https://pedroacosta.dev/blog/AI notes | Pedro AcostaExplanations and reflections on artificial intelligence, from its foundations to the systems we build with it.2026-10-18T00:00:00ZPedro Acostahttps://pedroacosta.dev/https://pedroacosta.dev/blog/what-is-a-decision-model/What is a decision model?2026-10-11T00:00:00Z2026-10-18T00:00:00ZA broken mug connects typed AI answers, calibrated probabilities and the cost of acting. Why a likely answer does not choose the action for you.https://pedroacosta.dev/blog/do-more-parameters-make-a-better-llm/Do more parameters make a better LLM?2026-10-04T00:00:00Z2026-10-04T00:00:00ZWhat 27B really means, why total and active parameters differ, and how to read context, memory and reasoning limits without mistaking a specification for a quality score.https://pedroacosta.dev/blog/what-does-a-model-learn-from/What does a model learn from? Training data2026-10-04T00:00:00Z2026-10-04T00:00:00ZFollow a collection of handwritten digits into a text corpus to see how sources, selection, duplicates and missing examples shape learning and evaluation.https://pedroacosta.dev/blog/what-is-a-transformer/What is a Transformer?2026-10-04T00:00:00Z2026-10-04T00:00:00ZFollow the fisherman from an English sentence to its translation and discover how a Transformer connects information to build an output, without needing formulas.https://pedroacosta.dev/blog/what-is-an-llm/What is an LLM, and how is it trained?2026-10-04T00:00:00Z2026-10-04T00:00:00ZFollow a fisherman's sentence through predictions, loss and parameter updates to understand what pretraining changes, why scale matters and how learning is checked.https://pedroacosta.dev/blog/from-words-to-vectors/From words to vectors: tokens, embeddings and why meaning becomes geometry2026-10-03T00:00:00Z2026-10-03T00:00:00ZFollow a sentence from tokens to identifiers and learned vectors, then see why context is still needed to interpret a word.https://pedroacosta.dev/blog/what-is-attention/What is attention, and what is it for?2026-10-03T00:00:00Z2026-10-03T00:00:00ZFollow an ambiguous word through queries, keys and values to see how attention combines information from context, and what its weights can tell us.https://pedroacosta.dev/blog/what-is-a-neural-network/What is a neural network?2026-09-28T00:00:00Z2026-10-03T00:00:00ZFrom a handwritten digit to a prediction: how neurons and layers work, how a network learns, and how its architectures fit into artificial intelligence.https://pedroacosta.dev/blog/is-ai-really-new/Is AI really new? A history of ideas becoming useful2026-09-21T00:00:00Z2026-09-21T00:00:00ZA brief journey through the ideas behind artificial intelligence, what changed along the way, and why a new capability is not the same as a reliable system.