Half year after the big AI boom

It has been nearly six months since I wrote my last topic: AI Challenges. Additionally, it's been slightly longer since agentic coding, beyond just intelligent autocomplete, became the new norm and standard. In this post, I aim to reflect on the sustainability of my predictions and highlight the daily changes I observe in the role of an Engineering Manager.

I am a fucking oracle

So far things I wrote about half year ago are actually happening

• We are doing what we did not before: classic backend engineers are doing bunch of frontend and vice versa.

• The idea that you have to wait for someone to do something for you becoming insane: in most of the cases you can solve your issues with agent.

• Tech debt? Not an issue anymore.

• The manager's role has been reduced very much, at least from where I stand: doesn't mean that I have less things to do.. actually the opposite, and even impact in the result is much bigger than half year ago. It's just does not require to plan for year ahead anymore, so it feels like there is no scale in what we do, yet everyday's results worth weeks or even months from before.

• However, reaching agreement, communicating, and aligning with multiple parties is still a bitch and is still the most inefficient aspect, despite improvements: people are now less afraid to commit to significant projects, as 'significance' is not scary anymore, and they are more open to changes.

I am helping one project as engineer here and there, and I was focused for a few weeks on some other stuff and when I got back, project architecture (and it is not a small project: several backend services, 5-6 independent fronted modules, connected in several different ways one to the other) was completely rewritten and in a good way: it became way faster, codebase structured way better etc.. Absolutely impossible thing to happen just 9 months ago.

More with less

I survived the wave of layoffs, but many around me didn't; a lot of them are brilliant engineers, managers, and just good people. I am not an expert in business and economy, so I won't speculate on the reasoning. The only thing that is clear is that there is an idea that fewer people can now do more, and this idea is heavily tested by everyone around. From where I stand, it is true in general. And less experienced people could achieve things that seemed impossible recently. So, can we release all the seniors now? No. Because today's paradigm still requires you (yes, human, not agent) to keep the context in your head, to see the big picture. As I said before in several topics: the more decision-making you can do on your own, the more senior you are. And this hasn't changed a bit; it is just that you don't have to be a brilliant coder anymore.

We still have a significant need for individuals who can rapidly comprehend problems and anticipate numerous implications, who can mentally process a complex task, break it down into smaller steps, and ensure that decisions won't disrupt other elements.

Decision making and operations

In parallel to increased ability for individual to take many things end-to-end, the decision making is also comes on theirs shoulders more. In previous topic I had an idea:

It might be that the AI revolution in high-tech is so powerful with its pace and overwhelming nature, that it might just demolish old-school corporate decision-making bureaucracy. Many people might lose control over production, and as a result, many roles might become extinct.

It's evident that a shift is occurring. What used to be 'I'm just an engineer, ask the product team what I should do' is now turning into 'okay, here's what I've done'.

As a result, operational questions and project management, the most challenging and unpleasant aspects of tech, are gradually being handed over to yesterday's engineer or those responsible for agentic coding development.

It looks like businesses expect to have an 'army' of 'one-man army' individuals.