As I explained to Wired readers back in the day: “Meme is my newsletter, delivered once a month via electronic mail to 4,400 subscribers… What counts is who reads it.” I was, in today’s language, an early influencer. The early internet offered “an idealized picture” where “ideas flow freely” — and I desperately wanted to believe it could work.

What began as a land of artisanal knowledge-makers cultivating gardens of wisdom became an industrial farming operation powered by social media platforms. On the internet of 2026, people speak of “infobesity” — gorged on processed information from unknown sources, an intellectual abattoir where bits of everything are thrown together with one purpose: to hook us on the feed. We know it’s bad for us. We can’t put it down.

In this slurry of addictive information, people craved the authentic — and influencer culture emerged as a tonic. Here were native guides doing the research for us, helping us make sense of the low-stakes (handbags, smoothies) and increasingly the high-stakes: longevity, retirement, parenting. Yet as with so much that starts as sincere human-to-human connection online, the higher the stakes, the more conflicted the influencer’s role became. Is it entertainment or discernment?

In spring 2020, it became a survival technology. A terrifying pandemic with no vaccine and no easy explanations sent millions of us to our screens for answers. Into this panicked information vacuum stepped people with charisma — “rizz” — who did their own research on COVID and taught us, maybe, how we could survive. Distrust of expertise became the feature, not the bug.

Into this landscape, ChatGPT burst onto the scene in November 2022, famously becoming the fastest product to reach 100 million people. Now the power to “do my own damn research” reached new heights. More information processed by machines whose provenance was even more mysterious than that of an influencer. A synthetic persona — a machine influencer — could now compete with the human influencer, feeding more processed intelligence into already-processed intelligence. It came to be known, aptly, as “AI slop.”

The Suspicion Didn’t Land on AI Alone

The human in the loop didn’t diminish in principle. AI just poured gasoline on an already-burning fire. When the provenance of everything online became suspect, the suspicion spread to everyone — influencers included. “How do I know you know what you’re talking about?” became the question directed at every voice online. When it came to health, wealth, and love, rizz was increasingly insufficient to carry the day.

What is your credential to tell me how to parent? What gives you the authority to advise me on my GLP-1 shots? What do you really know about the use of psychedelics to treat trauma?

Expertise Is Becoming the Scarcest Resource Online

The wheel is turning back toward expertise — credentialed professionals with original, hard-won knowledge, not merely fast and fabulous opinion. But with a twist: AI can now powerfully augment that expertise. Imagine world-class experts pairing their lifetime of knowledge with a custom AI trained exclusively on their know-how — the juicy stuff never put online for the big models to scrape. Think of this as “Whole Knowledge” — like a whole food, spiky bits and all, with real provenance. The intellectual terroir of a human mind, made available to those who need it most.

In this realm, the role of the charismatic influencer bifurcates. Where the stakes are low — fashion, travel, where no one is likely to die from misinformation — it will remain entertaining to follow an influencer. Where the stakes are high — illness, money — rizz won’t carry the day. Experts holding a world-class asset — their proprietary lifetime knowledge — will find themselves in a privileged position. No longer constrained by the physical limits of their daily schedule, they can scale their wisdom in ways previously unimaginable.

Already, smaller AI models trained on a narrower body of knowledge — a clinician’s lifetime of writing and speaking, for instance — are outperforming frontier models on domain-specific tasks. Smaller models also enable something anathema to the frontier model: a distinct voice and perspective. The public appears ready for this change — a recent Gartner study found that half of consumers now actively prefer companies that avoid generative AI in their marketing.

Take Dr. Becky Kennedy, the clinical psychologist who became the parenting whisperer to an entire millennial generation. Kennedy recently announced she had sold over 100,000 subscriptions to her Good Inside platform — which includes 24/7 access to an AI model trained exclusively on her knowledge — grossing $34 million last year. She raised a reported $10 million to make it possible.

The point is not to automate human connection. It is to rebuild trust in a world where trust has been broken. When synthetic slop makes the illusion of expertise free, the value of genuine, verifiable wisdom skyrockets. We are moving out of an era where technology rewarded those who captured the most attention, and into an era where technology scales those who hold the deepest knowledge.

In my 1998 Wired piece, I wrote that the internet was a place where “power comes not from wealth, but from thought.” In 2026, that idea’s time may have finally arrived.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.