Tech & Future
Understand the machines rewriting everything
Novice
What is AI & Machine Learning
Old-school software was a chef following a recipe, every step written out by a human. Machine learning is more like a child watching a million people cook, then trying their own version.
You don't tell it the rules; you show it examples and it figures them out. Modern AI, the chatbots and image generators everyone's talking about, is this idea cranked all the way up. Trained on most of the internet, it gets shockingly good at imitating us. But it isn't magic, it isn't thinking, and it's wrong more often than people realise. Once you see how the trick works, you can use it like a power tool instead of treating it like a wizard.
Why bother
Make it concrete
When you type a question into an AI chatbot, it doesn't "look the answer up" anywhere. It looks at the words you wrote, then guesses the most likely next word. Then the next. Then the next. Each guess is based on patterns it noticed in everything it was trained on. That's it. That's the whole magic trick. This is why an AI can write a beautiful confident paragraph and casually state a fact that's flat wrong: it isn't trying to be true, it's trying to be plausible. Once you really feel that distinction in your gut, everything about working with AI changes, including when to trust it and when to verify.
Polymath thread
Map