I am a very responsible engineer.
For years, I reviewed other people's code carefully — I wanted to understand every detail. So when agents started writing code for me, I reviewed theirs the same way: Is the design right? The interface? The variable names? Is it sound, without being over-designed?
Then I burned out.
AI produces more code than a human can consume. Reviewing all of it line by line made me the bottleneck — the slowest step in my own workflow.
So I had to admit something uncomfortable. Before, code was written for humans to read; readability mattered. Now, most code is written for agents — comments may not even be useful anymore.
It looks like humans and agents use the same language. They don't. I was trying to read the agent's language as if it were mine.
So I gave up traditional code review. Now I focus on two things: tests that prove the behavior, and design at the high level.
Some background on where I'm starting from.
I climbed from E4 to E8 at Meta on the back of a few strengths:
— I'm extremely detail-oriented. I know most of the codebase by heart.
— I think everything through before I write a line of code.
— I'm good at organizing team work.
— I can hold an enormous amount of complex business logic in my head — and keep it all straight.
— I create the problem structure and break the work down for others. At Meta, we call this the "Uber TL" role.
If you're an experienced engineer in this industry, you'll recognize this archetype. You probably have someone exactly like this working right next to you — it might even be you.
And these strengths are exactly why moving to AI-native has been such a struggle for me.
Here's the uncomfortable part: almost every strength on this list is exactly what coding agents are now coming for.
I will document my journey here....
I'm an L8 engineer at Meta.
Like most experienced engineers in this industry, I've hit a decisive moment in 2026: coding agents shocked us — and made it impossible to ignore that an AI tsunami is about to reshape everything we do.
So I decided to embrace AI with everything I have. Honestly? It has been an extremely painful experience.
I'm going to document my learning journey here — and share my thoughts on agents, and how they'll transform our day-to-day lives.
If you're a senior engineer feeling the same shock, you're in the right place.