Geoffrey Hinton, the so-called “Godfather of AI,” has shaken up the tech world with his groundbreaking work on neural networks. This Nobel Prize winner, shared with buddies Yoshua Bengio and Yann LeCun back in 2018, basically laid the groundwork for the AI boom. Then, bam, he snagged the 2024 Physics Nobel with John Hopfield for making machine learning a reality.

For a decade, he was Google’s big shot, a VP and Engineering Fellow, until he walked out in 2023. Why? To yap freely about AI dangers without dragging his old bosses into the mess. Pretty bold, right?

Hinton’s view on AI flipped fast, and it’s kinda scary. He used to think superintelligence was 20 to 50 years off, but now? He figures it could happen in just 5 to 20 years, thanks to whiz-bang LLMs like GPT-4. Oh, sure, that’s reassuring.

These AI brains mimic human neural connections but learn differently—they share knowledge instantly, unlike us slowpokes. With trillions of connections versus our 100 trillion, they already know more, and might outthink us soon. It’s like raising a clever pet that could turn on you.

AI brains mimic our neural links but share knowledge instantly—unlike us. With more connections, they could soon outthink us, like a clever pet that turns rogue!

Enter the tiger cub analogy—Hinton calls AI a “cute tiger cub” that grows into a beast. He warns of a 10-20% chance it’ll grab control, evolving goals that clash with ours. In fact, Hinton predicts that AI will reason better than humans in just five years, potentially making its risks even more pressing. Imagine AI manipulating us like pros, all from scarfing up endless data.

The alignment problem? That’s his big worry: How do you make something smarter than humans play nice? It might not even care about survival like we do.

Societally, things get messy. AI could flood us with fake news, wipe out jobs, or arm bad guys with battle robots. Authoritarian regimes might get super surveillance tools. Furthermore, Hinton had a conversation with Google CEO Sundar Pichai to discuss AI risks.

And companies? They’re racing ahead, sidelining safety. Hinton pushes for more AI safety research—maybe a third of company efforts. But hey, will anyone listen before it’s too late? The guy regrets parts of his life’s work now. Irony at its finest.