The Paradox of AI Productivity: Why Working Faster Feels Like Burning Out

Here’s an uncomfortable truth about AI-augmented work: the productivity boost is real, but so is the exhaustion. New research explains why “getting more done” with AI might actually be unsustainable.

The Core Insight

A study from Berkeley Haas School of Business, tracking 200 employees at a U.S. tech company from April to December 2025, reveals something that resonates uncomfortably with anyone deep into AI-assisted workflows: AI doesn’t reduce work—it intensifies it.

The researchers found that AI introduces “a new rhythm in which workers managed several active threads at once: manually writing code while AI generated an alternative version, running multiple agents in parallel, or reviving long-deferred tasks because AI could ‘handle them’ in the background.”

Sound familiar? That’s the seductive trap of AI productivity. You feel like you have a partner. You feel momentum. But what’s actually happening is “continual switching of attention, frequent checking of AI outputs, and a growing number of open tasks.”

Simon Willison, one of the most thoughtful voices on practical AI development, captures the personal experience: “I’m frequently finding myself with work on two or three projects running parallel. I can get so much done, but after just an hour or two my mental energy for the day feels almost entirely depleted.”

This isn’t just anecdotal fatigue—it’s a systematic pattern the research confirms across their entire sample.

Why This Matters

We’ve built up decades of intuition about sustainable working practices. 8-hour days. Task batching. Deep work sessions. These weren’t arbitrary—they reflected genuine limitations in human cognitive capacity.

AI tools don’t change those limitations. They just make it easier to exceed them.

The HBR piece calls out a troubling dynamic: organizations are struggling to “distinguish genuine productivity gains from unsustainable intensity.” When someone ships twice as much code, is that a productivity win or a burnout trajectory? The metrics can’t tell the difference until it’s too late.

Gergely Orosz, author of The Pragmatic Engineer newsletter, adds an emotional dimension often overlooked in productivity discussions: the grief of watching AI write most of your code. The skills that took years to develop—”being ‘locked in’ and balancing several ideas while typing them out, being in the zone, then compiling the code, running it and seeing that ‘YES,’ it worked as expected”—those experiences are being displaced.

Perhaps, he suggests, “being in the zone will shift to thinking about higher-level problems, while instructing more complex code to be written.” But that’s speculation. What’s certain is that something is being lost alongside the productivity gains, and we haven’t fully processed what that means.

Key Takeaways

• Productivity and burnout can coexist: Getting more done doesn’t mean working sustainably. AI amplifies both output and cognitive load
• The “partner” illusion is dangerous: Feeling like you have help makes you take on more work than you would solo, but the cognitive overhead remains yours
• Parallelization is the culprit: Running multiple AI threads simultaneously feels efficient but creates “always juggling” mental states
• Organizations need “AI practices”: Just as we developed practices around email and meetings, we need structured approaches to AI usage that prevent burnout
• The grief is real: Beyond productivity, there’s an emotional reckoning happening—skills that defined professional identity are being automated

Looking Ahead

The Berkeley researchers call for organizations to build structured “AI practices” that manage how these tools are used. This might include:

• Explicit limits on parallel AI sessions
• Required breaks between AI-intensive work blocks
• Metrics that capture sustainability, not just output
• Recognition that “one more prompt” is the new “one more email” at 11pm

For individual practitioners, the challenge is developing new discipline. The old intuitions about “a good day’s work” were calibrated to tools that limited your pace. AI removes those governors. Without conscious effort to rebuild boundaries, the default mode is unsustainable intensity.

Willison frames it perfectly: “I think we’ve just disrupted decades of existing intuition about sustainable working practices. It’s going to take a while and some discipline to find a good new balance.”

The first step is acknowledging the problem. If your AI-powered productivity feels simultaneously amazing and depleting, you’re not imagining it. That’s the feature and the bug, wrapped into one.

Based on analysis of “AI Doesn’t Reduce Work—It Intensifies It” via Simon Willison and Harvard Business Review research