I'm not sure what the article actually has to say about the topic it presents. It seems to want to be something about not hiring junior talent (something by the way not restricted to the tech industry) but instead it ends up more a critique about people wanting to use a waterfall pipeline.

very hard for junior engineers out there now.

Hiring seniors is a waste of money. You need someone who knows the jargon well enough to drive LLMs, but not too much more. The LLMs are good enough to do the debugging, the coding, and a big part of the design work and system architecture.

Most of the work today is manual testing and feeding the bugs back into the LLMs.

I'm not sure I understand the logic that we wouldn't hire Juniors because now we have AI.

What jobs are Juniors doing that AI can now do? I don't get juniors to build me entire web sites or implement complicated agentic pieces of work. Most of the time, we hire Juniors to help fill the talent pool with the hope that they will be productive in a few months. If we think that we don't need to refill the pool now because of AI, then that is daft although I can understand a period where companies are deciding whether they still need 50 Developers instead of 20 + AI.

For me, the worst change for Juniors is remote working. I don't want the experience of newly qualified engineer being sitting in their bedroom all day chatting to people on Teams and not knowing when they can interrupt and when they can't. We used to have an office so we could at least do 2 days per week together but we don't any more. The banter, the office, the observation, the overhearing things is a critical part of both learning engineering and learning how to be part of a workplace.

I would have liked the article to address more specifically what skills will be needed in the future and how engineers of all levels will learn them. Senior engineers in my circle feel their own brainrot, and at the same time feel like super wizards combining our deep software knowledge with the amazing AI dev tools. How will the AI native engineers learn the things, do they need to learn the things, what does anybody need to learn anyway? Let’s talk about it.

This blogpost didn't really address the elephant in the room that increasingly the bottleneck is no longer the part that jr devs could help out with.

Every startup I know that hired some level of jr's and encouraged them to use AI found themselves in a code review bottleneck for basic code quality and architecture decisions. Many startups I know are largely forgoing jr devs.

Code quality regarding evolvability, reliability, maintainability is still part of the development process. All of the people who claimed src code is just assembly on Twitter some months ago have started chiming in that they were wrong.

I do wonder what this does to the talent pipeline like mentioned in the post.

There’s an irony too, in that junior engineers are more likely to understand and use AI tools, and quickly become more performant than their senior engineers asking questions on stack overflow. It’s what’s happening at my company and it feels silly since certain projects are way behind and constantly hiring because the very senior lead refuses to use AI

> The AI version of this argument is not a new insight. It’s an old (and mistaken) preference, wearing new clothes.

Everything sounds like AI now when you squint just a bit.

Wouldn't the logic to only hire Junior engineers if AI could do the work, what do you need the Senior salary for if they're both just asking AI to do things.

This is obviously assuming the ideology of an organization that would consider something like "We shouldn't be hiring engineers anymore" in the first place so they obviously have absolute trust in AI.

Pre-AI juniors were valuable because on many tasks, it was faster for me (senior eng) to write a quick spec have a conversation with a junior, and have them go off for a few hours or a few days and write the code, and then come back with it for review. I could do that with several junior engineers and pretty reasonably paralellize the work.

Now a junior takes my spec, drops it into Claude, and submits a PR a few minutes later. So I'm back to being the bottleneck -- there's constant pressure to provide specs and review code, and ultimately that process is just me having an indirect conversation with Claude (em dash is mine).

The junior is probably providing negative value, since it would be more efficient for me just to talk to Claude, and they're not learning anything, because it's really hard to learn anything by skimming code that's being pooped out by Claude.

I'm not sure what the solution to this is, I still think we need to train junior engineers. I think my best advice to juniors right now is stop using Claude so much. Use it to plan and answer questions, but you should still be writing code even it's slower. Because that's the only way they're going to learn enough to effectively guide AI and review output and get past being a net negative.

I don't understand when he says "the job has changed to using agents but human judgement is still required" and then goes on with that assumption will remain true for our lifetimes basically. But the elephant in the room is that's going away very soon right?

Think back to 2024 and how your job has changed over the last two years. It went from science fiction to reality. And as engineers surely we understand the concept of expotential growth. It's not just us outsiders using claude to make our code but internal anthropic engineers are using it to make claude itself better. (Probably they also use openai and gemini to double check stuff!) And that's the expotential loop. Does anyone really believe in 2028 human judgement will still be the thing?