@cwebber i saw that the other day. I found the author's aside about how they use it to summarize "hundreds" of papers a bit alarming.

@eniko @aeva @cwebber He even talks about this with children - children cannot spot problems they never encountered before, but an adult learns with every new thing they see too which has the same exact problems and summaries will inevitably omit certain things you may find problematic as a human.

@emi @eniko @cwebber there's also the problem (which iirc the author overlooks) which is what happens when someone repeatedly internalizes noise that statistically resembles information. even if you have the skill to theoretically spot a specific inaccuracy, what happens when you read misinformation restated hundreds of different ways? what happens if you repeatedly scan it without much thought because you're in a hurry?