Read Zack Shapiro's The 10x Lawyer — it's the legal mirror of what's happening in the engineering world.
He writes about a re-sort happening in the legal industry. The market no longer ranks by firm brand; it ranks by individual judgment. Swap "lawyer" for "engineer," swap "BigLaw" for Big Tech.
"AI removes that ceiling. A senior lawyer working with AI is no longer constrained by the linear mechanics of production."
The same thing is happening in engineering. After coding agents arrived, the scope one person can hold expanded dramatically. Agents take on the bulk of information retrieval and execution work, freeing humans to spend time where real thinking is needed.
10x is happening. Most people just haven't noticed yet. Or they noticed, but they're in a meeting and don't have time to think about it.
Judgment
Zack says judgment is the bottleneck, not production. This holds in engineering too. People with structured thinking ability, architectural taste, clear communication skills, and enough expertise to correct direction — they seem to be the first cohort using coding agents well.
This isn't about seniority. Some people have written code for a decade with mediocre judgment. Others figure out where to focus within two years. The re-sort doesn't care about tenure; it cares whether you have judgment. Unfortunately, that's hard to put on a resume, and interviews rarely test for it.
But Is Judgment Necessarily Human?
The current division of labor is roughly: agents do retrieval and execution, humans provide judgment. Zack's entire essay is built on this assumption — AI amplifies human judgment, but judgment itself remains a human domain.
I'm not sure that's a long-term conclusion. Maybe it's just a phase. Maybe one day agents will have solid judgment too. What's humanity's advantage then? Probably that we're better at worrying.
From a humanist perspective though, I hope the transition is slow and gentle. Rather than building a digital agent to replace humans, I'd rather see another possibility: a tool that makes people better. One that learns human capabilities while giving humans feedback, helping them make better judgments, helping them learn to collaborate with agents. Not replacement — co-evolution.
I'm working on an exploration called BetterMe, and this is where the idea comes from. Not building a better agent that can replace humans, but building a system where humans and agents evolve together.
Re Zack Shapiro's The 10x Lawyer