in the context, maybe they have something. If what they're doing is looking for the one correct narrative, that's worse. Can we agree on that?

If it turns into a Wikipedia-like narrative-telling situation where there's one truth that comes out of it, that's worse. If it shows both sides and doesn't try to play favorites, just hey, here's the argument, make up your own mind, that's better. Because if somebody tweets something, let's say they believe something they saw on CNN, and I put a comment on it in our current model without this new feature, who's going to see it? Right? The people I want to persuade or inform, not likely they're going to see it.

But what would happen if I made the best argument against the tweet that I thought was a wrong narrative or interpretation? Well, if I make a really good counterpoint, the odds of my counterpoint floating up to the top should improve, right? Because it's a good counterpoint. Other people see it. They say, yeah, that's what I would say. He's saying this the way I would say it. So I have some optimism for this feature, but I'll acknowledge there's a million ways it could go wrong.

All right. Somebody's building a face mask that has a COVID test indicator built into the face mask. So I think something would change color based on the droplets coming out of your mouth all day. Eventually they might build up to the point where it could measure them. And I said to myself, if we can do that, shouldn't we have rapid tests already? If you can build that technology into a mask, you can't just put it in a pack of gum and sell it to me over the counter?

And if the problem is that the mask would be, let's say, a lower sensitivity than a proper COVID test, then don't