I deny your undeniability. Prove it by showing a real application vibe-coded, something useful that is not yet another electron POS. And where are all the bugs fixed in open-source projects? I have asked this question a hundred times.

Curl is spammed by fake bug reports for example.

Every AI user has something to sell but nothing to show. Make a difference for once.

AI won't. Executives who drink the koolaid will.

Nah. They'll all be needed to clean up the inevitable mess.

In any technology career, you are going to have to adapt to changes in technology. We went through the same transition from writing machine code to writing object oriented code. And from writing monoliths to writing micro services. You may not write code letter by letter the same way you used to. But your ability to adapt is your real value, not the code you can type from memory.

AI code generators still need very clear directions. They need guidelines and restrictions. They need architectural designs and decision making. They need code review. We still need the experience of qualified people, we just don’t need you to be the one typing out every letter. Delegate that part to the AI and up level your contribution.

My company dabbles in AI. A hardly abridged conversation with a bot I’ve read recently:

Hallo, Ich bin ein Bot. Wie kann ich Ihnen helfen?
(Hi, I’m a bot. How can I help you?)

Speak English, please.

Gerne antworte ich in Englisch. Wie kann ich Ihnen helfen?
(I’m happy to speak English. How may I help you?)

Caller disconnected

Yep, AI definitely is the future.

There will always be stuff to code. Even if the task of coding itself were completely removed, being able to use AI for coding will require knowledge of the subject matter and an ability to express one's needs.

If you go onto any bug list, most people are unable to express the most basic information to dill in a bug report form. "What are you using", "What did you do", "What did you expect to happen", "What happened instead". Just that. It doesn't require any technical knowledge, just the rudimentary ability to describe events and desires and yet they still fail.

AI assisted coding requires the ability to understand ambiguity too. "Solve world hunger" can be solved by killing all of humanity so that nobody is hungry anymore, putting porridge in front of evey human being until the end of time, hooking up every human to a feeding tube and harnessing energy from it, modifying human genes to not know what hunger is, or rebuilding society to give everybody equal access to nutrition that they enjoy.

Brandon Sanderson talked about it in the context art, "Making art with AI doesn't make you an artist, it makes you an art director." People who only understand grunt level coding work are at risk. The work in the future will require higher level skill sets in engineering, architecture, and project management.

I don't share your concerns about the profession. Even supposing for a moment that LLMs did deliver on the promise of making 1 human as productive as 5 humans were previously, that isn't how for-profit industry has traditionally incorporated productivity gains. Instead, you'll just have 5 humans producing 25x output. If code generation becomes less of a bottleneck (which it has been doing for decades as frameworks and tooling have matured) there will simply be more code in the world that the code wranglers will have to wrangle. Maybe if LLMs get good enough at generating usable code (still a big if for most non-trivial jobs), some people who previously focused on low-level coding concerns will be able to specialize in higher-level concerns like directing an LLM, while some people will still be writing the low-level inputs for the LLMs, sort of like how you can write applications today without needing to know the specific ins and outs of the instruction set for your CPU. I'm doubtful that that's around the corner, but who knows. But whatever the tools we have are capable of, the output will be bounded by the abilities of the people who operate the tools, and if you have good tools that are easily replicated, as software tools are, there's no reason not to try and maximize your output by having as many people as you can afford and cranking out as much product as you can.

Well, I think I agree with you a bit.

My story: six years ago in France I followed a three months cours to become a fullstack dev.

We were two or three out of twenty to know some basic stuff about programmation. There rest were smart people but without knowledge about computer at all.

The job market was so hungry for dev, that even us found a job immediately.

But I saw that some of my colleagues were 'not good'. And I think that they can be replaced at some point by Nocode/lowcode tools, or by a dev that use LLM.

But at the same time, like said @6nk06@sh.itjust.works that something to write some code from scratch, and something else to develop a software on the long run.

Eventually, I expect that advanced AI will. The same is true for many human professions. But I'm dubious that pure LLMs will be the route this takes. Too many limitations.

It's possible that such an advanced AI might make use of or incorporate LLMs.

At best it's going to make a lot of developer jobs insufferable because they're going to be cleaning up the large pile of horseshit that an AI produced to "reduce costs".

Though besides the "AI coding independently" area which I believe is still science fiction for the time being, I do think Software Engineer jobs will be reduced and wages lowered due to the massive productivity increase each developer has from using AI to deal with the repetitive stuff or faster troubleshooting and learning, that previously had to be done by looking at many Stackoverflow posts or parsing through other messy sites to find how something is done.

Kekw we'll see