


4 episodes
by TWiT · English
Hands-On AI explores and explains artificial intelligence — and shows you how to actually use it. Because AI is exactly that: a tool, and like any tool, what matters is what you do with it. In short, focused episodes, Mikah Sargent takes you inside the technology reshaping how we work, create, and communicate. Learn what's really happening inside a large language model, run AI entirely on your own hardware, put the AI already in your pocket to work, and discover uses you hadn't thought of. No buzzwords, no doom — just practical, hands-on help from one human to another. Submit your AI questions through email at [email protected] or via TWiT's social media. You might hear your question answered on the show! New episodes every Thursday.




Using AI to turn appliance/vehicle/thermostat manuals into a “home manuals” question-answering notebook with citations, using Gemini Notebook (Notebook LM) and a local alternative (AnythingLLM +…
19 Nov 2026 · 42m · 17 chaptersHow to choose Claude’s model (Haiku, Sonnet, Opus, Fable) and “effort” level (low→max) instead of always cranking to max; explains that higher effort increases internal “thinking” scratch work,…
12 Nov 2026 · 50m · 14 chaptersAI “sycophancy” (overly agreeing/flattering users) and whether modern ChatGPT resists when pushed.
5 Nov 2026 · 32m · 12 chaptersMicah Sargent tests why ChatGPT responses can sound generic, using repeated prompts and then improving results with more specific details and examples (prompt engineering).
29 Oct 2026 · 32m · 8 chapters