Why language models hallucinate, revisiting Amodei’s code prediction and AI in the job market
Do we need more AI hallucinations? In episode 72 of Mixture of Experts, host Tim Hwang is joined by Kate Soule, Chris Hay and Skyler Speakman to talk about OpenAI’s paper, Why language models hallucinate. Next, in March 2025, Anthropic’s CEO, Dario Amodei, predicted that AI will be writing 90% of code for software developers. How is that turning out? Then, is AI making the job market hell? Finally, you can now run an LLM on a circuit board the size of a business card. Where does that take us? All that and more on today’s Mixture of Experts. 00:00 – Intro 1:18 – Oracle, data center construction, iPhone 17 and tech saint 3:03 – Why language models hallucinate 15:59 – Dario Amodei’s prediction 22:45 – Job market is hell 29:57 – LLM on a business card The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.