Yes, I think and write with AI

Of course I use AI to help me think and write.

I also use GPS, search engines, calculators, spellcheck, IDEs, spreadsheets, databases, and occasionally other humans.

None of those tools decides what I believe.

I've spent decades developing native skills across engineering, technology, business, entrepreneurship, healthcare, investing, and the messy work of building things with real people under real constraints. Those experiences determine the destination. They shape the arguments, the examples, the judgments, the stories, and the things I choose to leave out.

AI helps me get there faster.

That should not be a controversial idea.

The map is not the destination

If I ask Google Maps to route me from Washington to Boston, nobody seriously argues that Google Maps deserves credit for choosing Boston. The map did not decide why I was going, whether the trip mattered, whom I should meet, or what I should do when I arrived.

It helped with navigation.

If I use adaptive cruise control, lane keeping, or a self-driving system, the car may perform a great deal of the mechanical work of getting me there. It still did not choose the destination.

The same distinction applies to writing.

AI can help research a topic, expose weak reasoning, generate alternatives, reorganize an argument, test an analogy, tighten a paragraph, or point out something I overlooked. Those are useful capabilities. Pretending otherwise does not make someone more intellectually rigorous. It mostly demonstrates confusion about the difference between a tool and the person directing it.

We don't ask this about other tools

• We do not ask architects whether AutoCAD makes their buildings less authentic.
• We do not ask engineers whether simulation software means they did not design the system.
• We do not accuse accountants of outsourcing arithmetic to Excel.
• We do not tell photographers that autofocus invalidated the photograph.
• And nobody looks at a pilot using modern avionics and concludes that the airplane must have chosen London.

Yet somehow, when a writer uses AI, otherwise intelligent people suddenly develop a philosophical crisis about tools.

So, yes

So, yes, I use AI extensively.

I use it deliberately and unapologetically.

I use it as a research assistant, editor, challenger, navigator, simulator, sounding board, and occasionally as a very fast intern with extraordinary recall and questionable judgment.

• The ideas are mine.
• The experiences are mine.
• The judgment is mine.
• The mistakes are mine.
• And most importantly, the destination is mine.

AI does not know what I am trying to say until I tell it. It does not know which ideas matter until I decide. It does not know which advice I would actually follow, which argument I find ridiculous, or which lessons from thirty years ago suddenly matter again today.

That is authorship.

The rest is tooling.

Three kinds of thinkers and writers

Right now there are roughly three kinds of people producing written work, and it is worth being honest about which one you are reading.

• Those using AI and telling the truth about it. They say so openly, explain how they use it, and stand behind the result because the thinking and the judgment are theirs.
• Those using AI and lying about it. They pretend every word came unaided from their own head, which is its own kind of dishonesty and usually the easiest thing to spot.
• Those not using AI at all because of ethical concerns. I respect the conviction, even if I think the reasoning is mostly confusion about tools. At least the position is honest.

The first group and the third group can disagree in good faith. The second group is the one that actually deserves your skepticism.

I am comfortably in the first group.

If you still believe otherwise

Let me be plain. If you still believe that using a powerful tool invalidates the work produced with it, you are not being rigorous. You are confusing the tool with the person holding it.

Nobody congratulates a carpenter for refusing a power saw and doing it all by hand. Nobody calls a surgeon less skilled because she used an imaging machine instead of guessing. Nobody doubts a lawyer who looked up a precedent instead of recalling it from memory. The tool did not make the cut, the diagnosis, or the argument. It made the person faster and more accurate, and the judgment was still theirs.

Somehow we forget all of that the moment a writer opens a laptop.

If the objection is really about authenticity, then be consistent about it. If you want to read my books, the articles on this site, the posts I publish elsewhere, and everything else I write without any technological assistance at all, then by all means commit to it fully.

• Turn off your phone and close every browser tab.
• Put away the search engine and stop looking things up.
• Do the arithmetic by hand, the long way.
• Print out a paper map and navigate by landmarks.
• Write your own reply with a fountain pen, not a keyboard.
• And perhaps read it all by candlelight, since electricity is a tool too.

If that sounds absurd, it is because it is. But it is exactly the same reasoning, applied honestly instead of only to the one tool that happens to make you uncomfortable.

The position is not principled. It is selective. And the only thing it proves is that the person holding it has not thought very hard about which tools they already depend on without a second thought.

Electricity has been giving writers an unfair advantage for quite some time. So has the printing press, the typewriter, the ballpoint pen, and the chair you are sitting in. Nobody asks whether any of those invalidated the work.

We have had this argument before

Here is the part that should give the skeptics pause. Every one of these objections has been made before, in print, by serious people, about tools we now consider completely ordinary. I am not paraphrasing them from memory. The sources are in the notes at the bottom.

For each one I also note how AI for thinking, writing, and creating is the same as that case, and where it honestly differs.

Read those in order and a pattern shows up. The objection is always the same shape. The tool does part of the mechanical work, therefore the person did nothing, therefore the work is not real. It has been wrong about the alphabet, the printing press, painting, photography, the typewriter, the phonograph, the ballpoint pen, the calculator, the sampler, and the search engine. Nobody today asks a novelist whether the words are really hers because a machine put the ink on the page.

Where AI actually is a little different

I am not going to pretend AI is just another typewriter. There are real differences, and the honest version of the objection deserves an honest answer. Here is what is genuinely new.

• Earlier tools moved the hand. This one can propose the thought. That raises a fair question about whose judgment is running, and the only good answer is to keep the judgment yours and say so.
• It is generative, not deterministic. A calculator is always right about arithmetic. A model can be fluent and wrong at the same time, so anything it hands me gets checked before it goes out with my name on it.
• It was trained on other people's work. That raises credit and consent questions a typewriter never did. Those questions are worth taking seriously, and they are separate from whether I wrote the thing.
• It makes plausible volume cheap. The cost of publishing something you never actually examined has dropped to nearly zero, which is why so much of what you read now feels weightless.
• It flattens voice unless you fight for yours. Default AI prose is smooth, hedged, and forgettable. If my writing sounds like me, it is because I keep pushing it back toward how I actually talk.

Every one of those is a reason to be disciplined about how you use the tool. Not one of them is a reason to pretend you do not use it. And notice that all five are questions about the person, not the machine.

They also do not explain the reflex. The reflex is not about training data or hallucinations. It is that the tool is new, and new tools always feel like cheating until they feel like furniture. My standard is simple. Own the output, check the facts, keep the judgment, and say out loud that you used the tool.