Five AI terms people keep using wrong
Model, algorithm, training, prompt, agent. What each one actually means, so the headlines stop being a blur.
Our resource library covers the ideas that don't change. This is where we write about what's happening right now: new tools, news worth actually understanding, and answers to the questions you send us.
A plain-English guide to hallucinations, featuring the lawyer whose court filing cited six cases ChatGPT invented, and the airline chatbot that got its company fined for a discount policy it made up.
Two loud claims dominate this conversation, and they're both wrong. What the actual jobs data says, and what to actually do about it.
The Pope's fake puffer jacket fooled millions. A $25 million deepfake scam and a fake damage claim show what's at stake when you can't tell the difference.
Four things, in rough order. If one of these is what you need, say so and it jumps the queue.
Model, algorithm, training, prompt, agent. What each one actually means, so the headlines stop being a blur.
A short, practical routine for verifying anything a chatbot tells you, useful whether you're doing homework or reading the news.
Not a panic piece. What's genuinely worth worrying about, what isn't, and how to have a useful conversation about it.
A regular round-up of real questions people send us, answered properly. Send us yours and it might be in the first one.
We announce every new post, guide and video on Instagram the day it goes up, and that's the most reliable way to catch them right now.
A proper email newsletter is on our list, but we haven't built it yet and we'd rather not collect addresses we can't email.
Send it in. Reader questions are the whole reason for one of our first planned posts.
Ask us somethingThe free courses page is live and full of material you can start right now.
Free coursesTell us what you're confused about. We genuinely choose topics based on what people ask us.