Building a 0-human company in public. I test AI agents, automation stacks, and enterprise workflows so you can see what actually works. FAANG | Future of AI

Dubai
Joined December 2024
Building a 0-human company in public. I test AI agents, automation stacks, and enterprise workflows to see what actually works. Here I share: 1) agent experiments 2) tool teardowns 3) research notes 4) failures and lessons Follow if you care about autonomous companies and AI agents.
1
5
550
If it takes 1 hour to generate 1 image then it is not unlimited. Do companies understand this ?
1
88
True !! When we got the annual plan we didn’t have this info. @suno you can’t be doing this to us
I’m SO PISSED about what @suno just pulled on us. They suddenly change the ToS, and now, just because they fucking feel like it, I can only download 60 songs a month. Excuse me? I signed up for the ANNUAL PLAN for a reason It has to be one of the dumbest moves in history.
2
5
586
I was about to get this sub but then I saw MJ Niji being offered. As far as I know MJ doesn’t give away API access. Can someone confirm ? @DavidSHolz
Seedance 2.5 UNLIMITED up to 7 days on Buzzy. Live now. In 7 days, at ZERO credit cost: 1. Direct first Hollywood-ready film and attract Millions of views on social 2.Launch and grow a content channel from the ground up 3.Plan and execute an end-to-end marketing campaign 4.Build a standout showreel and land your first client For a limited time only.
1
1
1,506
There is still no nano banana pro moment for video generation yet. Nano banana pro gives you results in 1 shot. But even seedance 2.5 you find issues with physics
3
229
I’m laughed so hard on this !!
1
160
There will be a time in 2-3 years you will have LLMs in your brain as a chip and then you will have your brain + an LLM with 1 tillion token context. @OfficialLoganK this is what I want 👆
119
I had 6 fingers until Nano Banana 2.5 came out in 2025. Now I have 5 consistently on each
1
151
Is eveyone still a prompt engineer or did they graduate to Agentic builder ?
172
In hindsight everything is possible. Please focus on the present
165
Grok 4.7 in 4 weeks.
2
189
Moderation level at @grok and @imagine seems worse than Google these days
1
9
496
The 5 fingers at the end is a flex of a good model
I love the smell of Flux in the morning.
1
2
361
It’s done now !! People in states have easy access to money. I hope agents are good at technical analysis
Robinhood is now open to AI agents. You can open an agentic account and connect an AI agent to it. From there, you can let it research, trade, and manage a portfolio on your behalf. Here's how to get started. 🧵
228
Grok for excel. PPT and Word. You can finally adjust images on word. Such an opportunity for copilot to lose here.
Grok for Excel is live. Use Grok 4.5 to build financial models, analyze market data, and generate charts and graphs. Try it now x.ai/grok/excel
213
Anyone who makes or solves any use case from this do let me know. I can’t think of any use case.
Replying to @openshipio
OpenShip is built by the community, for the community. We welcome contributions of all kinds - code, bug reports, feature requests, documentation improvements, and feedback. github: github.com/oblien/openship
1
203
How many things did Claude or codex kill this week ?
1
184
1 prompt 2 platforms see the results 1) Grok Imagine 2) Google Omni
🤖 Made with AI
3
470
Prompt: Create a 15 second ultra photorealistic 16:9 landscape video that looks like authentic spontaneous smartphone footage of a highly skilled young male tricking athlete performing a continuous sequence of extreme aerial acrobatics outdoors. LOCATION: A quiet suburban American public park on a bright afternoon. Slightly worn orange red outdoor court, simple wooden rail fence, grass, mature trees, small houses and a partially visible parked red car in the background. Natural blue sky, strong realistic sunlight and accurate long shadows. SUBJECT: One athletic young adult man with short dark hair, wearing a loose brown patterned plaid style long sleeve shirt, relaxed light blue jeans and white sneakers. Maintain identical face, hairstyle, body proportions, clothing and shoes throughout. Clothing must react naturally to gravity, wind and rotation. 0–1.2s: Wide full body handheld smartphone shot. The athlete stands several meters from camera, focuses, bends forward, swings his arms and compresses his knees to prepare for an explosive jump. 1.2–5.5s: He launches upward and sideways into a powerful aerial trick. Perform a fast sideways-backward twisting rotation with a brief wide scissor-leg split in mid-air. Arms, shoulders, hips and legs move with realistic athletic biomechanics and angular momentum. He completes the rotation and lands firmly with natural knee compression and weight transfer. 5.5–8.5s: Using the landing momentum, he immediately rebounds into a second explosive corkscrew style backflip. His body becomes briefly horizontal, tucks into fast rotation, then opens naturally and lands with realistic gravity, speed and impact. 8.5–11.8s: He flows directly into a third combination: sideways takeoff, aerial rotation, a brief natural one hand-assisted cartwheel like inverted transition, then one final fast twisting flip before landing safely on both feet. Every movement must follow real world physics with no floating or impossible hang time. 11.8–14.6s: He absorbs the final landing, regains balance and casually walks a few steps toward the camera, showing subtle natural breathing and physical exertion. 14.6–15s: Hold briefly on the athlete, then naturally fade to black. CAMERA: Single continuous smartphone style shot using a 26–28mm equivalent wide lens at chest height. Keep the athlete's full body visible during all tricks. Natural handheld micro shake, subtle stabilization, small reactive pans and tilts from a real camera operator. No robotic tracking, impossible camera movement or dramatic cinematic zooms. Realistic directional motion blur and slight smartphone rolling shutter only during fast rotations. REALISM: Must look like genuine viral sports footage recorded by a real person. Natural daylight exposure, subtle automatic exposure adjustment, realistic skin texture, fabric wrinkles, cloth physics, shadows and foot to ground contact. Background objects remain completely stable with consistent perspective and scale. Avoid all AI artifacts: no face morphing, extra limbs, broken fingers, changing clothes, duplicated objects, background warping, flickering textures, rubber like anatomy, floating, teleporting, unnatural rotation, CGI appearance, artificial slow motion, excessive depth of field or overprocessed cinematic grading. Natural consumer smartphone image quality with realistic sharpness, mild digital compression and true to life colors. AUDIO: Synchronized natural outdoor ambience only: light wind, distant birds, subtle fabric movement, sneaker friction, realistic landing impacts and quiet breathing. No dialogue, voiceover or music. Aspect ratio 16:9. Duration exactly 15 seconds. No text, subtitles, logos or watermarks.
420
Never accept the 1st work presented by an LLM. Always ask it to refine it.
3
178
I do not think Anthropic consults claude before taking business decisions.
2
1
3
267