I recently came across an interesting story about one of the key figures in the AI world. Andrey Karpaty — a Slovak with a Canadian passport — has gone from being an ordinary academic researcher to someone who truly influences the development of artificial intelligence. And his story is worth understanding.

What makes him special? Karpaty was one of the founders of OpenAI, then led the development of autopilot at Tesla for several years, and now works on educational AI projects. But the main thing is his contribution to the patent database. He has registered six patents, five of which belong to Tesla, one to Google. These cover predicting characteristics for autonomous driving, creating data for machine learning, video annotation using deep neural networks, and more. His former employers will continue to benefit from his innovations for a long time.

He was born in Bratislava in 1986, moved to Canada at age 15. Studied computer science, physics, and mathematics. By 2015, he defended his dissertation on neural networks and computer vision, and taught a course on convolutional neural networks at Stanford. Then he worked at Google, followed by OpenAI, but his stint there was very brief. The main part of his career was at Tesla, where he worked on autopilot for a full six years.

Now Andrey Karpaty records video lessons on neural networks on YouTube and founded the startup Eureka Labs. The idea is simple — to integrate AI into education so that teachers spend less time on routine tasks and more on interacting with students. By the way, he expressed an interesting idea: schools need to stop catching students using AI for homework and just reshape the educational process to fit the new reality. His words: "You will never be able to detect the use of artificial intelligence. You have to assume that any work outside the classroom can be written with its help."

Karpaty also wrote an article called Software 2.0, where he explained how neural networks are changing the very nature of programming. In his view, they will enable the creation of software so complex that humans simply cannot understand it. He even introduced the term "vibe coding" to describe development with AI assistants.

His observation about how quickly AI is evolving is particularly interesting. Just a few years ago, experts predicted that true artificial intelligence would appear in 2043. After the release of GPT-4, that date shifted to 2028. Karpaty himself admitted that for the first time in his life, he feels behind as a programmer. He compared modern AI tools to "powerful alien technology distributed to everyone, but without instructions for use."

It’s quite possible that we are really on the verge of something big. And people like Andrey Karpaty are laying the foundations of this revolution. It will be interesting to see what happens next.