Sunny Madra’s Post

One of the more impressive GPT-4o demos: full analysis of multiple CSVs from a single command. I'm pretty sure there's a large number of folks around who are paid a decent wage to do these type of analyses in a few days and this example does it in ~20 seconds - these capabilities will only get better, faster and more accurate. Think about that.

AI can now generate videos!! I was blown away by the quality of the renders and how refined everything looks in the demo video that openAI has released. The time when you would search for a video clip in stock footage and have to compromise with what you want and what’s available is gone. Although the platform isn’t available for public use at the moment, the possibilities are truly limitless! A video snippet like this - https://lnkd.in/ezgntArn

For this break, continued down the rabbit hole of running ML models and AI completely in the browser. Found out I need to enable WebGPU to get some demos running. Full video: https://lnkd.in/eKbMUDU7 #WorkinOnIt

Watch as #OpenAIo1 brings a video game to life from just a simple prompt, showcasing its impressive capabilities. With its unique ability to think carefully before responding, OpenAI's latest model outperforms all others. The more time it takes to reason, the more accurate and effective its solutions become. #strawberry #strawberryAI #GPT4O #OpenAI

Loneliness has become an epidemic. Many people are spending more time with digital companions and there are fast growing startups solving for that. One of the most dangerous use of AI gonna be lonely people being influenced by AI algorithms. In last decade - we have seen how algorithms (on social media platforms) have the power of shaping our opinions. The biggest danger we gonna face is - AI companions manipulating people's emotions to believe them in something. It would be worse than any kind of religious extremism we have seen or any dictators ability to manipulate people. Tech amplifies our ability. My general view is - screw up more than less. Life is messy but beautiful. Just go out and meet people. Have fun, mess around, get hurt, fall in love, experience life. Real-life human connections are much more valuable than anything else. Ultimately, that's what makes us human.

One of the best ways to accelerate your machine learning pipelines faster, cheaper, and easier is using zero-shot prediction. Daniel G. shows how easy it is to pre-label your dataset with FiftyOne. Key highlights: ✅ No prior training necessary – just define your labels and apply them. ✅ Utilize models like CLIP, YOLO, and Segment Anything. ✅ Compare zero-shot predictions with ground truth to see their effectiveness. Experience it for yourself through live interactive demos with Try FiftyOne —no install required! ➡️ Get started here: https://hubs.ly/Q02HvgMP0 #MachineLearning #ComputerVision #DataScience #VisualAI #ArtificialIntelligence

"Excited to unveil my latest project: a movie recommendation system! 🎬 Using collaborative filtering and machine learning, I've developed a model that suggests personalized movie recommendations based on user preferences. Check out the demo video to see how it works! Big thanks to everyone who supported me along the way. Let's connect to discuss the potential of recommendation systems in entertainment! #DataScience #MachineLearning #RecommendationSystem #bharatintern " 🚀📽️

I have something really cool to show you! OpenAI had a bug that “accidentally” gave full access to o1 (not just the preview) for a bit. They’ve fixed it now, but there are still a bunch of examples online showing off what it can really do. One example is this video where codes a simulation of gravitational lensing—the way photons move around a black hole. Should I go into detail about how huge this is going to be? Probably not, but one thing’s for sure—my company will never be the same. We’re already shifting our strategy to adapt right away since a lot of the things we’re coding manually now probably won’t make sense to do at all in the next 6-12 months.

Blogger of the Week: Ambarish Singh⭐✨ This week's blog takes us on an insightful journey into the world of Reinforcement Learning (RL), a fascinating branch of machine learning. In his blog, he unpacks the fundamentals of RL, explaining how agents learn through trial and error to optimize their actions based on rewards. From foundational concepts like Markov Decision Processes to advanced topics like Multi-Agent Reinforcement Learning, the blog serves as a complete guide for anyone looking to understand how RL models and algorithms work. Beyond the theory, the blog delves into real-world applications of RL in industries like robotics, gaming, finance, and healthcare, showcasing its transformative potential. Read the complete blog to know more. Link in the comment 🔗

AI will never be as bad again as it is right now. Everything it generates will only get better.