Six new people this week, which doubles the size of this place. Welcome Zan, Paulette, Bhanu, Nagamani, Dhanush and Mohan. Rather than a generic welcome, here's the one thing worth knowing about what this community is for. Almost all AI training teaches you to get output. Very little teaches you to tell whether the output is any good. That gap is where people actually get burned - not because AI is wrong sometimes, but because it's wrong in exactly the same confident tone it uses when it's right. There's no tell. You can't spot it by reading; only by checking. So that's what the free AI Literacy course in the Classroom is built around. About an hour, no coding, no maths. Half of it is prompting; the more useful half is evaluation - spotting fabricated facts, knowing what never to paste into a chatbot, and breaking work into steps you can actually verify. If you're new, one question, and a one-line answer is fine: what's the last thing AI got confidently wrong for you? Could be a made-up statistic, a citation that didn't exist, code that looked right and wasn't, or advice that fell apart when you checked it. I'm asking because the answers decide what I build next. The pinned intro post is still there if you'd rather do a proper introduction.