Computing & AI

How computation works and how modern AI systems represent language, make predictions, run locally or remotely, and reuse older algorithmic ideas.

Topics

Tokens and EmbeddingsLLMs do not normally read text one character or one word at a time. A tokenizer converts text into integer token IDs representing chunks such as whole short words, word pieces, punctuation, or bytelevel patterns.
Transformer AttentionAttention lets each token build a new representation by selectively combining information from other tokens in the context.
What an LLM Actually PredictsAn autoregressive large language model is trained around an almost embarrassingly simple objective: given previous tokens, assign probabilities to the next token.
Why Local vs Cloud AI Models DifferRunning an AI model locally versus through a cloud service changes where the weights and computation live, which in turn changes privacy, latency, hardware limits, cost structure, and model choice.

Genetic AlgorithmsA genetic algorithm (GA) searches by maintaining a population of candidate solutions rather than improving one solution along a single path.

Gödel, Escher, BachThe book in one sentence
RecursionRecursion is when a definition or procedure refers to a smaller or simpler instance of itself.

WebAssemblyWebAssembly (Wasm) is a compact, portable binary instruction format designed as a compilation target for languages such as C, C++, and Rust. In a browser it runs inside a sandbox alongside JavaScript rather than replacing the browser or DOM.