The fact that you don't seem to realize how this is even worse is scary.

The API is basically the prompt LOL. I chucked in the API, and had o1-pro write tests + fix bugs until it worked.

Yes that is exactly what it is! It's basically a very small upgrade over my favorite database: A JSON-file :)

Literally did something similar just last week, was looking for a good redis gui on windows and couldn't find one that clicked (closest I got was Another Redis Desktop Manager but it sorted my keys like 1, 11, 2 etc) so turned to sqlite and implemented this. Also added a "sync to disk" method so I get both the benefits of ram and persistence, worked out great since the data I'm getting over tcp has a sequence number so in case of any errors I resume from the last sequence number in the db. Thinking of fully committing and moving some stuff from the language like decoding the raw bytes to build a json object to an extension.

What's it give you above just holding a big object in memory? I guess partial serialization is something, so updates are stored on disk... But then why not just store a json file per key on disk? It's not like the serialized blobs allow you to have indexes or a particularly efficient full text search, so why bother with the SQL statements at all?

dang -- why was this flagged? Seems like a perfectly reasonable post.

Users flagged it. We can only guess why users flag things, but there are usually clues in the comments. I've turned the flags off now.

p.s. @dang doesn't work - you have to email hn@ycombinator.com if you want mostly-guaranteed message delivery.

Maybe because it’s written with a LLM

The implementation is very clearly included, if you... scrolled down.

And the point is, you could easy do `const db = new Database("./database.sqlite")` instead.

The wrapper makes it so manipulating your database is just like manipulating a plain Javascript object.

It's since been edited. At the time of my original post it was importing from a file not present on this page.

100% ;)

Is the thing that's going on here that softmax can't push a value to 0, but by subtracting 2 softmax maps we can output 0s?

Follow-up question is: Isn't it extremely unlikely to output 0?

Or negatives

Thanks for the help. After upgrading my instance, all the query times dropped dramatically. The thing is, I did not even upgrade the instance that much (micro -> medium).

I'm guessing some important bottleneck was being hit, but I have no idea what it was (maybe 1GB RAM was causing spillover to disk in a query or something? like that?).

It seems important to understand what bottleneck I hit -- but frankly, I have no idea.

It seems like the modern solution is to use something like temporal.io -- (related to windmill). Alas, surely people have been solving this problem for ages now without temporal.io

This has a context window of 65K for the storywriter version.