← Videos
How AI Answers - Models

AI from the Ground Up8 Nov 2026 · 4:41

How AI Answers - Models

When an AI assistant answers your question, where does the information come from?

From the episode description

When an AI assistant answers your question, where does the information come from? We follow one simple workshop question through the chat application, the language model and the information supplied with the request. The goal is to make those parts visible, so you can distinguish a plausible suggestion from an answer connected to an actual source. In this episode: • How the application and language model do different jobs. • What training changes, and what inference means. • How tokens contribute to a generated response. • Why supplying a document changes the current input without itself retraining the model. • What to look for before assuming an assistant searched for an answer. No coding background is needed. We use fictional Workshop Notes and deliberately written sample answers to explain the process, rather than claim results from a live model test. Try it with one short document: follow a specific detail from the source into the assistant’s reply, then check whether the reply actually supports what you asked. ResearchNode: AI from the ground up • Episode 01 of 10. Next: Why Does AI Forget? Tokens, Context and Memory Presented by Michelle, an AI-generated presenter.

Descriptions and chapter markers are supplied by the channel. They are not a transcript or an independent check of every claim in the video.