Of Dollars and Data writer frames AI risk as a test of originality

AI is getting better at producing the kind of work many people get paid to do, and that puts a premium on what machines cannot actually have: lived experience. In an essay on Of Dollars and Data, writer Nick Maggiulli argues that the next career challenge is less about resisting AI outright and more about deciding whether to chase easy rewards or protect original judgment.

Maggiulli builds the piece around HBO’s Hacks, the comedy series starring Jean Smart and Hannah Einbinder. He says the show captures a familiar split in creative work: one character is more focused on commercial success, while the other is more focused on truth and authenticity.

He uses that contrast to describe a broader divide between what he calls “hacks” and “artists.” In his framing, hacks optimize for attention, status or money, while artists care more about the quality and meaning of the work itself. Maggiulli says most people sit somewhere between those two poles, including himself.

The idea is not limited to writers or performers, according to Maggiulli. He applies it to other jobs, including financial advice, where he says the difference shows up in whether clients are treated mainly as people to serve or as revenue to capture.

Why easy rewards can narrow creative work

Maggiulli points to “mode collapse,” a term from artificial intelligence, to explain why both people and machines can start producing similar work. Mode collapse refers to a generative model losing variety and repeating a narrower set of patterns that reliably succeed.

He cites writer Henrik Karlsson, who has argued that newer language models often sound more polished but less surprising than earlier systems. Karlsson also argues that humans can experience a similar narrowing when they are rewarded for acceptable work rather than creative risk-taking.

For workers watching AI improve, Maggiulli says the issue is practical. Large language models, or LLMs, are AI systems trained to generate text and other outputs from patterns in large datasets. They can be fast, inexpensive and knowledgeable, but Maggiulli argues they do not possess personal experience.

That distinction is central to his disagreement with a comparison made by writer Dan Koe. Koe has argued that AI resembles Gutenberg’s printing press because it can make an old skill set less valuable, much as the press reduced demand for scribes who copied manuscripts by hand.

Maggiulli accepts Koe’s broader point that people who learn to use AI well may outperform those who do not. But he argues that the printing press replaced a task where the individual identity of the worker mattered less, while AI is now being applied to fields where a creator’s background, taste and experience shape the output.

The career takeaway

Maggiulli says he cannot control how AI develops or how people choose to use it. He says the controllable part is personal behavior: whether a worker produces interchangeable output or work shaped by their own judgment and experience.

His conclusion is pointed but not anti-AI. He agrees with Koe’s view that the next 24 months could influence a person’s career path for many years, while arguing that the choice is also about standards. In Maggiulli’s framing, the question for workers is whether they use new tools in a way that makes their work more generic or more distinctly their own.

This story draws on original reporting from Of Dollars and Data.