Technology

Problem with the theory is that people believe in LLM strongly enough that whatever pressure there is within a market to be vaguely similar evaporates. SQL certainly has dialects, but at least the basics are vaguely similar, as an example.

Working with a vendor that is oddly different from every other vendor in the space and we applied pressure to implement more typical interfaces. Their answer was "just have an LLM translate for you and use our different and frankly much weirder interface". When we did cave and use it and demonstrated the biggest LLMs failed, they said at least they give you the idea. Zero interest in consistent API with LLM as an excuse.

On the write your code for you, it has to be kept on a short leash and can be a nightmare if not overseen, though it can accelerate some chore work. But I just spent a lot of time last week trying to fix up someone's vibe coded migration, because it looked right and it passed the test cases, but it was actually a gigantic failure. Another vibe coded thing took 3 minutes to run and it was supposed to be an interactive process. The vibe coded said that's just how long it takes, if it could be faster, the AI would have done it and none of the AI suggestions are viable in the use case. So I spent a day reworking their code to do exactly the same thing, but do it in under a second.

For the jira ticket scenario, I had already written a command line utility to take care of that for me. Same ease of use instead of using jira GUI and my works torturous workflows, but with a very predictable result.

So LLM codegen a few lines at a time with competent human oversight, ok and useful, depending on context. But we have the similar downside as AI video/image/text creative content: People without something substantial to contribute flood the field with low quality slop, bugs and slow performance and the most painful stuff to try to fix since not even the person that had it generated understood it.

It will end up as an assistant rather than take over the world at least while it is affordable

I truly believe an AI winter is coming

I am in the same boat, long time infrastructure automation engineer as well. Sometimes it’s faster to explain how terraform or whatever needs to act and then fix the issues rather than having to sift through the docs for every provider.

I also do a similar thing to you with code, I also have to read a lot of other people’s code in languages I don’t know to help troubleshoot things and while I can usually follow the logic it is such a time saver to have AI to read the docs for the libraries and languages for me to at least find the part of the docs I need to read faster than searching myself.

Overall, I also agree with the sentiment on AI most of the time and all of its criticisms are definitely valid but I think too many people try to use AI to do their work for them instead of using it more like a rubber duck you can program with normal language.

Or they'll make apis shittier because they don't want AI using it.

However, Copilot has made it a lot easier to navigate through Azure's incomprehensible menu structure.

The closest facsimile I have in my work is occasionally running an Excel formula I've written through Copilot in order to find a formatting error or to help fix an Access query, but If fundamentally understand what I'm doing, can validate that the produced result is correct, and can fix it if I have to somewhere down the line.

It's good you've found some simple ways to use it, but in the vast majority of work I do, it would take longer if I used AI because everything produced using an LLM has to be human-validated regardless, so I might as well not skip the important step of learning and understanding it.

I never use it to ideate and never use it for anything that isn't eminently simple, like creating a sheet with x number of columns and rows or something like that. I hate the idea of the environmental impact and that helps me avoid it.