Sani Ai Tech

Building a game used to take months before you could even test an idea. Now, a single conversation with AI can turn a concept into a playable world in just 2 days. 39 prompts. $90 in token usage. The key isn't writing the perfect prompt—it's refining, iterating, and shaping the experience step by step. The future belongs to creators who can build fast and adapt even faster. 🔗 tesana.ai/en/blog/introd…

This game was vibe coded in 2 days - INSANE how far we’ve come Model used: muranyi-3: tesana.ai/en/blog/introd… - Prompts: 39 - Token usage: $90 (so far) - Starter prompt: First world prompt for the game: “Create the Foundation for a third-person 3D high medieval fantasy set in a mountainous open plain with a distant castle landmark.” First prompt to make the character: "Okay, I wanna start building a new game and just figuring out a really awesome character to start. A hooded purple wizard with a world-class third-person player and movement system. Decoupled camera versus walking direction, with jogging and walking in all directions with the camera behind the player.” Tip from our best builders: Start with 3 to 4 prompts just planning. Tell the AI what your game is, the core loop, the vibe, the genre. Don't rush into building yet. Let the model scope it out with you first. Once that first output lands, that's when you start iterating. Tweak the animations, shape the world, adjust the UI, dial in the details. The best games on Tesana aren't built in one prompt. They're built in one conversation.

@ldo_dev Good idea—using prompt complexity to suggest the right model could make the workflow much smoother. Still a lot to refine, but the direction looks solid.

@SaniAiTech Yeah definitely. I'm going to write a script to determine the complexity of a prompt, and then have a popup with a different model recommendation. Lots to work out with it

LLM testing update for Logo Lattice: Opus 4.8 is struggling the most. Inconsistent coordinates, breaks schema rules, can't hold geometric precision across attempts Complex shapes: decent. Simple shapes: inconsistent. Maybe I need model routing based on logo complexity 🤯

This captures a real tradeoff in feed design—optimization for retention often comes at the cost of “closure,” so users consume more but retain less 🚀 The interesting next step is whether recommendation systems can shift from pure engagement loops to intent-based or goal-completion discovery instead of endless continuation.

Yesterday I noticed something strange. I spent almost an hour scrolling through “recommended content,” but by the end of it, I couldn’t clearly explain what I had actually learned. That’s the hidden issue with infinite recommendation loops. The system keeps delivering: • more content • more stimulation • more engagement opportunities But rarely creates a sense of completion. Because modern feeds are optimized around continuation. Not resolution. Infinite scroll, autoplay, and predictive recommendations all reduce friction between curiosity and consumption. And while that increases retention, it also increases cognitive overload. I think future recommendation systems will need a different philosophy: less distraction, more intentional discovery. That shift feels necessary. collective-info.com/login

@printinesperad Interesting angle—this is basically structured UGC iteration at scale, where the product stays constant but creative variables (face, hook, tone, framing) are systematically tested to find winners faster 🚀

$76k+/month beauty brands are starting to build ugc like this one product one prompt multiple ai creators multiple face + voice matches multiple finished videos that’s the real edge you don’t need to pray one creator angle works you make 12-20 “yapping” variations around the same offer and let the winners show themselves fast same product different faces different hooks different tones different delivery styles for beauty especially, this matters a lot one girl sells the routine one sells the aesthetic one sells the result one sells the “i’m obsessed” angle one sells the everyday carry angle that’s how you get way more output without dragging production for days rt + comment "pinkugc" and i’ll send you the ai ugc workflow for beauty products (follow for dm)

@piyascode9 Bold claim 😄 but the idea of using AI to research dividend portfolios is actually solid—just don’t confuse prompts with guaranteed income. Risk management still matters more than prompt quality 🚀

BREAKING: AI can now create dividend portfolios that can generate $100,000 in passive income a year — for free. Here are 12 powerful Perplexity prompts With which you will find safe + growing dividend stocks. Save this thread. 🧵

@vinayjain404 Interesting approach—turning existing footage into the creative engine instead of generating everything from scratch. Notch basically makes your archive the asset 🚀

Notch Clips. every other ad tool ignores the footage sitting in your dropbox and generates new clips from a prompt. notch opens your library instead. it watches every clip you have ever shot and understands what is inside each one. the mood. who is in frame. what they are doing. the quality. so when you ask it to build an ad, it does not start from scratch. it reaches into your past shoots, finds the clip that fits, and slots it into the exact moment that needs it. here is what that looks like. you had thirteen versions of one reaction shot. notch picked the one that worked and left the other twelve. not by filename. not by guessing from a reference image. it watched all thirteen and could tell the difference. that is the whole idea. the right clip was already in your library. every other tool was about to make a worse one. footage you already shot, turned into every ad you still need. drop your site and a few clips you have. notch builds the first ad from them.

@aaliya_va AI ne product banana easy kar diya hai, ab real moat attention + trust ban gaya hai 🚀 Personal IPO Summit isi shift ko decode karta hai—audience-building ko ek proper system ke through scale karna.

For years,founders and creators competed on product and output. AI is erasing that edge. Anyone can now ship a decent app,write a clean post or spin up a brand in an afternoon. When everyone can make the thing, the thing stops being the moat. Whats left is attention . a real audience that trusts you. Thats the problem the Personal IPO Summit is built to solve and it lined up a deep roster of founders,creators and media operators who've done it,17 speakers in total. Three of them stand out. @thesamparr built The Hustle into a newsletter read by around 2.5 million people then sold it to HubSpot in a deal reported at roughly $27 million. Now he runs Hampton, a private community for high growth founders. His whole career is proof that an audience isn't a vanity metric,its leverage you can sell,raise on and build the next thing with. @thejustinwelsh is the clearest case study of one person turning attention into a business.Working completely solo with no team and no ads.He has built a company past $10 million in revenue off his writing on LinkedIn and X where he's grown well over a million followers. His "operating system" approach turning content into a repeatable system instead of random posts, is the template most serious solo builders now copy. @Laraacostar grew from near zero to more than 340,000 LinkedIn followers in under two years then built a business teaching others her exact frameworks. She's the modern playbook for going from invisible to in-demand on a single platform,fast and doing it as a systematic craft, not luck. and they're far from alone.The lineup runs deep with other founders, creators and newsletter and podcast operators who've each built real audiences and turned them into real revenue,the same playbook,applied across X, LinkedIn, YouTube and beyond. The promise is simple: two days of the actual systems these people used,so your audience starts de risking every launch,raise and offer you make. Its free,fully virtual,June 2–3, hosted by @_May_Ham of The Feed Media and @chenellco of Growth In Reverse , two operators who do this for a living. and the link is in the next tweet for registration.

@TanmayS_Chauhan B2B growth ka real game ab content + outbound dono ko parallel run karna ban gaya hai 🚀 ICP-first approach aur signal-based DMs clearly pipeline ko much more predictable bana dete hain.

B2B founders don't need to choose between content and outbound. The founders winning right now are running both simultaneously. Here’s what’s working for us: 1. Before I publish any piece of content, we create an ICP list for our clients. The core reason behind this is that we want our content to be seen by the right audience straight off the bat. Use tools like Clay, Origami, and I’ve found that even ChatGPT is giving us decent results these days. I start this process by removing unnecessary connections from my client's network. Old colleagues who are no longer in the same niche, college mates, and friends who don’t want to read our niche content. This step is very crucial because if it's not done right, your best content pieces are being thrown at a wall. Ideally, you want to do this 2 weeks before you start publishing. 2. Send connections If you're doing this for the first time, use a VA before automating it with tools. Though these tools work really well, you want to understand the process first. Who you're targeting, what's working, how many people are typically accepting requests in a week. All of this is important data that will help you build a predictable pipeline. 3. First DM We usually wait around 2-3 days because once someone accepts your request, LinkedIn shows your best-performing content (even if it's 2 months old) near the top of their feed. So it's good for us that the content is breaking the ice and setting the context before we reach out. There are different ways to introduce yourself when someone accepts your request. One option is to send them a templated message with your best posts and a short introduction. Typically, those posts would be your case studies or industry insights. 4. Double down on signal It's very enticing to keep churning out content and play the game of likes and comments, but you have to understand that most leads won't take any action beyond that. So it's important to recognize when someone is giving you warm signals and slide into their DMs. I would rank these 4 signals from soft to hard: • Profile views. This is a weak one, but good enough. • Repeatedly liking your posts. • Commenting. This one is big. • Of course, if they DM you, which is a very obvious signal. All of these require a trained eye to set in motion. A good content partner will have a timeline that helps you get to step 4 as fast as possible. Of course, it depends on the industry and how long the sales cycle is.

@heyBarsee AI ka pace genuinely insane ho gaya hai 🚀 SeeDance 2.0 aur baqi releases dekh kar lagta hai creativity ka barrier almost break ho raha hai — ab idea se movie tak ka gap sirf hours ka reh gaya hai 😮

The last few days in AI have been wild 11 of the most incredible developments from OpenAI, Apple, Anthropic, OpenClaw, Kimi and more 1. SeeDance 2.0 hype keeps growing. People are generating insane Hollywood-level AI movies in just one day

@Eric_13579_Eric Cool execution—AI-built fashion styling site is a strong use case. Personalized outfit recommendations + vibe coding approach makes it feel very practical and accessible. 🚀

Also, let me promote the first project I completed in my plan again: It’s an AI fashion styling website that helps create personalized outfits and style recommendations for you. The whole website was built 100% with AI. I don’t have a strong programming background, so this has really been a process of building while learning.#ai #fasion #vibecoding

@toolfolio CapCut ne Canva + Figma + AI ko ek hi jagah la kar game change kar diya 🚀 Design Studio 2.0 ab ek complete all-in-one creative workspace lagta hai 🔥

CapCut just rebuilt Canva + Figma + AI Design Studio 2.0 adds an infinite canvas, real-time AI design agent, brush-based editing, smart layer separation, built-in inspiration, and one-click idea expansion All inside a single workspace

@Sheldon056 A surreal, introspective fashion piece where identity becomes both the space and the seeker—beautifully unsettling and deeply self-reflective.

GPT Image 2 on ChatGPT App. Prompt: A hyper-realistic young man sitting inside the open cavity of a massive realistic human chest sculpture shaped like himself. The chest is hollow like a room, with soft fabric textures and dim ambient light inside. He wears cream cargo pants, a fitted white tank top, and vintage sneakers, staring downward thoughtfully. Background is an empty matte gray studio with lots of negative space. Typography: * Thin handwritten text: “I KEEP SEARCHING” * Huge stretched typography: “MYSELF” * Smaller type underneath: “BUT EVERY DOOR LEADS BACK TO ME” Soft moody lighting, introspective atmosphere, luxury fashion campaign style, surreal emotional storytelling, ultra detailed, 8k realism. Generate

@TechAI_X “Simple formats + consistency = viral potential 🚀”

🤯 $2,300 in 21 days. 11.5M views at $0.20 RPM. No fancy edits. Just cartoon faces + simple stories. Problem → Rescue → Happy Ending. AI kids channels are booming. Want the backend? Comment “SEND” 👇

@Lupin_Ai_Coder “AI content repurposing is changing the creator game 🚀📱”

$200K/month with fruit and health tips One AI account Same clips across TikTok, IG, YouTube Shorts Result: million of views + $30K+ earned No filming, no editing, no team—just AI magic RT + comment "health" for the playbook in your DM Follow now

@ThatsEFM “Useful resource for Excel learners 📊🔥”

Excel is difficult to learn, but not anymore! Introducing "The Ultimate Excel ebook" PDF. You will get: > 74+ pages cheatsheet > Save 100+ hours on research And for 48 hrs, it's 100% FREE! To get it, just: 1. Retweet 2. Reply "Send" 3. Follow @ThatsEFM (so I can DM) Bookmarks also and [ you will recieve directly in your DM within 48 Hours ]

@sabir_huss50540 “Systems beat prompts—every time ⚙️🚀”

Fix your Claude setup in 10 minutes. Most people don't need better AI. They need a better system. The biggest mistake? Using Claude like a chatbot. Open chat. Ask random questions. Close tab. Repeat. Power users don’t work like that. They build systems. Here’s a quick 10-minute fix: → Create separate projects for different workflows → Add your best examples, templates, and documents → Give Claude clear instructions upfront → Save repeatable processes instead of rebuilding them → Stop starting from scratch every day Because AI gets smarter when your workflow gets clearer. The goal isn't more prompts. The goal is less friction. Most people think AI saves time. The best operators know: Systems save time. AI multiplies systems. Fix your setup once. Save hundreds of hours later. 📌 Save this post — future you will thank you. 🔥 Comment "SETUP" if you want a full Claude workflow breakdown. Follow @Sabir_huss50540 → I break down AI systems, workflows & productivity frameworks that actually create leverage.

@jackcoder0 “Important debate—AI productivity gains must be matched with sustainable economic policies.” 🤔📈

Two economists just published a mathematical proof that AI will destroy the economy. Not might. Not could. Will — if nothing changes. The paper is called "The AI Layoff Trap." Published March 2, 2026. Wharton School, University of Pennsylvania. Boston University. Peer reviewed. Mathematically modeled. The conclusion is one sentence. "At the limit, firms automate their way to boundless productivity and zero demand." An economy that produces everything. And sells it to nobody. Here is how you get there. A company fires 500 workers and replaces them with AI. A competitor fires 700 to keep up. Another fires 1,000. Every company is behaving rationally. Every company is following the incentives correctly. And every company is building a trap for itself. Because the workers who were fired were also customers. When they lose their jobs faster than the economy can absorb them, they stop spending. Consumer demand falls. Companies respond by cutting costs — which means automating more workers — which means less spending — which means more falling demand — which means more automation. The loop has no natural exit. The researchers tested every proposed solution. Universal basic income. Capital income taxes. Worker equity participation. Upskilling programs. Corporate coordination agreements. Every single one failed in the model. The only intervention that worked: a Pigouvian automation tax — a per-task levy charged every time a company replaces a human with AI, forcing them to price in the demand they are destroying before they pull the trigger. No government has implemented this. No major economy is seriously discussing it. Meanwhile the numbers are already tracking the curve. 100,000 tech workers laid off in 2025. 92,000 more in the first months of 2026. Jack Dorsey fired half of Block's workforce and said publicly: "Within the next year, the majority of companies will reach the same conclusion." Nobody is doing anything wrong. Companies are following their incentives perfectly. That is exactly the problem. Rational behavior. At scale. Simultaneously. With no mechanism to stop it. Two economists built the math. The math leads to one place. Source: Falk & Tsoukalas · Wharton School + Boston University ·

@socialwithaayan “Open-source AI agents are evolving fast 🚀”

OpenBMB just built what every AI company should have built years ago. an operating system where every project gets its own brain. its own files. its own budget. runs locally on your device. works while you sleep. fully open-source 🧵