In partnership with
I’ve written before about AI as leverage, and I still believe that framing is broadly correct. I use AI constantly in my own work, and my teams use it every day. At this point, opting out of AI doesn’t feel thoughtful or principled. It feels like pretending a major shift isn’t happening.
So this isn’t an anti-AI post.
What I’ve been wrestling with instead is something harder to quantify than time saved or tickets closed. It’s a subtle shift in how people experience their work when AI is introduced without much intention. Not dramatic enough to trigger alarms, but consistent enough that once you notice it, you start seeing it everywhere.
In a few cases, the way we’re using AI is quietly pulling people away from the parts of their work they actually enjoy. And that has consequences we’re not really talking about yet.
A thank you from this week’s sponsor:
AI is exfiltrating your data in fragments. See how to track and stop it on Feb 3.
Every week, AI makes your team faster—and your data more exposed. Files move to new tools, models train on sensitive inputs, and traditional DLP is blind to the context that matters most.
On February 3 at 11:00 AM PT, we’re pulling back the curtain on Cyberhaven’s unified DSPM & DLP platform—and showing how a single, AI-native platform can finally keep up with how data actually moves.
Join the live launch to see:
The first public demo of our unified AI & data security platform, built for the realities of 2026 and beyond—SaaS sprawl, shadow AI tools, and data that never sits still.
How security teams gain x-ray vision into data usage, so they can spot the risky handful of actions hidden in millions of “normal” events—and stop them in real time, not after the damage is done.
Candid stories from security leaders on where legacy DLP and standalone DSPM break down, and how they’re rethinking data protection around context instead of static rules.
A look ahead at what’s next for DLP, insider risk, AI security, and DSPM from Cyberhaven’s product and leadership teams—plus where we’re investing next.
Don’t wait for the next AI-related incident to show you the gaps in your data security stack. Reserve your spot and be among the first to see how a unified DSPM + DLP platform can change the way your organization protects its most critical data.
Onboarding feels different with AI
We’ve had new engineers ramp faster than I would have expected even a year ago. They’re productive quickly. They’re shipping code. They’re contributing in ways that look impressive on the surface.
But when I sit down with them and talk about how they’re feeling, I hear something different. They don’t feel grounded in the system. They don’t feel like they understand the codebase so much as they know how to move through it.
They’re using AI to explain unfamiliar code, generate changes, and answer questions they don’t yet have enough context to evaluate. The output is often fine. Sometimes it’s very good. But the understanding underneath it is thinner than it should be at that stage.
This isn’t a personal failing. It’s a structural one.
Onboarding isn’t an execution problem. It’s a learning problem. And learning, especially early on, requires friction. It requires sitting with confusion long enough to form mental models, making mistakes, and slowly building intuition about how a system behaves.
AI is extremely good at removing that friction. When it does so too early, people move faster but learn less. The engineers themselves notice this. When someone says, “I don’t feel like I’m really learning, I’m just doing,” they’re not being self-critical or insecure. They’re accurately describing what’s happening.
Research on AI in knowledge work supports this pattern. AI can improve short-term productivity, but when it substitutes for the cognitive work required to build expertise, it undermines learning and engagement. Speed goes up. Depth doesn’t.
This is also affecting experienced engineers
This time, it wasn’t a new engineer. It was someone experienced, confident, and very good at their job.
They genuinely enjoy coding. Not in a nostalgic or precious way, but in the sense that writing code is how they think through problems. It’s where they get into flow. It’s where they feel ownership and pride in what they build.
As AI tools became more embedded in their day-to-day work, their role shifted in small but meaningful ways. Less time building things from scratch. More time reviewing, validating, and correcting AI-generated output. The work still got done, often faster than before.
But their engagement dropped.
What changed wasn’t productivity. It was the relationship they had with the work. They went from being a builder to being a supervisor of building, and that distinction mattered more than anyone expected.
There’s research emerging here as well. When AI replaces core, identity-defining parts of a role, job satisfaction declines even when performance improves. People are generally comfortable with AI assisting their work. They’re far less comfortable when it replaces the parts of the job they care most about.
This isn’t resistance to change. It’s a reaction to losing craft.
Where this conversation usually gets stuck
At this point, discussions about AI tend to slide into unhelpful extremes. Either you’re fully embracing the future, or you’re clinging to the past. Either speed matters, or meaning does.
That framing misses the real tension.
The actual tradeoff leaders are navigating isn’t speed versus joy. It’s leverage versus meaning. AI is a powerful leverage multiplier. It compresses time, reduces effort, and increases output. Meaning comes from something else entirely. It comes from learning, from mastery, from judgment, and from feeling genuinely responsible for the work you’re doing.
The research is clear on this. AI increases engagement when it supports autonomy and learning. It decreases engagement when it replaces them. The same tools can lead to very different outcomes depending on how the work itself is designed.
Which means AI adoption isn’t just a tooling decision. It’s a job design decision.
This isn’t an anti-AI argument. It’s an anti-default one.
AI will always optimize toward the metrics we implicitly or explicitly reward. If speed is the only thing we value, AI will optimize for speed. That’s not a failure of the technology. It’s a reflection of our priorities.
But default optimization rarely optimizes for humans.
When we automate the parts of work people find draining, work improves. When we automate the parts they identify with, work starts to feel hollow, even if it looks efficient from the outside. You don’t see this immediately. You see it over time, in disengagement, in shallow learning, and in people quietly losing interest in work they used to care about.
So where can we continue to find joy in the work we do with the presence of AI?
We find it in the work we intentionally protect.
As leaders, that means being explicit about what AI should accelerate and what it shouldn’t replace. It means recognizing that some friction is productive, particularly when people are learning. It means preserving space for craft, judgment, and deep thinking, even when those things are slower and harder to measure.
AI can absolutely make work faster.
Leadership is making sure it doesn’t make work empty.