viso.ai

@Krishnaik06 Just ran a live computer vision and object detection workshop, so curious what you would make of a different path to the same result: describe the scene in plain language, no model training required. See it now: viso.ai

@Sentdex Curious what the Director of AI and Engineering at Lucky Robots makes of this: a vision system that understands a scene it has never seen, no training data, no model per use case. Adjacent to the consumer-hardware work you have been writing about. See it now: viso.ai

🎓Viso Now 101- How to get started! In this video, I walk you through the entire process. From creating an account to building your first AI application. You'll learn how to: ✅ Upload your first video ✅ Create your AI application in minutes ✅ Analyze more footage using the same app If you're just getting started with Viso Now, this is the perfect place to begin.

@pabbeel have long needed labeled data to make sense of a scene. Viso Now skips that step: describe what matters, and the system understands it. Curious what a roboticist makes of that claim. See it now: viso.ai

@ecsquendor . VGI-1 is the engine behind Viso Now, and it's built on a claim your panel would probably want to interrogate: that a system can infer context and reason about a scene it's never seen, without a labeled dataset or a retrain per use case. Viso Now, running on that engine, went live this week. I'd rather put whoever built VGI-1 in front of your panel and let them defend the architecture directly: how it handles zero-shot context inference, where that claim is most likely to break under adversarial cases, and what "understanding" means here versus pattern-matching at scale. If that's worth a real technical argument on air, I'll set it up. If you'd rather pressure-test it over email first, send me your hardest question.

📢 VisoTalks Corner. What made our first Build Vision AI workshop special wasn't just the technology. It was the conversations, the collaboration, and seeing people from different industries come together to build solutions that matter. This is just the beginning. 🚀 #AICommunity #BuildVisionAI #VisoNow #ComputerVision

Turns out people don't want another AI presentation. They want to build. 50+ builders. 20+ real-world Vision AI apps. One unforgettable night. This is just the beginning. 👀 #BuildVisionAI #VisoNow #AI

🏆 Congratulations to the winners of our first Build Vision AI workshop! Use Case Winner: Emergency Arrival Severity Monitor: An AI vision agent that analyses CCTV footage in emergency rooms to assess patient severity and help healthcare teams prioritize urgent cases.

thanks to the @viso_ai team for hosting, excited to play around with the product have a couple of dormant projects which are itching for incredible computer vision capabilities sportwarren.com imperfectbreath.netlify.app imperfectcoach.netlify.app

@UNgethe Thank you for joining us @UNgethe ! Really excited to see those projects!

Tomorrow night in London 👁️ Come build something you've never built before. Visual AI Workshop LDN — last few spots left. 👉 luma.com/pxldog4w @LumaHQ @AITinkerers

3 days out and we're setting up for something we've never done before 👇 Our Co-CEO and CTO on stage. Live camera feeds pointed at the streets of London. 60 minutes to build something real with @visoai Viso Now. If you work in AI, ops, safety, retail, logistics — there's a use case in this room for you. 🗓️ Sunday 6 July · 6PM · London 👉 luma.com/pxldog4w @TechLondonAdv @seedcamp @techUK

30 spots left for our Visual AI Workshop in London on Sunday. Free to attend. Prizes. Drinks. And a platform that lets you build computer vision apps from a prompt. If you've been sitting on it. Now's the time 👇 luma.com/pxldog4w

Computer vision has followed the same playbook for 15 years 🧵 1/ Collect data → label it → train a model → deploy something brittle that breaks when the lighting changes. Every new use case means starting over. 2/ Visual General Intelligence changes this. No training data. No custom models. You describe what you want to understand — it sees, reasons, and acts. 3/ Where CV reads the frame, VGI understands the situation. Context-aware. Zero-shot. Any domain. 4/ This is what we built at @visoai. And on 6 July in London, you can build with it yourself. 👉 luma.com/pxldog4w cc @karpathy @ylecun @fchollet

We're hosting our first ever event in London 👁️ Visual AI Workshop LDN — a free hands-on evening where you build real computer vision apps using nothing but plain language. No code. No training data. No ML background needed. 🗓️ 6 July · 6PM · Central London 🏆 Prizes for the best use cases 🎟️ Free — spots are limited 👉 luma.com/pxldog4w @LumaHQ @swyx @AITinkerers @LdnTechJobs @londonai

@londonai This is going to be huge. Viso Now, our product, will do for vision what CHATGPT has done with text.

Build computer vision apps in minutes. London 6th of July.

Most companies think they're doing Visual AI. They're actually doing Visual Data Collection. There's a massive gap between having cameras and having intelligence. The companies winning in manufacturing, retail, and infrastructure know the difference. Thoughts?

@latentspacepod @nikhilaravi @josephofiowa @EthanHe_42 The open question for us is whether task-specific CV workflows eventually become implementation details rather than the primary user experience.

@viso_ai see our SAM 1/2/3 pods with @nikhilaravi and @josephofiowa, the rise of concepts is definitely part of it, although imo "intent" is more of a videogen problem (see our @EthanHe_42 pod) than pure CV

A question for @latentspacepod: If LLMs moved us from programming to prompting, will computer vision move from training models to describing intent? Feels like a fascinating shift.

Champions don't just react to the game. They understand it. From football pitches to factory floors, the ability to see what others miss changes everything. #VisoNow #VisualIntelligence