How AI actually works
by William Craig
0 of 23 solid · next:
Learning from examples
Find my starting point
Teach me everything
Learning from examples
Everything is numbers
Training vs. using
How wrong you are
What a 'neuron' is
Text as tokens
Memorising vs. learning
Getting less wrong
Why depth helps
Meaning as position
Holding data back
What parameters are
Blame, spread backwards
Attention
The transformer
What the model sees
Next-token prediction
Retrieval (RAG)
Sampling
Becoming an assistant
Hallucination
Tool use
Agents
Foundations
How learning works
Neural networks
Language models
Why they behave as they do
Systems built on top