I’m a xennial developer. I"ve been coding for 30 years. AI now codes better (and a thousand timed faster) than most mid-level developers. The company I work for has not hired a single junior dev for months now. The new paradigm is a senior dev controlling a team of AI agents. It feels like it doesn’t even make sense to think of training juniors, because at this rate even seniors will be obsolete in a year or two.
AI in the software dev world is not hype.
How is it a positive for you to be required to deliver 10x the work in the same timeframe as before, while earning the same salary, and while having your job change without warning or negotiation? AI makes it possible to do much more work in a shorter amount of time, but that doesn’t translate into more free time - it’s just that it becomes expected that you should deliver much more work.
Let me give you an example: imagine you work for a magazine. Before, you worked on a team of 5 designers, who each had a week to come up with two or three sections of the magazine. Now, most of the team is gone, you are creating the whole magazine by yourself using AI, your job changed from writing copy and using photoshop to create art to “prompt engineering”. The company expanded its business and now they publish 10 magazines (mostly AI slop) instead of one, because they can.
This is great for those who sell tokens and maybe for your boss, no one else. Workers end up being expected to increase their output multiple times; the fun parts of the job are taken over by AI and you’re left doing basically QA all day; the market is runover by AI slop; your boss now has to compete not only with very specialized people, but also with kids using AI.
I can tell you from personal experience working in the consulting world that many people who have been heavy AI users for the last year are now ending up with burnout. I can personally see everything the article mentions going on in my real world bubble.
The codebase is large, or you have files with thousands of lines of code? Because there’s a big difference. AI agents are perfectly able to work with large codebases, but they do have problems with large files (which usually are a sign of terrible code).
I honestly don’t get what your asking. There already are loads of “vibe coded” software out there. The fact you don’t notice it, only makes my argument stronger.
For the last six months I have personally been working on pretty complex software project involving multiple user-facing frontends, a managing backoffice for admins and a backend/API in which not a single line of code has been written by a human. I can guarantee you that, in my 30 years of professionally developing software, this is the most rational, documented and test-covered codebase I have had the pleasure of working with.
You need to understand that most developers are, honestly, not very good. You can ask any dev who has ever worked consulting if they ever saw a good codebase, and most likely people will tell you that all they ever saw was spaggheti. You talk of “cheap and terrible AI code” as if the current “human code” was great, but you’d be surprised. Code, when all is said and done, is just a mean to an end. Users and stakeholders don’t give a fuck about how nice the code is, what they want is working features.
You also need to understand that I’m talking about agentic state of the art AIs that are not cheap, not copy/pasting from chatGPT. The company I work at has spent dozens of thousands on Cursor tokens for Claude Opus 4.5, GPT-5.2, etc. in just the last few months.
Regarding “how will new devs become experienced”, yeah, I have no idea. The truth is that, right now, an experienced dev + a budget for AI tokens can be more productive than a room full of juniors. Those juniors also don’t learn shit, because they just go to chatGPT to get code and understand mostly nothing of what they’re doing. I don’t know how this will evolve.
Software will still be a thing, obviously; People will still be building software. But in the same way that you code without needing to know anything about the electronics of the machine, future programmers won’t need to know nothing about the syntax and rules of programming languages, that much seems obvious to me.
Programmers will turn into project managers, with AI doing the coding itself.
One of the last jobs to be replaced
How can you see what’s going on and say this? Did you read the article?
Have you heard of Claude Code, Cursor, Windsurf, etc? Where is the equivalent to those tools for those other jobs that will be replaced first?
After plain text, code is probably the biggest share of training data. Github alone hosts billions of lines of open source code. Where is the same type of data for those other jobs?
Skill issue :shrug:
State of the art AI can indeed code better than most programmers. I can get weeks of work done in a day with $100. I’ve been programming in multiple areas for 30 years.
I’m not saying that a noob can ask an AI to build Google and it will be done; But you’d better believe that an experienced programmer using AI will deliver weeks of high-quality work in a single day.
In this scenario, people with no experience are simply dead in the water. Things that took an experienced team months to build can now be done by a single guy in a few days, and there’s no way that won’t mess with employment.
If I was a young person I would be getting as far from software development as possible. Coding as a job is basically obsolete, people like you will just take time to accept it.
The same is true of most software built by humans, and AI is able to refactor, if you ask it to.
The variable you’re missing is time. There was a big shift in quality by Christmas, and the latest models are much better programmers than models from one year ago. The quality is improving so fast, that most people still think of AI as a “slop generator”, when it can actually write good code and find real bugs and secutity issues now.