That thing where somebody breathlessly talks about a “great” AI feature being added to an app they use, then show that said feature immediately breaks and doesn’t work properly, but will almost certainly be fixed soon, despite the flaw in question being baked into how LLMs work.
🤨