pull down to refresh

I'm sure there's a market for consumers that value accuracy over time. Like asking an important question and letting the model compute for a few minutes to get a more accurate answer instead of spitting some BS out in a few seconds.