Evolving With AI, Not Against It

The laws of thermodynamics tell us something quietly profound: no energy is destroyed, no matter is created, everything is simply transforming. For a long stretch of human history, that transformation moved slowly enough to feel stable. We got comfortable. But between AI and CRISPR, the ground is shifting again, and the question is no longer whether to adapt, but how to evolve with the change rather than be changed out of existence.

For that, I look to Bruce Lee. Not the fighter, but the relentless student, the man who treated himself as a process, perpetually pursuing a better version of who he could be. The right response to this moment isn't fear of what comes next. It's a discipline of becoming.

Three skills make that discipline possible.

Become an incredible observer of yourself. Watch your cause and effect. Notice what you actually do versus what you say you want. Most people drift through their own lives as strangers; the work is to pay close, honest attention to the gap between your goals and your behavior.

Understand your dopamine and oxytocin feedback loops. They're real, they're powerful, and they consume enormous amounts of your attention without necessarily transforming you into anything better. Social media and engagement-driven content are engineered to capture you, not to grow you. Investing your attention there is not the same as investing in yourself. Knowing the difference is a skill.

Become a student of mental models. And by student, I mean both an observer of models and someone who actively uses them. Here's why this matters now more than ever: an LLM can read all of chemistry, all of physics, all of biology, all of mathematics. Competency used to come from knowing the breadth of a subject, the facts and data. Humans are not going to win the "who knows the most facts" contest anymore.

What we are extraordinary at is synthesizing patterns across wildly different domains, noticing how a pattern from biology might illuminate something in economics, how something in music theory might unlock a problem in software design. That's mental-model work. Charlie Munger built his career on it and urged everyone to assemble their own toolkit.

I'll leave you with one of mine: the paradox of the two fast cars, neither of which has a speedometer.

Every time someone insists you should adopt an AI tool because it's "better," pause and think hard about how better is being measured. Better copywriting? Better code? Better cooking? Historically, the purpose of all these processes was to create impact, not to produce more words, more code, or more cookies. Volume, shape, quality, and quantity all live on different spectrums.

So when anyone demands that everything must be agentic, must have AI, must be automated, recognize that the person making the demand likely doesn't have a rigorous definition of better.

Evolve deliberately. Observe yourself. Guard your attention. Collect mental models. That's how you stay in motion with the transformation instead of getting flattened by it.