Defeatism at Work

AI, anxiety, and the quiet erosion of trust in engineering culture

Earlier today I had a meaningful conversation with a close friend from a previous job, and as we shared some of the challenges we’ve both been experiencing professionally, he introduced me to a term that immediately resonated with me. It perfectly captured not only the atmosphere I’ve been observing since the beginning of 2026, but also a feeling I’ve struggled to articulate in myself: defeatism.

I don’t mean defeatism as laziness or lack of ambition, but the quiet erosion that happens when people who care deeply about their work start to believe that their judgment no longer matters. When every decision is second-guessed, every activity is measured, and every team is pushed to move faster without being trusted to decide how.

Over the past several months, I’ve noticed a growing sense of tension spreading through many organizations, especially within software engineering. Teams are being pushed to move faster than ever before, largely fueled by the rapid adoption of AI tools that promise dramatic gains in productivity and efficiency. Everywhere you look, organizations are racing to integrate AI into their workflows, hoping to accelerate development, reduce costs, and ship faster.

At first glance, that sounds reasonable enough. Right?

The problem is that once large amounts of money start being spent on AI platforms, token consumption, subscriptions, and infrastructure, someone eventually begins asking difficult questions.

Are we actually moving faster?

Are teams more productive?

Can we prove the return on investment?

That is where things begin to get complicated.

A Harvard Business Review article from 2022 titled “Quiet Quitting Is About Bad Bosses, Not Bad Employees” by Jack Zenger and Joseph Folkman points out that employees do not necessarily need to resign from their jobs in order to disengage emotionally. Many simply retreat into doing only what is necessary to survive professionally.

According to Gallup’s 2024 State of the Global Workplace report, only 23% of employees worldwide describe themselves as engaged at work. Let that sink in: TWENTY THREE PERCENT. Sixty-two percent are categorized as “not engaged,” while another 15% are considered “actively disengaged.”

Perhaps the most important statistic in the entire report is this one:

Seventy percent of the variance in team engagement can be attributed directly to the manager.

That matters because I don’t think what we are seeing right now is simply an “AI problem.” I think it is a leadership anxiety problem amplified by AI.

It’s hard not to notice that many senior managers, directors, and above are now under enormous pressure to justify the growing investment in AI tools across the workplace. The message coming from many organizations seems clear enough: adopt AI quickly, show measurable gains quickly, and prove that all of this increased efficiency is translating into visible business results before the costs become difficult to defend.

The problem is that measuring creativity, software quality, collaboration, mentorship, or long term maintainability has never been easy. Faced with that ambiguity, many organizations unfortunately seem to gravitate toward the things that are “easiest to count,” like pull requests, lines of code or, comments on reviews. “Metrics” that feel objective because they produce charts and rankings. And that’s where things begin to drift into dangerous territory.

Sadly, I’ve seen reports ranking engineers based on how many pull requests they submitted in a week. I’ve seen dashboards highlighting the number of lines of code written, as though software engineering were a typing competition instead of a craft built on judgment, design, communication, and restraint. One particularly surreal example tracked how often engineers reviewed pull requests and replied with “LGTM” or “looks good to me.”

To be fair, I don’t think most leaders are doing this out of malice. In many cases, they are trying to navigate unfamiliar territory while being asked to explain, quantify, and justify something that is still evolving in real time. Many senior leaders today grew into management long before AI tooling became part of everyday engineering work, and some have understandably drifted away from the technical details over the years.

I sometimes joke that a few probably couldn’t exit Vim if their lives depended on it (it’s a nerdy joke; look it up), but the deeper issue is not technical skill alone. It’s that the pressure to produce immediate evidence of productivity can push organizations toward simplistic measurements that miss the actual purpose of the work.

It reminds me of the old Dilbert strip where employees optimize themselves around the incentive system instead of the organization’s real goals. Once people realize what is being measured, behavior inevitably shifts toward improving the metric itself, whether or not the underlying work is actually improving.

caption…

And maybe that is the part that concerns me the most.

When organizations become overly focused on quantifying every visible activity, people slowly begin retreating emotionally from the work. Creativity narrows. Curiosity declines. Risk taking disappears. Engineers stop experimenting and start optimizing for whatever keeps them off the next dashboard.

Eventually, you can feel something subtle but important draining out of the culture. The strange thing is that this can happen even in jobs people genuinely love. That’s probably the hardest part for me personally.

I still love software engineering. I still love building things. I still love solving problems with teams of creative people. I still love mentoring engineers, brainstorming ideas, running workshops, designing systems, and watching people grow into roles they once thought were beyond them.

But lately, I’ve caught myself feeling emotionally tired in a way that feels unfamiliar. Not because the work itself has lost meaning. But because it sometimes feels like the surrounding systems are slowly squeezing the humanity out of it.

The irony is that AI, at its best, should probably be giving us more room to think deeply, collaborate creatively, and focus on meaningful problems instead of repetitive work. Instead, in many places, it seems to be accelerating a culture of surveillance, optimization theater, and performative productivity.

And I think people feel it, even if they don’t always say it out loud. That, to me, is where defeatism begins. Not with anger. Not with rebellion. But with the quiet realization that effort, craftsmanship, and creativity no longer seem connected to how value is measured.

Once people begin believing that, something important starts breaking long before productivity metrics ever show it, because the real danger is not that AI will replace software engineers.

The real danger is that organizations, in their rush to justify AI investments and optimize every visible activity, may unintentionally strip away the very things that made great engineers, great teams, and great organizations effective in the first place: trust, autonomy, curiosity, experimentation, mentorship, and pride in the craft itself.

Sure, you can measure pull requests. You can measure tokens. You can measure lines of code. But the most important parts of engineering culture were never easy to measure to begin with.

And once those disappear, they are much harder to rebuild than any dashboard will ever be able to explain.