Maybe was thinking of https://arstechnica.com/tech-policy/2025/01/ai-haters-build-tarpits-to-trap-and-trick-ai-scrapers-that-ignore-robots-txt/ ?

Yeah we finally set up a workflow where we get production data available in a staging environment. This has saved a lot of trouble via “well it worked on my local where there were 100 records, but prod has 1037492 and it does not”

The other day I was updating something and a test failed. I looked at it and saw I had written it, and left a comment that said like “{Coworker} says this test case is important”. Welp. He was right. Was a subtle wrong that could’ve gone out to customers, but the wrong stayed just on my local thanks to that test.

I would have questions about how they work with a team and structure.

Are they going to be okay with planning work out two weeks ahead? Sometimes hobbyists do like 80% of a task and then wander off (it’s me with some of my hobbies).

Are they going to be okay following existing code standards? I don’t want to deal with someone coming in and trying to relitigate line lengths or other formatting stuff, or someone who’s going to reject the idea of standards altogether.

Are they going to be okay giving and getting feedback from peers? Sometimes code review can be hard for people. I recently had a whole snafu at work where someone was trying to extend some existing code into something it wasn’t meant to do*, and he got really upset when the PR was rejected.

Do they write tests? Good ones? I feel like a lot of self taught hobbyists don’t. A lot of professionals don’t. I don’t want to deal with someone’s 4000 line endpoint that has no tests but “just works see I manually tested it”

I’ve definitely had some coworkers that in retrospect we should not have hired. But I’ve also had people I was iffy on that turned out great. Hiring is hard.

I usually squash my local into a single commit, then rebase it onto the head of main. Tends to be simpler

I’ve never had a complaint about logging stuff in python. It generally does what I expect.

“Create a copy of your object and print that” is what I ended up doing, but I don’t think most people would say that’s intuitive. I expect if i print something at a particular time, I get what it is at that point in time.

Some languages are just worse to work with. like JavaScript. Console.log is like sure I’ll log your object but I’ll tell you what it is now, not what it was when you logged it.

I am going to get a lot of use out of that URL.

I’ve been telling people they need to put a dollar in the jar when they do that, but I haven’t actually been enforcing it.

I understood that reference

I mostly work in Python, but we use types at work. For a hack day project I skipped typing stuff for like an hour, and then went “wait this sucks” and added types. It was easier overall.

Maybe it’s part of an elaborate vampire hunting scheme. It seems like something idiot PCs would do in a role playing game, and the protagonists of “What We Do In the Shadows” would fall for.

Kind of off topic but some people are really bad at writing jira tickets.

“Show the user a list of projects [eof]”

Ok but like, only their projects, right? Do they need to be ordered? Searchable? Paginated? Only active ones or soft deleted ones, too? Do you just need the name or do you need metadata too?

Somehow product doesn’t love my stance of “if it’s not on the ticket or in a sop, the behavior is undefined and you get what you get” stance.

401 is “I don’t know who you are. Get fucked”

403 is “I know who you are and you’re not allowed here. Get fucked”

I was a QA engineer. I think one of the guys on the team I was on developed a stress response from hearing me walk over to his desk.

Lots of “page crashes if the user doesn’t have a last name”

“Why wouldn’t they have a last name??”

“No idea, but 372 users in the DB don’t, and 20 of them were created this month so it’s not an old problem”

“incoherent muttering and cursing”

Mouselook, huh? More like Quake. :old man yells at cloud:

I don’t trust anyone not using semver.

A whole train of thought happened in my head here about transitioning and semver.

Because I didn’t want someone to yolo connect to production, and we don’t have infrastructure in place for running arbitrary scripts against production. An http endpoint takes very little time to write, and let’s you take advantage of ci/cd/test infrastructure that’s already in place.

This was for a larger more complicated change. Smaller ones can go in as regular data migrations in source control, but those still go through code review and get deployed to dev before going out.

I have several times insisted that a migration be done via an ad hoc endpoint, because I’m a jerk, but also it’s much easier then to test, and no one has to yolo connect directly to prod.

There’s that old saying ‘everyone has a development environment. Some people are lucky enough to have a separate production environment, too’