Ideas

Ideas · Why it matters

The level pages show how each sample works. These essays are about why it matters: what it means to put a thought into a machine, what goes wrong when a system speaks without knowing what its words commit it to, and why the most impressive reasoning systems of recent years pair a neural network with a logic that checks it.

They assume no Hilbert, and little logic. Read them before the samples, or after, or instead.

Essay	The question it answers
Putting a thought into a machine	Why is simulating belief, knowledge and intention hard, and how do possible worlds solve it?
The modal systems, explained	What are K, T, D, S4, S5 and KD45, and how do you choose one for a real problem?
When a chatbot makes a promise	Why did an airline pay for its chatbot's answer, and what kinds of problem does semantics solve?
Neuro-symbolic AI	Why does a network that proposes plus a logic that checks beat chain-of-thought on its own?

Where the ideas run as code: Minds for knowledge, belief, duty and time; Agents for intention; the capstone for several agents deciding together.