PART H · VOLUME 5

Games and Machine Intelligence

How a game runs sixty times a second, and how a language model is built, judged and used.

45. Games - How They Are Made and How They Run 46. Machine Learning From Scratch - Fitting, Loss, Gradients and Neural Networks 47. Parameters, Tokens and Embeddings 48. The Transformer 49. How a Model Is Actually Made - Pretraining to Release 50. Benchmarks, Evaluation, Versions and Dates 51. Serving, Running One Yourself, RAG, Agents and the Limits

CHAPTERS IN THIS PART

45. Games - How They Are Made and How They Run

You will be able to say exactly what makes a game different from every You will be able to write a correct game loop with a fixed physics step,...

46. Machine Learning From Scratch - Fitting, Loss, Gradients and Neural Networks

You will be able to say exactly what machine learning is, in one sentence, You will be able to tell the honest history: the perceptron, two winters,...

47. Parameters, Tokens and Embeddings

You will be able to say exactly what one parameter is: a single number in a You will be able to take an architecture description and compute the...

48. The Transformer

You will be able to say what came before the transformer, why it worked, and You will be able to explain attention in ordinary words to somebody who has...

49. How a Model Is Actually Made - Pretraining to Release

You will be able to name every stage that turns raw web text into a shipped You will be able to explain where training data comes from, how it is...

50. Benchmarks, Evaluation, Versions and Dates

You will be able to say exactly what a benchmark is, in five mechanical You will be able to describe the main benchmarks by name, year, size and...

51. Serving, Running One Yourself, RAG, Agents and the Limits

You will be able to say what changes when a model stops learning and starts You will be able to read a real API request body field by field, and say...