Nobody knows what the future holds.
You're not helping with blog posts like "the apocalypse is coming," or "be like that, not like that," or "the weakest will disappear."
The truth is you don't know.
Of course, you'd better have strong skills, and be on top of your game, whatever happens.
However, you'd better start building things, testing tools and services, leveraging technology to improve features and workflows, and ultimately learning new concepts or fixing bad habits.
You don't control the market. There's no magic recipe, or hidden secrets.
If you don't know how to take the most of AI, then ask your favorite LLM, and you'll get accurate tips.
Most of the time, it will recommend breaking big problems into smaller problems or processing by iterations.
I read many posts comparing "lazy devs" to "magic AI wizards," explaining all the differences between the two "characters."
I'm not saying it's not true, but it would be the same conclusion for many other tools.
Top comments (16)
Yeah, the amount of bullshit discourse around AI is overwhelming, no matter which side one takes
Indeed.
I don't think you need to be at the top of your game. Just make smart decisions, or the smartest decision at that time. AI can assist you, but in the end you give the final OK.
Magic AI wizards got me laughing. I feel like those are the people that before AI spend a day programming to do a five minute job.
At the moment AI is still cheap because the providers want to attract people. I want to see what happens once the prices go up.
That's a huge stake for "AI providers." I don't think early adopters would be willing to lose their new toy, but it does not seem to be a sustainable business, for now.
By "top of your game," I mean "learn your craft," know what you are doing, don't bloat your code, etc, not "the best of the best of the best."
I've noticed those tools often leave dead code and unused variables, even the best ones, so if you don't review and ask for a clean up, you'll likely get unnecessary blocks of code that would be pure debt.
AI is not meant to do your job.
I read that as the best of the best. thank you for the clarification.
I'm constantly asking myself, at what time does the AI code generation stops being productive. Sure it can generate code/text faster than any typist. But the more content it generates the more people have to review.
When people need to review big changes from other people it is a task people tend to avoid. And the "ideal" AI workflow is to only do reviews. Is that an exciting look for the future of programming?
In the beginning of the year people were making fun about only needing a single button on a keyboard, because all you need to do is review and confirm. Are we still laughing now?
I think the truth is AI can't replace any jobs, I know you used job in a different context but I want to make it broader, because it not responsible for its output. Isn't taking responsibility the main part of doing a job? Whatever job that might be; collecting garbage, serving people food, educating people and writing code.
that's because it's my definition of "top of your game" :)
You have a point, and the speed looks cool but the code produced by these platforms can be hard to evaluate, especially for complete beginners.
It can even hide nasty errors under the appearance of a result that “almost” works for you.
Yeah. It seems over simplistic!
Good supervision requires much more time.
Tools are not responsible for anything, we are.
perfect
I agree, even in my company a dude that was a vibe coder was fired because he didn't know what he was doing... before I joined a dude that was like an AI prophet also was fired because he was only good with words but not with hard skills... to be honest AI is like an amplificator, if you don't have the skill, you won't neither with AI... otherwise a lot knowledge jobs such as lawyers, accountants, finance, psychology, etc would be cooked even before us engineers.
Another thing is that most engineers think they only job is to create code and that's so wrong... it's like saying the job of a basketball player is to score points, that's only part of the job.
But we can have multiple responsibilities such to communicate with the PM, PO, client, engineers, check requirements, change, create new requirements, document code, PRs, commits, and many... many more.
Great point! They usually call that "soft skills" vs. "hard skills" technical.
You cannot skip those.
I think you could mitigate very bad scenarios by forcing some context like "act like a senior engineer," and ask the tool to explain its work step by step, but it's not magical, and in the end always limited.
Agreed. Prediction is just reference which will never perfectly tell the future 🙂
Murphy's law :)
To me, those prophecies are kinda lazy. It does not mean there's no threat at all.
It's a constant threat, but it's not the only one.
Well put. I'm glad to see more and more people able to cut through the doomsday prophecies.
Agree with you! Besides, it does not help.
We all can feel how dangerous it could be in the wrong hands, but what can you do. You'd better learn how it works.
Don't trust into the hype of any new stuffs , be a critical thinker , do 1st principle thinking at first , understand into it's core then u can if u like it.
Definitely. I'm a developer. I want to understand how things work.
AI is a great tool but I dont think AI will be downfall of development.