In 1996, in a PBS documentary called Triumph of the Nerds, Steve Jobs said something that has become one of the most quoted lines in business history.
The line, paraphrased: Picasso had a saying — good artists copy, great artists steal — and we have always been shameless about stealing great ideas.
The line landed. Three decades later, it is the unofficial motto of Silicon Valley. Almost everyone who quotes it gets two things wrong.
The first thing that’s wrong
Picasso almost certainly never said it. There is no documented Picasso source for the line. The actual documented source is T.S. Eliot.
In a 1920 essay about the Elizabethan playwright Philip Massinger, Eliot wrote something significantly more interesting than the version we remember.
The full Eliot quote: “Immature poets imitate; mature poets steal; bad poets deface what they take, and good poets make it into something better, or at least something different.”
Read that last clause again. Bad poets deface what they take. Good poets make it into something better, or at least something different.
That second clause — the one almost nobody knows — is the part of the framework that actually matters. The Jobs/Picasso version drops it entirely.
The second thing that’s wrong
When Jobs invoked the quote, he was talking about Apple’s 1979 visit to Xerox PARC. What Apple did there is usually described as “copying the graphical user interface.” That description is wrong in a specific way.
The Xerox mouse had three buttons and cost $300. Apple’s had one button and cost $15. The Xerox GUI required a $30,000 workstation. Apple put the same concepts on a $2,495 Macintosh. The Xerox document editor required weeks of specialized training. Apple’s word processor could be used within hours.
Apple stole the idea. Then they made it into something better. The stealing was the entry condition. The transformation was the achievement.
The framework that matters
The simplified Jobs/Picasso version says: great artists steal.
In the hands of an executive who hasn’t read the Eliot original, this becomes a permission structure: we can copy what other companies are doing without doing the harder work of transforming it for our context.
The Eliot framework says the opposite. The stealing is the entry condition. The transformation is the achievement. Bad poets deface what they take. Skip the transformation work and you produce the defaced version — the AI pilot copied from someone else’s successful pilot, deployed without context-specific adaptation, which fails publicly and produces internal cynicism about AI more broadly.
The 2026 application
Every senior leader running an AI transformation in 2026 is facing this choice. Should we develop original approaches? Or should we openly steal what’s working from everywhere and adapt?
The Eliot/Jobs framework says: steal openly. Steal the use cases, the architectures, the workflow redesigns, the change management approaches, the vendor selections — whatever is working, wherever it’s working. Do not waste organizational cycles on false originality.
Then do the harder work.
Context-specific adaptation. What worked at a tech company will not work in a bank, a hospital, or a manufacturer without substantial modification. The vendor case study you are copying from was built for someone else’s context.
Integration into your existing system. The pilot that succeeded at another company succeeded because it was integrated into their specific workflows. Your workflows are different. The integration work is the transformation.
Aesthetic and human-experience resolution. The Apple-vs-Xerox difference was overwhelmingly about human experience. One-button mouse. Consumer-affordable price. Hours instead of weeks of training. Your AI implementation needs the equivalent.
Deliberate differentiation from the source. If your implementation looks exactly like the source you stole from, you have not transformed. You have defaced.
Three practical questions
One: what specifically are you stealing, from where? Name three sources. Articulate what you’re taking from each. The specificity forces honesty about the origin of your approach.
Two: what is the transformation work that has to happen? Different industry, different data, different culture — name the specific gaps between the source implementation and yours.
Three: who is responsible for the transformation work? Not the technical implementation. The transformation. If the answer is “no one,” your pilot will produce the defaced version.
The closing thought
Eliot understood something in 1920 that Silicon Valley has repeatedly forgotten. Stealing is not the achievement. Stealing is the entry condition. What you do with what you stole is the achievement.
Apple stole from Xerox. Then they made it into something better. The Macintosh sold. The Xerox Alto did not.
In 2026, thousands of AI implementations are being copied from vendor case studies and deployed with no meaningful transformation work. Most will fail. The ones that succeed will be the ones that stole openly and then did the harder work Eliot pointed at: making it into something better, or at least something different, for their specific context.
The world has changed. The leaders who notice will be the ones the next decade is built around.
