I'm a programmer. AI bots like chatgpt, Gemini and Copilot are so incredibly useful for me because it's becoming harder and harder every passing week to find usable documentation on the libraries I'm using. The speed, accuracy and willingness to answer are unmatched by any other service online.
I feel like the greatest win for copilot is that Google is dogshit now and copilot is good at pulling results and citing them. I wish I didn't have to agree as an AI hater that it's often better than Googling BUT still has to be checked for lies.
I don't think I've ever got a good answer out of any of them, and trying to figure out which of its 'facts' are just made up is a pain. But why promote the use of a tool that will eventually take over your job? (albeit badly)
One of the things that development teams around the world are figuring out now is that even with AI, you still need experienced developers who know what questions to ask the AI and how to interpret the answers into a workable solution. Just randomly copy/pasting things just won't cut it.
Developers will be incentivised to use AI code because it's quicker. From a management perspective, there's no difference between a dev writing code they don't understand or an AI doing it. But with no one who understands how or why it works, they'll become more reliant on AI to debug and update it.
Devs who don't understand the code they're writing are terrible devs. It is borderline impossible, even, to give AI prompts to fix code you don't understand yourself. You won't know what the "fix" will do or whether it'll work. A dev who is nothing without AI shouldn't have it until they're not.
A dev who needs AI at all isn't a very good dev. But that's the problem, only about 10% of devs are good, 30% are ok, 30% are crap but might improve, and the rest have chosen the wrong career. Good devs are out numbered. To the rest, AI will level the playing field - downwards.
Would you say the same of a dev who uses documentation? Stack Overflow (useless as that site might have become lately)? Google? an IDE? All those things are standard tools in a developer's toolset, and the only real difference is that AI is just newer and more controversial.
Not at all comparable. Documentation is for architecture and syntax. AI will soon be taking over the logic. I like my job. I create elegant logical structures in my perfect universe of 1s and 0s. Within the next 5 years, a lot of that will be taken away.
Here's my thing: I'm a frontend developer. I like writing the HTML, Javascript and Java code that's part of my code, but I'm bad at CSS. I usually ask our web designer to do that stuff. With ChatGPT, I can ask the AI "here's my HTML and CSS, how do I get it to do X?" without interrupting coworkers.
Like, the stuff I ask usually isn't explained in documentation or Google, and gets your questions closed on Stack Overflow because of their overbearing content moderation policy. I can get my work out the door faster just by asking AI targeted questions, where before I'd need copious experimenting.
So instead of hiring a CSS developer, you've been incentivised to outsource their job to AI. No doubt there will be CSS developers who will be doing the same for html and javascript. And eventually management will realise they can get something kind of workable by just feeding in the UI/UX design.
Eventually AI-speak will become the high-level language of choice and a developers job changes from logic and structure, to persuading an idiot machine to do what you want without making stuff up. Developers may still be employed, but they won't be programmers www.theregister.com/2025/04/12/a...