Kevin Rose

But it's part of the Vercel ecosystem if you're into that ecosystem.

But it's a way to monitor, retry, and have these long working tasks.

As you know, with edge functions, that's been the challenge, right?

You don't really get a whole hell of a lot of time and things get stale.

Anyway, this is nice.

I like it.

You're right though, it is an extra expense, but I think it is relatively inexpensive for this type of task.

All right, so back to the clusters.

So,

Now what we've done is we've run an algorithm on top of that.

We've got all the vector embeds and we're starting to build clusters.

So as you can see here, this is today.

I don't know that everything's caught up.

It looks like we've enriched.

Oh yeah, we're close.

2,284 of 2,288 stories have been, so 99.8% has been enriched in the last 24 hours, meaning that they've gone through that entire pipeline.

Of that...

ones that were more or less failures that we actually had to reach out to Gemini and get some extended data from them.

353 stories were done there.

And you can see here we're starting to get some actual weight.