Explained
The AI terms behind the news, in plain language. A minute or two each, a new one most mornings.
- A
- AI agentWhat makes an agent different from a chatbot, and why that changes the risks.Oct 19, 2026
- B
- BenchmarkHow AI models are tested and scored, and why a high score is not the whole story.Oct 18, 2026
- C
- CVEThe shared ID number every known security hole gets, and what the score next to it means.Oct 25, 2026
- ComputeWhy the chips and data centers behind AI are the biggest bill an AI company pays.Oct 27, 2026
- Context windowHow much a model can read at once, and why it runs out.Oct 16, 2026
- Coordinated disclosureHow security holes are kept quiet until a fix is ready, and why that matters now that AI finds more of them.Oct 22, 2026
- Cross-site scripting (XSS)How a stranger's code can end up running inside a website you trust, and why that is so dangerous.Nov 1, 2026
- F
- Formal verificationHow a math-style proof can show that code behaves correctly for every input, not just the ones you tested.Oct 23, 2026
- M
- Mixture of experts (MoE)Why huge models only use a slice of themselves for each word.Oct 16, 2026
- Model routingWhy AI tools increasingly pick the model for you, and how that saves money.Oct 29, 2026
- O
- Open sourceWhy open source is about rights, not just being able to see the code.Nov 2, 2026
- Open-weight modelWhat you get when a model's weights are public, and what you don't.Oct 16, 2026
- P
- P vs NPThe biggest open question in computer science: if an answer is easy to check, is it also easy to find?Oct 28, 2026
- Prompt injectionWhy an AI can be steered by text it was only meant to read, and what helps against it.Oct 26, 2026
- Q
- QuantizationHow AI models are shrunk to fit on smaller computers, and what it costs.Oct 17, 2026
- R
- Recursive self-improvementWhy AI that helps build better AI could speed up progress, and why labs watch for it.Nov 3, 2026
- Red teamingWhy AI labs pay people, and other AIs, to attack their own systems.Oct 24, 2026
- Reinforcement learningHow AI models learn by trying, failing and being rewarded, and why labs use it to make them better at coding.Oct 30, 2026
- S
- Supply chain attackWhy attackers hide malware in software you already trust, and what you can do about it.Oct 31, 2026
- T
- TokenThe small pieces of text an AI model reads and writes, and why you pay per token.Oct 20, 2026
- W
- Word error rateThe standard score for speech-to-text: how many words a model gets wrong, and why lower is better.Oct 21, 2026