Here is something worth sitting with for a moment. You have a job ladder. Junior, Mid, Senior, Staff Engineer. Probably written a few years ago, maybe updated once or twice since. It defines what good looks like at each level, what you expect from each person, and how you make decisions about hiring, promoting, and compensating your engineers.
That ladder is now out of date. And the gap between what it says and what reality demands is growing every single quarter.
I had this exact conversation with Jordan Cutler, Staff Engineer at Pinterest and author of the High Growth Engineer newsletter (95,000 subscribers), in Episode 166 of the Leman Tech Leadership Podcast. What Jordan said stayed with me long after we stopped recording: “The expectations are now going to be raised. There is going to be a lot more output per engineer that you hire. And not only the raw output, but the type of output.”
That is not a warning about AI. That is a warning about your mental model of what engineering looks like. And if it has not been updated since 2022 or 2023, you are evaluating people against a standard that no longer exists.
The Ladder Has Moved. Nobody Sent a Memo.
Jordan put it visually at the Infoshare conference, where we first met. He showed a slide: Junior, Mid, Senior on the left. Then the same slide on the right, with AI in the picture. The Mid is doing what the Junior used to do. The Senior is doing what the Mid used to do. The Staff engineer is operating in a territory that Senior Engineers rarely touched before.
This is not about people suddenly becoming better at their jobs. It is about the available leverage changing fundamentally.
Here is how Jordan defines each level right now, in mid-2026, with AI as a standard part of the toolkit:
Junior Engineer:
Historically, juniors needed hands-on mentorship for most problems. Today, Jordan expects them to be self-sufficient for the first 70 to 80 percent of any task, using AI to get further before asking for support from a more senior person. The moment you hire a junior and let them come to you for everything, you are both wasting time. The junior who gets hired in 2026 is the one who has already built something independently and can prove it.
Mid-Level Engineer:
Independent problem-solving is table stakes. A mid is expected to own delivery, contribute to team decisions, and bring well-formed questions, not raw confusion, to their manager. With AI in the mix, a mid today can prototype, research, and ship things that would have required a senior two or three years ago.
Senior Engineer:
No check-ins, end-to-end delivery, informal leadership within the team. That part has not changed. What has changed is the scope of impact expected. Seniors are now shipping proof of concepts in 15 minutes that used to take days. Which means the bar for what counts as senior-level output is genuinely higher. The same effort that earned a senior title in 2022 earns a mid-level review today.
Staff Engineer:
This is where the role fundamentally changes. A senior owns their team’s work. A Staff Engineer expands scope across multiple teams, organizations, and strategic initiatives. Jordan described it as being “outsourced to company-wide problems,” the person who gets pulled into the room when something critical needs cross-functional leadership, standards-setting, or architectural thinking at scale. It is senior leadership, without the management title.
The Identity Problem Nobody Is Talking About
Here is where it gets more complex than just updating a job description.
During our conversation, Jordan and I went into the psychological dimension of this shift. He said something that resonated deeply with me: the resistance to AI adoption is not primarily a skills problem. It is an identity problem.
Think about who is most threatened by this change. It is not the junior who is still building their identity. It is the engineer with 15 or 20 years of experience who thought they had figured out how to do their job well, who built a sense of self around mastering a craft, who believed the next decade would look broadly similar to the last one. Their identity is under enormous pressure right now.
I see this pattern constantly in my work with tech teams and their leaders. People who are not scared loudly. People who are quietly resistant, who frame their hesitation as concern for quality, who say “AI gives sloppy outputs” not because it always does, but because it is psychologically safer to critique the tool than to admit the tool is changing what their expertise is worth.
This is connected to what I work with in the Process Communication Model (PCM). A Base Thinker whose identity is built around competence and technical mastery will feel distressed when that mastery is suddenly portable to a tool anyone can use. A Base Persister whose identity is built around doing things the right way will push back on shortcuts, even effective ones. The distress is real. And unaddressed, it becomes a quiet blocker inside your team.
If you want to go deeper on what distress looks like in tech leadership, and what to do about it, read: Why Are We So Frustrated as Tech Leaders?.
The Race Toward the Cliff (And What Your Job Is as a Leader)
Jordan shared an image from Ethan Evans that has stayed with me. Engineers are racing toward a cliff. The cliff represents full automation of their role. The ones furthest from it are still writing every line by hand. The ones closest to it have built what Jordan calls a software factory, where they define the idea, get alignment, and let AI handle implementation in a fraction of the time.
Your job as a leader is not to push everyone off the cliff. It is to understand where each person is in that race and set expectations accordingly.
What does the software factory actually look like in practice? Jordan described it precisely: a Staff Engineer identifies a performance opportunity across multiple teams, spins up a proposal document in 15 minutes with Claude, ships a working proof of concept in roughly the same time, and then delegates the 80% that remains to engineers who can work through the validation, edge cases, and QA that AI cannot reliably finish. The Staff Engineer moves to the next high-impact problem before the first one is even shipped.
That is a fundamentally different way of structuring work than most teams are using today.
The good news for leaders: this is a systems problem, not a talent problem. Which means it is fixable. But it requires you to think about how work flows, not just how people perform.
If you want to explore how anticipatory leadership applies here, this is exactly the kind of Soft Trend Jordan is describing:
Three Things You Can Do This Week
1. Audit your expectations, not your people.
Pull out your current job ladder definitions. Junior, Mid, Senior, Staff. Read what you wrote for each level. Now ask yourself: if an engineer at that level today had access to Claude, Cursor, GitHub Copilot, and MCP integrations for their toolchain, would those definitions still be accurate? If you wrote them before 2024, they almost certainly are not. Rewrite what “great” looks like at each level with AI as a given, not an exception.
2. Stop measuring the wrong thing.
Jordan flagged something that I found sharp: leaders who measure AI adoption by token usage are optimizing for the wrong output. You can burn tokens all day and ship nothing valuable. Jordan described cases at other companies where engineers ran AI in infinite loops overnight, not to produce useful work, but to top an internal leaderboard. When you measure activity instead of value, you get optimized activity with no value. Rethink your proxies. Measure the quality and scope of delivery, not the volume of AI interaction.
3. Create a space where learning happens publicly.
Jordan created a Slack channel at Pinterest called “How I AI.” It became one of the most active channels in the company. No budget. No formal program. Just a prompt: this is how I used AI to do this faster this week. People shared what worked and what did not, and the learning compounded organically.
You can do this tomorrow. One channel, one weekly prompt: “What is one thing AI helped you do faster this week?” Watch what happens.
One More Thing: Do Not Write Off Juniors
Jordan and I both addressed what is happening in the junior market, and I want to name it clearly: junior engineering roles have dropped to somewhere between 6 and 9 percent of all engineering job offers globally. Companies are scared of the investment required to onboard someone who needs a lot of guidance.
Jordan’s position is that we should still be hiring juniors, and I agree with him. In fact, juniors are more valuable now, not less. Someone needs to QA the proof of concept that a Staff Engineer spun up in 15 minutes. Someone needs to do the last 20 percent that AI cannot reliably finish. That is meaningful, real work. And it builds the next generation of seniors.
But the junior who gets hired today is not the one who waits to be taught. It is the one who shows up having already built something, who demonstrates they can think logically, solve problems independently, and use AI as a multiplier, not a crutch. Critical thinking is now the junior’s job interview.
If you are a leader who has written off juniors entirely, I want to challenge that assumption. The cost of not developing the next generation is a problem you will feel in four or five years, when your seniors have moved on and there is nobody in the pipeline.
The Question Worth Sitting With This Week
Pick one level on your team. Just one. Junior, Mid, Senior, or Staff. Write down what “great” looks like at that level in 2026, with everything that is on the table right now.
Be honest. Is that definition different from what you had in your head twelve months ago?
If yes, you have work to do. Not on your people. On your expectations.
The team you have is probably better than you think. The standard you are measuring them against might just be out of date.
If you want to keep exploring these ideas, listen to the full conversation with Jordan Cutler on Episode 166 of the Leman Tech Leadership Podcast!
