Disclaimer: I hate writing. I’m using AI to get my ideas onto paper. The opinions, experience, and numbers are mine. The grammar is not.
Musings of an overtired CTO dictating into his phone with a sleeping baby in his arms.
There’s a big conversation happening in tech right now about what happens to junior developers. The framing goes something like this: AI is eating the junior role, the entry-level pipeline is dying, and we need to figure out how to get new people into the industry.
I think the conversation is stated wrong.
AI Is a Very Fast, Very Autistic Junior Developer
Let’s be honest about what we’re dealing with in mid-2026. The frontier models are fast. They’re productive. They can crank out tactical code at a pace no human matches. But they are, at their core, a very fast junior developer with a very specific flavor of autism. Sometimes it’s the good kind: laser focus, relentless throughput, zero ego. And sometimes it’s the other kind: context blindness, pathological literalism, an inability to hold the shape of a codebase in its head while it works on one corner of it.
I saw this firsthand this week. I was trying to communicate how I wanted a refactor to go inside a Rust codebase, and the real skill wasn’t Rust. It was articulation. Being able to express both the intent and the substance of the refactor, the why and the what, in a way the model could act on without drift. That’s not a junior skill. That’s not even a senior skill, traditionally. It’s a new skill, and most people are bad at it.
The person behind the prompt matters more than the model behind the response.
The Thing Money Can’t Buy
“Senior developer” is nearly meaningless as a title. I’ve met senior developers with five-plus years behind a keyboard who can’t debug a production incident at 3 AM without hand-holding. And I’ve met people two years in who’d run circles around them. The real axis isn’t years. Can you think architecturally? Have you been burned badly enough to know what not to do? Can you hold the full shape of a system in your head while making a decision about one piece of it?
What a truly experienced senior brings to the table is something specific: a mixture of been there, done that, got the t-shirt and fucked around and found out. That’s scar tissue. It’s the memory of getting paged at 3 AM because a decision you made six months ago just detonated in production. Money can’t always buy that. Bootcamps can’t teach it. It’s earned the hard way.
But here’s where it gets interesting. AI can help short-circuit the timeline.
What AI gives a motivated new developer is the ability to experiment and try things at 10x the speed. Ship something, break it, learn why it broke, try again. If they’re an active participant, if they’re consciously looking for the lessons instead of just copy-pasting outputs, they can accumulate the kind of experiential knowledge that used to take years in a fraction of the time.
Be Scared of the Dangerous Junior
Here’s the part most people get wrong. As a general rule, yes, junior developers are going to have a harder time. The old path of joining a big team, getting assigned narrow tactical tickets, learning under supervision, is drying up. AI handles those tickets now.
But there is a specific type of junior developer who is going to absolutely wipe the floor with your “senior” engineers. They have an above-average IQ. They have genuine curiosity and an appetite to learn. They have a $200/month Claude subscription. And they are relentlessly, obsessively working and learning, every single day.
Everyone should be scared of those people. They are going to upset entire industries. Not because AI replaces experience (it doesn’t) but because it compresses the timeline to acquire it for anyone willing to do the work.
If you just graduated: be one of them. Nobody’s going to invite you.
Stagnate and Die
Here’s the part for the seniors. And for the juniors who want to become them.
The playing field has been leveled. Not completely, but enough that it matters. Those mythical million-dollar engineers at Facebook and Google used to sit at the top of Everest. Most of them aren’t even on K2 anymore. The path to those heights used to be a decade of paged nights and burned fingers, and it was steep. Now it’s a lot less steep. A lot smoother. If you’re the right person doing the work, the climb is a fraction of what it was.
Which means the answer for seniors is the same as the answer for juniors: hone, think, learn. Every day. The scar tissue you earned five years ago is still worth something, but it’s worth less every quarter you coast on it. The tools change. The patterns change. The shape of what’s hard changes.
If you stagnate, you die. That’s been true forever in this industry. The clock just runs faster now.
AI didn’t kill the junior path. It killed the lazy one. The mountain came down, and the juniors who noticed are climbing past every senior who got comfortable. Stagnate and die.