"Temperature" isn't heat. "Inference" isn't deduction. "Weights" aren't in a gym, and an "epoch" takes an afternoon. AI talk is full of everyday English carrying very specific — sometimes counterintuitive — technical meanings. This series decodes 37 of the worst offenders. Take the quiz first to get your baseline, work through the decoder, then take it again and watch the jargon lose its power.
Each question drops you into a real situation — a meeting, an incident, a pricing page — and asks what the jargon actually means there. Take it cold before reading: your first score is saved as your baseline. Come back after the decoder and the same quiz shows you exactly how far you've moved.
No signup, nothing leaves your browser. Twelve questions, about four minutes. Guessing is encouraged — the wrong answers are where the learning is.
Every term follows the same shape: what it sounds like, what it actually means, how it's used in AI talk, and one concrete example. Terms link to each other where the concepts connect — follow the trails. Cards you've opened get a green dot, so you can pick the series back up where you left off.
All 37 terms in one plain-English table — for the meeting where someone says "our inference costs are asymptoting" and you have ten seconds. Click any term to jump to its full card.
| Term | Plain English |
|---|
Vocabulary is the entry ticket. These tools show you the concepts behind the words, interactively.
When your leadership team and your engineers finally speak the same language, AI strategy conversations get dramatically shorter. Helping teams build that shared vocabulary is part of what we do.
Talk to us about AI literacy for your team