AI Mental Models: A Framework for the Future of Work

Accompanying the discussion with John Davison, Lucas Draichi, Alec, and David Garber

As AI transforms how we work, the mental models that served us in the industrial age may no longer apply. In a recent discussion, John Davison shares 10 crucial mental models for thriving in an AI-driven world. These aren't just abstract concepts—they're practical frameworks for rethinking careers, businesses, and the nature of value creation itself.

The Core Mental Models

1. Specialization is for Insects

Originally from David Epstein's "Range"

The industrial age rewarded deep specialization, but AI changes this equation entirely. When machines can outperform humans in narrow domains, our advantage lies in multidisciplinary thinking. Instead of competing with LLMs on specialized knowledge, humans excel at connecting diverse mental models and applying the right framework to complex situations.

Key insight: Focus on developing competence across multiple skill areas using the 80/20 principle—learn the 20% of any skill that delivers 80% of the results.

2. Tailored Suits Required

Just as a tailored suit fits perfectly compared to off-the-rack alternatives, AI-driven coding makes it economically feasible to build exactly the business processes you need. Rather than adapting your business to existing SaaS solutions, you can now create custom operational stacks that match your specific requirements.

The shift: From "What cloud services should I buy?" to "What exact process serves my customers best?"

3. Good Devs Are Lazy Devs (And Now You Are Too)

Bill Gates: "I choose a lazy person to do a hard job, because a lazy person will find an easy way to do it."

The programmer's instinct to automate repetitive work is expanding beyond engineering. As problems scale to internet-size (millions to billions of data points), everyone needs to think algorithmically. The old "numbers game" mentality fails when the denominator reaches planetary scale.

4. The Lollapalooza Effect and Theory of Relativity

Charlie Munger's concept + Einstein's insights

When multiple systems interact, predicting outcomes becomes nearly impossible. Studying mental models deeply—not just superficially—gives us powerful tools for understanding complex situations. The difference between knowing "people have different perspectives" and understanding why those perspectives can be fundamentally different is transformative.

5. Paradox of Two Fast Cars with No Speedometers

When evaluating AI tools, beware of replacement claims without proper measurement. If someone says their solution is "faster" or "better," the first question should be: "How are we measuring this?" Look for defensible statements backed by specific, measurable attributes.

6. No Yelling in the Kitchen

From Brian Chesky's insights on organizational design

In dysfunctional restaurants, front-of-house staff get yelled at when they enter the kitchen. Similarly, dysfunctional tech companies have silos where sales can't talk to engineering. The future belongs to small teams cooperating intensely across all skill areas—engineers who understand sales, salespeople who can build with AI tools.

7. SeBS, Not SaaS (Services enabled by Software)

People don't wake up wanting to buy more software—they want problems solved. As AI makes software development cheaper, the differentiator isn't the code but the service it provides. Software is like a pill: we don't take it because we love pills, but because it solves our problems.

Focus shift: From building software to understanding what problems people actually want solved.

8. Cooperate vs. Replicate

From "The Second Machine Age" by Brynjolfsson and McAfee

Research shows that average humans working with average AI often outperform either the smartest humans or the most advanced AI working alone. The magic happens in human-AI cooperation, where different perspectives and the ability to make creative mistakes leads to better outcomes.

The Meta-Model: Studying Mental Models Intentionally

Perhaps the most important insight is that intentionally studying mental models is incredibly high-value work. Most people encounter these concepts casually, but deliberately building a toolkit of mental models—and knowing when to apply each one—becomes a superpower in complex situations.

Pro tip: Learn mental models through stories rather than abstract definitions. The emotional context helps them stick and makes them easier to recall when needed.

Practical Implications

These mental models aren't just intellectual exercises. They suggest concrete actions:

• Professionals: Develop competence across multiple disciplines rather than deepening narrow specialization

• Business leaders: Focus on creating exact-fit processes rather than adapting to generic solutions

• Teams: Break down silos and encourage cross-functional collaboration

• Everyone: Think algorithmically about repetitive tasks and embrace "productive laziness"

The Bottom Line

We're living through a transition where the rules of work are being rewritten. Those who adapt their mental models—moving from industrial-age thinking to AI-age frameworks—will be best positioned to thrive.

The question isn't whether AI will change your work, but whether you'll update your mental models fast enough to stay ahead of the curve.

What mental models are you using to navigate the AI transformation? Which of these resonates most with your current situation?