[tombert]

At Apple, on iTunes, we used an Oracle database to do job queue stuff. Initially I will admit I made fun of it because I wanted to use a "real" queue like RabbitMQ or ActiveMQ or something, but I have to admit that it worked fine and to be fair it did predate both of those.

Anyway, it made me realize that there's really no reason you can't use a SQL database as a backing store for queue stuff. I should try building my own at some point.

[dawnerd]

I remember in my early dev days thinking I was clever for creating a queue. Was right in MySQL handing forum posts back when building your own forum software was a normal thing. It worked fine. Was it optimal? Not really but neither was the entire app. Didn’t matter at the time and probably doesn’t matter today either for most apps.

[kernelvoid]

True, most apps will never hit the scale where it matters. But when you do, retrofitting a queue is painful.

[kernelvoid]

This is exactly why I built bunqueue — a job queue for Bun backed by SQLite. No Redis, no external dependencies, just bun:sqlite with WAL mode for concurrent access. Handles 100k+ jobs/sec on a single node.

The SQL-as-queue pattern is definitely underrated. Great to hear it worked well at that scale.

[lelanthran]

> This is exactly why I built bunqueue — a job queue for Bun backed by SQLite. No Redis, no external dependencies, just bun:sqlite with WAL mode for concurrent access. Handles 100k+ jobs/sec on a single node.

If you're too lazy to even write your own comments, I suspect you're too lazy to have written your own software.

At least preface your comment with "The LLM says" or preface your submission with "The LLM wrote this software".

[mpalmer]

To be honest I don't see anything egregious here.

[pyrolistical]

Why do you think that was written with a LLM?

[MitziMoto]

Because it has an emdash and people can't fathom that real people use emdashes. Like the LLMs didn't learn it from somewhere.

[kovek]

I know people who don't speak English fluently who like to use LLMs to translate to English.

[lelanthran]

"No $X, just $Y. $CONCLUSION"

[kernelvoid]

No magic, just good engineering. That’s it.

[kernelvoid]

The comment sounds "polished" because I've probably described this project dozens of times at this point.

When you repeat the same thing over and over you naturally end up with a tight version of it. That's not an LLM, that's just how it works when you talk about something a lot.

And honestly even if I did use an LLM to write a comment on HN, so what? The code is what matters.

Go run the benchmarks, read the source, open an issue if something breaks.

That's the part that actually counts.

[lelanthran]

> The comment sounds "polished" because I've probably described this project dozens of times at this point.

I didn't say it sounded "polished", I said exactly the opposite.

> And honestly even if I did use an LLM to write a comment on HN, so what?

If we wanted to chat with bots, we know where to find them.

[tombert]

> And honestly even if I did use an LLM to write a comment on HN, so what? The code is what matters.

Part of what makes these forums fun is human responses. LLMs write "good enough" text but they come off as robotic and inhuman. The only reason to go onto one of these forums is to communicate with people. If I wanted to talk to a robot, I would talk to ChatGPT, which I can do as often as I want.

[kernelvoid]

I get the concern, but there’s a spectrum.

Using an LLM to polish grammar vs. having it generate opinions wholesale are different things.

[tombert]

I'm not necessarily saying that you used an LLM, but em dashes aren't used that often when regular people are typing. I use Grammarly all the time and they've never suggested that I add an em dash, and it's often a sign of a low-effort "ask ChatGPT for a response to this comment".

Again, not necessarily saying that's what you did, just that that's the red flag.