Did I Write This, or Did AI? That May Be the Wrong Question

For the past two months, writing has been harder than it has ever been in my professional life.

Following a broken arm from a crash on my bike, a total reverse shoulder replacement surgery, and the nerve issues that came with it, my right hand became nearly unusable. I could still think. I could still advise clients. I could still form opinions. What became difficult was getting those thoughts onto a screen quickly and comfortably.

So I adapted, starting with typing using my left hand only.

Then voice-to-text became an important tool. Artificial intelligence became another. It helped organize fragmented thoughts, improve clarity, and structure ideas more efficiently.

Which raises a question that seems to be on many people's minds these days:

Did I write this article, or did AI?

The more I think about it, the more I believe that's the wrong question.

The Druckenmiller Debate

Recently, investor Stanley Druckenmiller found himself in the middle of a public debate after acknowledging that he uses AI to help write. Defending his approach, he compared AI to a calculator, arguing that he sees no issue with using technology to help communicate ideas more effectively. He reportedly remarked that he uses AI for writing in the same way he uses a calculator for math.

The controversy wasn't really about his views on Treasury policy.

It was about authorship.

A discussion on a popular podcast followed, where one participant compared AI-assisted writing to lip-syncing. The argument was that readers expect to hear directly from the author and that content generated by AI somehow dilutes that connection.

I understand the concern.

I just don't think the analogy fits.

Every Generation Thinks the Next Tool Is Cheating

Long before AI arrived, similar debates surrounded calculators.

Then spreadsheets.

Then word processors.

Then spell-check.

Then grammar-checking software.

Now it's AI.

The criticism often sounds remarkably similar:

The old way was authentic. The new way is cheating.

That's why I found myself nodding along while listening to a conversation between legendary producer Rick Rubin and the team at Andreessen Horowitz discussing his book The Way of Code. Rubin's perspective is both simple and profound:

"Tools will come and tools will go. Only the vibe coder remains."

That statement resonated with me because Rubin has spent an entire career working at the intersection of creativity and technology.

Sampling machines.

Drum machines.

Digital recording.

Production techniques that many traditionalists initially viewed as somehow less authentic than older approaches.

Yet history has a funny way of changing our perspective. Technologies once dismissed as shortcuts often become accepted parts of the creative process.

A Beastie Boys Lesson

One of my favorite examples comes from Rubin's own history.

Rubin has often said he isn't really a musician. Yet he famously appears playing guitar on the Beastie Boys' "No Sleep Till Brooklyn."

The lesson isn't that Rick Rubin is secretly one of the world's great guitar players.

The lesson is that creativity has never required mastery of every instrument involved in the final product.

Sometimes the contribution is vision.

Sometimes it's judgment.

Sometimes it's taste.

Sometimes it's simply recognizing how seemingly unrelated ideas fit together.

The person contributing doesn't need to be the best guitarist in the room to help create something meaningful.

That observation feels particularly relevant to today's discussions about AI.

Sampling, AI, and Human Creativity

The early days of hip-hop offer another useful analogy.

Sampling was controversial.

Critics argued that artists weren't creating something truly original if they were drawing from existing recordings.

But that criticism missed the point.

The magic wasn't in the machine.

The magic wasn't in the record being sampled.

The magic was in knowing what to sample.

What to combine.

What to emphasize.

What to leave out.

What story to tell.

The machine provided capability.

The artist provided judgment.

Large language models operate very differently from music samplers, of course, but I think there is a similar lesson.

The value doesn't come from the tool itself.

The value comes from the human deciding what matters.

This Debate Isn't New to Me

Years ago, I had a similar discussion with a former business colleague.

He believed that portfolio management software was almost a form of cheating.

When evaluating client portfolios, he preferred manually writing down current holdings, comparing them against target allocations, and calculating every trade by hand.

There was genuine value in that approach.

If you're managing your own account or perhaps a handful of portfolios, manually working through every position may deepen your understanding.

But when you're responsible for dozens or hundreds of client accounts, the equation changes.

My perspective was different.

The software wasn't replacing my judgment.

It was amplifying it.

The software could perform repetitive calculations with consistency and speed across large numbers of accounts. I remained responsible for the recommendations, oversight, risk management decisions, and client outcomes.

In hindsight, our disagreement wasn't really about software.

It was about trust.

He trusted the manual process more.

I trusted my ability to supervise the technology.

Neither of us was trying to avoid responsibility.

We simply drew the line in different places.

The Low Oil Light

I've often thought about technology this way.

I trust the low-oil warning light in my vehicle.

That doesn't mean I've surrendered responsibility for maintaining my car.

It means I don't stop every few miles to manually inspect the dipstick because a monitoring system is already watching for me.

The tool handles the detection.

I remain responsible for the outcome.

Calculators didn't eliminate mathematics.

Spreadsheets didn't eliminate accounting.

Word processors didn't eliminate writing.

GPS didn't eliminate navigation.

Portfolio management software didn't eliminate financial judgment.

Instead, each tool shifted where humans add value.

My Shoulder Changed the Conversation

This became much less theoretical over the last several months.

Following shoulder replacement surgery, there were days when typing left-handed were frustratingly slow.

Voice-to-text helped bridge the gap.

AI became an assistant that helped organize thoughts, improve transitions, and structure ideas.

But AI didn't experience the surgery.

AI didn't sit across from clients discussing retirement.

AI didn't help families navigate healthcare decisions, tax questions, market declines, career changes, or life's many transitions.

AI didn't form the opinions in this article.

Those experiences belong to me.

The technology simply helped me communicate them.

In some cases, AI's greatest value isn't that it writes something for me.

It's that it says something I disagree with.

When that happens, I'm often forced to clarify what I actually believe.

The predictive nature of AI can draw thoughts out of the author that might otherwise remain unspoken.

In that sense, it sometimes behaves less like a ghostwriter and more like a conversation partner.

What Is Authorship?

This is where I think many discussions about AI go off track.

We focus on who typed the words.

But authorship has never been that simple.

Writers use editors.

Writers use research assistants.

Writers use spell-check.

Writers use grammar-checking tools.

Many writers dictate.

Many writers collaborate.

Many use software that constantly suggests changes to wording and structure.

So where exactly is the line?

Do I need to disclose that I used voice-to-text?

Do I need to disclose that Microsoft Word corrected a spelling mistake?

Do I need to disclose that Grammarly suggested a better sentence?

Do I need to disclose that I used a calculator to verify a number?

Do I need to disclose that I typed this rather than handwriting it?

Somewhere along that continuum, the discussion stops being about tools and starts being about responsibility.

And that's where I think the focus belongs.

The Better Question

To me, authorship comes down to four simple questions:

Who supplied the ideas?
Who exercised the judgment?
Who verified the facts?
Who accepted responsibility for the final product?

If the answer to those questions is the human whose name appears at the top of the page, then the work remains fundamentally theirs.

So yes, AI helped me write this article.

It helped organize ideas.

It improved transitions.

It challenged assumptions.

It suggested alternative ways to structure an argument.

But it did not supply the experiences.

It did not supply the convictions.

It did not supply the judgment.

Those are mine.

Which is why I suspect the debate surrounding AI often begins with the wrong question.

The question isn't:

Did I write this, or did AI?

The better question is:

Who is ultimately responsible for the ideas, judgment, and truth contained within it?

Because calculators, spreadsheets, sampling machines, portfolio management software, voice-to-text systems, and AI will all continue to evolve.

The tools will change, just as they always have.

The human contribution remains.

The judgment remains.

The responsibility remains.

Or as Rick Rubin put it:

"Tools will come and tools will go. Only the vibe coder remains."

The tools change.

The responsibility does not.