But for whatever reason, in one hiring round the vast majority had cut and pasted answers from search results verbatim (we dealt with a new recruiter, and I frankly suspected this new recruiter was telling them this was ok despite the instructions we'd given).

These were not subtle. But the very worst one was one who did like the developer you described: He'd found a forum post about a problem pretty close to the question, had cut and pasted the code from the first answer he found.

He'd not even bothered to read a few comments further down in the replies where the answer in question was totally savaged by other commenters explaining why it was entirely wrong.

This was someone who was employed as a senior developer somewhere else, and it was clear in retrospect looking at his CV that he probably kept "fleeing the scene of the crime" on a regular basis before it was discovered he was a total fraud. We regularly got those people, but none that delivered such obviously messed up answers.

For ever developer like this, you're probably right there will be a lot more that are less extreme about it, and more able to make things work well enough that they're not discovered.

The problem? They were passing the input text as hardcoded plaintext, i.e. it wasn't even a string with quotes or anything -- just `foo(here is my raw, non-string input, no quotes necessary lol)`, and they could not conceive of what the issue might be.

I'm not sure why, maybe people are missing knowledge that would allow them to understand, so they just try random things in the hope that it works? It surprises me every time it happens.

Sounds like this guy understands concurrency. :)

Sure, it isn't good practice. Sure, it probably isn't what NASA should be doing. But if you're literally building yet another uber-like app, you probably shouldn't be spending too long thinking about details.

Correct and broken are black and white if you can divide the problem correctly, and there's no excuse for shipping broken code. At some point someone has to take responsibility for not shipping garbage. I get that you, me, or any engineer don't always have that luxury, but it should be a shameful thing not something you accept as normal or ok.

I would argue it's much worse "engineering" to have no product at all.

Is that a euphemism for wandering around aimlessly? committing random code to see what works? That's also not good engineering....

Not saying it won't end in the outcome you want people gamble all the time, I'm just saying it's bad engineering.

Those, among others, are the types of cases where spending extra time getting something exactly right (even if just a few hours) is just not worth it.

It seems that only in Silicon Valley startups and the like do people start companies with only the vaguest idea of what they are actually going to build and no idea of whether or not they're solving an actual problem that anyone cares about.

No it can’t. Quick and dirty? Sure. Take on some tech debt to get to market quicker. Blindly copying and pasting? You’re never going to build functional software that way. This guy was committing code with syntax errors that he’d obviously never even run. How are you going to get to market quickly that way?

Personally I can’t see how it would be faster to ask ChatGPT for an answer then carefully scrutinize the output to make sure I understand what it’s doing. Code is often easier to write than read - especially when it’s not your code.

In hindsight the solution is obvious, just run the code without reading it then try to fix it if it doesn’t produce acceptable results.