I'm no longer searching, I'm prompting.

This is hygge

Do I contradict myself? I contradict myself. My models are large.

This balance isn't always easy to square in public communications! Someone might read them and see me as a polemicist, an apologist, or a hypocrite who wants it both ways.

I think the issue with text generation AI in particular is, though it is useful, it also encodes some ideological and ethical choices. So, I'm actively working with news outlets to help them adopt it (in a reporting assist capacity), but I am concerned about effects writ large.

It's been interesting to see the big differences between YouTubers who pursue mass audiences for bigger ad deals and YouTubers who build Patreon memberships. In defense of the hyperscalers like OpenAI, most big brands in American life trend toward duopoly.

Lot of circumlocution of late.

Just two of the turkeys.

I didn't see it initially, credit goes to @jeremybmerrill.com for this catch.

If there are 5.7 million people in Minnesota, how can 32 million turkeys be 23 to 1?

Stranger Stranger Things

Y Combinator would call this a pre-revenue metric. Would your advice be to focus on other metrics, or do something else?

Turning the dial that says "mass psychosis" and looking backward for approval.

Like if we came into SMS contact with slaves daily (cursed world), how would we use that?

The question I've always had is, "Okay so this is a good joke, but is it appropriate if the person is operating under those circumstances?" Not to be a killjoy, though, it's just an ethical line I wrestle with.

How does it respond to prompt injection and stylometry are the only ways to know.