After a Break: The Life of an Engineer in the New Reality of 2026

Hello everyone. The last time I wrote here was in August 2025. Seven months of silence.

The break lasted not because I had nothing to say, but rather the opposite – there were so many events that it was difficult to fit them into short posts. London (where I moved from Cambridge a year and a half ago) sets a frantic pace, and the industry is throwing so many challenges at us that sometimes you just want to pause and take a breath.

A quick update: I'm still at FAANG. I've just completed another performance review cycle. Despite the general anxiety in the industry and the endless news about layoffs, I not only managed to stay afloat, but also received a high rating. But, to be honest, the joy I feel about this is mixed.

Here are my main thoughts on where we are now:

1. Code as a Byproduct

This sounds strange for an engineer, but since the beginning of the year, the nature of my work has changed dramatically. Almost all production code is now generated by AI. My role has transformed into a kind of wild hybrid of architect, reviewer, and "meaning integrator."

I still spend a lot of time understanding the generated code, finding suboptimal solutions or bugs, but there is almost no classic "writing code from scratch" anymore. The routine in the traditional sense has evaporated, but the cognitive load has only increased.

2. The Cost of Resources

One of the most interesting insights of the past few months is related to economics. If you calculate the cost of the resources (compute) that are now used to support the operation of AI agents for my daily tasks, it's a bit unsettling. It feels like we're approaching a point where the computing power that supports one engineer is starting to cost as much as, or even more than, their compensation.

We have become operators of incredibly expensive and powerful systems. This changes the perception of responsibility: you are no longer just "writing functions," you are managing a huge resource that must be profitable.

3. Exponential Expectations

Because the speed of generating solutions has increased many times over, the bar of expectations has soared into the stratosphere. "Working well and efficiently" is now the basic hygiene minimum. Developers are now expected to:

* Achieve speeds that are several times higher than they were a year ago.
* Focus on generating ideas with a huge business impact.
* Be able to orchestrate this AI flow so that the project doesn't turn into a mountain of unrelated technical debt.

4. Adaptation Through Anxiety

I won't lie: I don't always feel comfortable in this situation. There's a persistent feeling that the ground beneath my feet has turned into a treadmill that is constantly accelerating. In London, being at the epicenter of all these changes, you feel this pressure in your very bones.

What will it be like in six months? It's hard to imagine. But right now, the strategy is to adapt with all my might. My past success in the performance review was the result of good old discipline, but the next one will obviously depend on how effectively I can direct the power of AI in the right direction.

The industry is changing irrevocably. We are no longer just programmers; we are conductors of systems that are becoming smarter and more expensive every day.

We're holding on. And how are you feeling about the beginning of 2026? Do you feel the acceleration?

#ru #en