We have a no AI use policy at the company.

I had a contract developer added onto my team. I start to see AI generated comments in his code all the time. Point out that the code being contributed is def AI nonsense. I brought it up with my boss which reports to the CTO. Response: “As long as he doesn’t get caught I guess.”

He did get caught. This is me catching him and telling you.

Yup. Logic problem. I've seen similar examples.

Good luck in the stone times. Most people found ways how AI is beneficial. People that still didn't, have bad news for them. It's just a tool that requires ironically a lot of time to learn. And that is not the vision many people had for it. They thought they could lay back. That's not the case at all. AI compliments professional experience.

Working for a non AI policy company, holy fuck, never in my life. If you can get the expected results with AI, why would you ever not use it? It doesn't make sense. The only difference between me and AI is, is the fact that it adds more exception handling that I do myself. It just types for you. Don't let it technically design stuff, the programmer must do that himself.

Also, doesn't get caught? What the fuck does that actually mean? Caught with what? Efficient working? Getting caught, like if you're doing something wrong. You could say the same for autocomplete or something.

Nah, company really has to level up regarding AI instead of fighting it. Many use it for their benefit, so it is possible. It's not the tool, it's the user. And the tool, is advanced and not easy to use at all.

@retoor I’m assuming it means “getting caught slopping the codebase” which he did

@AlgoRythm regarding that, how can't you not get caught, review process right? I mean, with decent use of AI you won't see a difference. You force it to work as you do yourself. The comments are not end of the world - but you can disable those. I do not mind that much anymore. I accepted that the world of programming is changed and don't care anymore if someone knows that I use AI. I did do many years without, i know the struggles, and with that knowledge I am able to instruct LLM's now and they are just the best tool to work with for me. So, why wouldn't i use it :) It would be shooting in my own foot. Then i can type everything myself again. Why would I, I still type enough but at least not the 1000th docker configuration over and over again.

@retoor the biggest problem is AI does a “good enough” job even if you don’t use it right, but the slop creeps in.

Bugs that are hard to notice, hundreds comments that are full of emoji etc. These people don’t care enough to learn how not to slop the codebase with their chat bots. They just want the work to be done so they can jerk off all day.

And thus, it makes it a tool worth banning if people are using it incorrectly.

@AlgoRythm From experience, AI does better job then the 90% of the developers but I have to add: In someone's hand who knows what is he doing, as definitely it can result in a catastrophic garbage pile of code as well. But its not about the AI but about the person using it, If someone doesn't care and look after the basic codebase hygiene then yeah...

I don't see the problem in using AI as a helper, those policies make no fucking sense.

Yes, you use AI, it helps you, you clean it, but you don't just copy and paste it like a mother fucker!

Do you know who set the policy or is this a military style company where in absence of reporting channels your immediate superior has complete power within your contract terms?

So this company uses: Google search, forums etc... but no AI.

RIDICULOUS.

@SidTheITGuy you really should start to do nice again. What is wrong with you lately. You're sleeping on the couch or something?

@retoor @afaIk you guys are doing the exact same thing @c3r38r170 was complaining about in that other post. The AI policy is part of the premise, not the issue.

@SidTheITGuy

with all due respect

piss off

@lorentz I dunno, she comments on the consequences of the policy and how ridicules it is, that seems appropriate to the rant

Yeah... I don't really see how the phrase 'with all due respect' is being nice. It _always_ come right before the insult.

All in all, I like the American saying 'agree to disagree'. Saves us a lot of back && forth where for some reason people have this need to make others agree w/ their view.

So one person likes using LLMs - that's fine, same for !liking to use them.

Has this been clearly communicated to the consultant. I am consultant myself and most of my clients are demanding that i use AI in some form. I do but very sparsely.

It really helps with documentation.

@SidTheITGuy That's just your fantasy, now let's get back on topic shall we ;P

@SidTheITGuy What war? I hold the position of 'there's room for both'.

Snek's a nice chat room, currently frequented by a small number of folks, whereas devRant is an older, more established platform for posts - rants about tech, broadly speaking.

Why would it ever have to be one or the other?

OT: Personally I have no problem w/ a company making a determination of what they want / allow. It's their right to disallow the use of LLMs. Going against that is in this case warrants a disciplinary action as is the case w/ the guy that was caught.

@SidTheITGuy it's not my fault that you're sleeping on the couch! Relax a bit :) Happy happy joy joy! :)