender MCP → Blender builds whatever your agent asks for github.com/ahujasid/mcp-for-… Three.js → real 3D in the browser github.com/mrdoob/three.js GSAP → letters that fly, bounce and land github.com/greensock/GSAP regl → 120,000 particles in one draw call github.com/regl-project/regl Unslop → turns AI-looking UI into real design github.com/heydamir/unslop Impeccable → design polish, one command at a time github.com/pbakaus/impeccabl… Anthropic Skills → Claude's own skills, generative art included github.com/anthropics/skills $7,500 → $0 Send the list to your Claude Then read how motion engineers check every frame in the article below ↓
Your Opus 5.5 video is not bad Your checks are Everyone posts what Opus 5.5 can make. Almost nobody posts what it got wrong I went through 10 of my films: 100+ defects, 10 classes, one check for each ↓
17
20
1
320
21,317
AI Workflow Lab retweeted
Haven't bought GTA 6 yet 🌴? Build your own city instead I made a neon Miami police chase with Claude, all in a single HTML file: Night-time Miami coast with palms, art-deco hotels and glowing neon signs Police cars with flashing sirens chasing a red sports coupe Helicopter camera with a live HUD, minimap and news ticker A roadblock the cops crash into A jump over an open drawbridge at 180 km/h No game engine. No downloaded assets. Everything, from the buildings to the traffic, is generated by code and opens right in your browser. The wait for the release is a lot more fun when you build something yourself. What city would you make? Prompt in the comments 👇
24
13
41
1,228
AI Workflow Lab retweeted
I built an F-35 flight in pixel art, and the cockpit view is the part I'm proudest of. The real F-35 has no head-up display. Pilots see their speed, altitude and heading inside the helmet, so that's how the cockpit works here too, with the big wide screen below it. It starts from above over a chain of islands. Then the afterburner kicks in and the jet climbs into a full roll. After that you drop into the cockpit for a turn over the sea. Which jet should I fly next?
7
2
13
455
Three bears from one template of 21 bones and 28 shapes, built by Fable 5.2 The penguin in my article was a single model and it did not look much like a real penguin. So this time I asked Fable 5.2 for something harder, the bear family from the fairy tale with their den and proper fur It wrote one bear as numbers, then scaled it to 1.0, 0.82 and 0.42 for papa, mama and the cub. Each body has 13 layers of fur on it. The den is a rock mesh of 189K triangles with three straw beds inside The first test frames were rough. Mama was sitting in mid-air and the cub’s nose was buried inside his muzzle. The tree crowns looked like umbrellas, so they got cut. All of that was fixed before the 576-frame render
3
1
8
217
AI Workflow Lab retweeted
This watermelon slice is made of jelly, and I cut it up on camera. A star cutter goes in first. Then the knife makes five slices and one long cut across, so the slice ends up in 13 wobbly pieces with 19 seeds spread between them. After that they get tossed twice, and every piece bounces when it lands. What should I turn into jelly next?
4
1
11
470
Black Friday is 12 days away. Here’s how I’d use AI for it, based on what tests actually show: → Paste a suspicious text into a chatbot before you click. I tested ChatGPT, Gemini, Grok and Claude on 10 texts (6 scams, 4 real). All four got 10 of 10. Small test, but it’s a free second opinion. → Give it a budget, not “find the best one.” In a Cisco and Carnegie Mellon study, 8 of 13 AI models picked pricier options for users who seemed rich. Adding a number like “under $200” nearly fixed it in most top models. → Don’t trust its prices. In a small Core dna test, ChatGPT got 13 of 38 Black Friday prices wrong. Check the store’s own page. → If an agent buys for you, use a card with a low limit. ChatGPT Wallet has no spending limits of its own yet. And one thing never to do: paste your card number or passwords into a chat. Save this for the 27th. Which of these do you already do? Sources: my own test (Nov 15), Cisco & Carnegie Mellon, Core dna, OpenAI #AI #BlackFriday
10
2
14
378
AI Workflow Lab retweeted
This sphere spins in both directions at once, and your brain just picked one without asking you. Stare at it and try to flip it. The dots never change. What changes is which reading your visual system commits to, and once it commits it will defend that reading until something forces a switch. Language models do the same thing with an ambiguous prompt. They pick one interpretation early and build everything on top of it. The difference is you can feel yours flip. Which way was it spinning when you first looked?
8
2
16
185