What do Steve Jobs and Bryson DeChambeau have in common?
They are both wildly famous and genuine game changers in their fields. People copy them, critique them, and build entire debates around them. Steve Jobs is gone now, but his imprint is still everywhere, and Bryson is still out there turning golf into an engineering exercise in public.
And they share a habit that looks quirky until you understand what it really is. They both got tired of wasted energy.
People love to talk about Jobs’ uniform, the black turtleneck and jeans, as if the clothing was the point. It wasn’t. The wardrobe was just a symptom of something deeper, the same instinct that showed up in every product he touched. Jobs was obsessed with simplicity, not as an aesthetic, but as a discipline. Fewer buttons. Fewer options. Clean defaults. A controlled experience where the user was not asked to make many choices. Even the box mattered because it was part of the experience, the first signal that this was going to be intuitive, intentional, and hard to mess up.
Bryson did something similar in golf. Same length irons are not a gimmick, they are an effort to cut down variables that drain consistency. Less adjustment from club to club. Less mental overhead at address. More repeatability where it matters, with a swing that can be built around fewer moving pieces. From the outside, both choices can look a little crazy, but they weren’t crazy at all. They were considered, the product of mastery, not a shortcut around it.
That is the part most people miss. You don’t start by stripping away choices. You start by learning which choices matter, and only after years of work do you earn the right to simplify. That is when constraints stop feeling like limitations and start functioning like leverage.
That playbook maps directly onto AI in the legal system.
Right now, too many lawyers and judges are either overly cautious or overly enthusiastic. Some want to shut it down because it feels unfamiliar. Others want to hand it the keys on day one because it feels magical. I get both instincts. But both can be wrong.
The bigger mistake is going all-in out of the gate, not because AI is inherently dangerous, but because people use it before they can define what “good” looks like in their own domain. They want the tool to supply expertise instead of supporting it. They want speed before standards, automation before judgment.
We all know about hallucinations by now. The tool can produce something that looks polished, cites authority with confidence, and still be wrong, and the danger is not just that it is wrong, but that it is wrong in a way that sounds right, which is exactly how errors slip through when people start trusting style over substance.
AI can absolutely save time and energy. It can help lawyers sort discovery, generate first drafts, compare versions, and pressure test arguments. It can help judges and staff move through records faster, organize issues, and produce clearer writing. Used the right way, it can take the busywork off the table so the human time goes back where it belongs, to judgment.
But only if the humans already understand the work, because AI does not create expertise, it amplifies it. When a lawyer knows the record and the law, AI can strip away wasted energy and leave attention for strategy and persuasion. When a judge understands the issues and the standard of review, AI can reduce the clutter and keep the focus on the actual act of judging.
But when someone is not yet competent, AI removes decisions they still needed to make. It gives them something that sounds finished, and that is when errors stop being obvious and start being persuasive.
Jobs and DeChambeau did not succeed by being reckless. They succeeded by being deliberate. Jobs simplified the entire user experience, down to the box, because he understood what could be removed without sacrificing performance. Bryson simplified his equipment and his setup because he understood which variables were worth controlling. They reduced choices only after they had earned the ability to know what could be safely removed.
That is the playbook for lawyers and judges with AI.
Start small. Learn the tool. Learn your own workflow. Decide what you will never delegate. Build repeatable templates and checks. Then simplify, standardize and scale.
Not crazy. Considered.
