Everyone knows AI is often wrong, and the industry just doesn't care
Quote
We don't know. And it's probably always changing because every time these models are retrained, they get new things right and old things wrong.
Episode 237: FF-P-MEG (episode page, opens in a new tab) 00:40:30 (play audio from this timestamp, opens in a new tab)
Voice-matched
Summary
Nobody knows whether LLMs give false information 15% or 40% of the time, and it keeps shifting because each retraining gets new things right and old things wrong. Dave sees a software-industry mindset shift: instead of shipping what you believe is bug-free and patching later, everyone knows the output is often bad and says 'we'll fix it later somehow.'
Not in the audio. Guest episode (Daniel J. Lewis; 0.35 threshold). A 9-cue Dave run continuing his rant about being 'depressed' over AI hype.
Also said
Episode 237: FF-P-MEG (episode page, opens in a new tab) 00:40:41 (play audio from this timestamp, opens in a new tab)
Everyone understands that this stuff is giving bad information a large percentage of the time. But it's like, well, we just don't care.
Voice-matched