> Imagine a company that fires its software engineers, replaces them with AI-generated code, and then sits back, expecting everything to just work. This is like firing your entire fire department because you installed more smoke detectors. It’s fine until the first real fire happens.
I feel like this analogy really doesn't capture the situation, because it implies that it would take some event to make companies realize they made a mistake. The reality right now is: You'd notice it instantly. Product velocity would drop to zero. Who is prompting the AI?
The AI-is-replacing-programmers debate is honestly kinda tired, on both sides. Its just not happening. It might be happening in the same way that pirated movies "steal" income from hollywood: maybe companies are expanding more slowly, because we're ramping up per-capita productivity because engineers are learning how to leverage it to enhance their own output (and its getting better and better). But, that's how every major tool and abstraction works. If we still had to write in assembly there'd be 30x the number of engineers out there than there are.
There's no mystical point where AI will get good enough to replace engineers, not because it won't continue getting better, but because the economic pie is continually growing, and as the AI Nexus Himself, Marc Andreesen, has said several times: Humanity has an infinite demand for code. If you can make engineers 10x more efficient, what will happen in most companies is: we don't want to cut engineering costs by N% and stagnate, we want 10x more code and growth. Maybe we hire fewer engineers going forward.
> But with the AI craze, companies aren’t investing in junior developers. Why train people when you can have a model spit out boilerplate?
This is not happening. Its fun, pithy reasoning that Good and Righteous White Knight Software Engineers can prescribe onto the Evil and Bad HR and Business Leadership people, but its just not, in any meaningful or broad sense, a narrative that you hear while hiring.
The reason why juniors are struggling to find work right now is literally just because the industry is in a down cycle. During down cycles, companies are going to prioritize stability, and seniority is stability. That's it.
When the market recovers, and as AI gets better and more prolific, I think there's a reality where juniors are actually a great ROI for companies, thanks to AI. They've been using it their whole career. They're cheaper. AI might be a productivity multiplier for all engineers; but it will definitely be a productivity normalizer for juniors; using it to check for mistakes, learn about libraries and frameworks faster, its such a great upleveling tool for juniors.