Which AI model should you use, and how much effort should it put in? Using Claude as the example, Mikah Sargent explains its four model families, Haiku, Sonnet, Opus, and Fable, and its five effort levels, and shows that ChatGPT and Gemini have the same two dials. He explains what an AI model's "thinking" actually is and what changes when you switch models or turn up the effort. He compares an easy job on a light setup and a heavy one, tests a logic puzzle with one right answer at two effort levels, and explains why the top setting isn't automatically the best one. Then he shares a way to choose that works in any chatbot. Send your AI questions to [email protected] Host: Mikah Sargent (https://twit.tv/people/mikah-sargent) Download or subscribe to Hands-On AI at https://twit.tv/shows/hands-on-ai Want access to the ad-free audio and video and exclusive features? Become a member of Club TWiT today! https://twit.tv/clubtwit Club TWiT (https://twit.tv/clubtwit) members can discuss this epi
Chatbots have a reputation for telling you what you want to hear. Mikah Sargent tests that by telling ChatGPT it's wrong when it's right. Then he tells one made-up argument twice, with the same facts and only the teller changed, and pushes back on the verdict. He explains what researchers call sycophancy, one way training can encourage the habit, and why it can be easy to miss. Then he shares four habits for questioning the answer. Send your AI questions to [email protected] Host: Mikah Sargent (https://twit.tv/people/mikah-sargent) Download or subscribe to Hands-On AI at https://twit.tv/shows/hands-on-ai Want access to the ad-free audio and video and exclusive features? Become a member of Club TWiT today! https://twit.tv/clubtwit Club TWiT (https://twit.tv/clubtwit) members can discuss this episode and leave feedback in the Club TWiT Discord (https://twit.memberful.com/account/discord/authorize). ZipRecruiter.com/twit (http://ziprecruiter.com/twit)