The End of Average

I’ve gone back and forth on whether intelligence—the hard-to-quantify but definitely real attribute we all possess—is becoming more or less important in the age of AI. The internet and search engines effectively eliminated librarians as a profession. There are people who work at libraries, but there was a time when law and consulting firms would have in-house librarians. The implication is that when machines became capable of directing us to the right information, we no longer needed humans to do so.

It would be incorrect to extrapolate that processing information (e.g., learning) would become obsolete. Universal access to information didn’t result in the marginal utility of knowledge going to zero. What changed is that capable individuals with the right appetite were able to satisfy their informational needs, regardless of social class or pedigree. Hence, we’ve gone through a generation where formerly insular fields have been disrupted by outsiders. When information lived behind analog walled gardens, competition was limited to those with access. In a world of free information, competition becomes fiercer, and humanity’s best become better.

We now find ourselves in a transition toward intelligence as a commodity. To a first-order approximation, every human on Earth can outsource intelligent tasks to an AI. Does that imply that being individually intelligent ceases to be a differentiator? I believe it’s the opposite. As the tide of free intelligence rises, many people will be left behind.

In the age of free information, it was sensible to consult an attorney to prepare a legal document, even if we knew they were copy-pasting most of it. There was too much liability in the unknown unknowns. Today, one can find a standard contract, consult with an AI line by line, understand the motivations and implications, and recognize if something is unusual. It’s not perfect, but most lawyers are below average (power law), while the AI improves exponentially. My bar (pun intended) for consulting with a lawyer is moving up.

There are a couple of implications of this. First, I need to be smart enough to articulate my objective and understand the guidance. This was also true before AI lawyers, but now it’s more visceral. Second, in instances where a human is required (e.g., a litigator), it will be obvious we are not dealing with one lawyer among all lawyers, but one among a progressively more specialized subset. Naturally, they will get paid more. And because more and more humans will realize they have an average lawyer in their pocket, there’ll be more contracts, more lawsuits, and more demand for the special few human attorneys. Hence, “smarter” attorneys will get richer as AI improves.

In summary, before, intellectual tasks required either innate intelligence or a transaction with someone else. Today, that transaction is trending toward gratis. It was always true that one could borrow intelligence in proportion to one’s own, in the sense that an objective must be established and posterior actions taken. But it’s more apparent when the cost of borrowed intelligence goes to zero. Meanwhile, markets used to pay humans for intellectual tasks that can now be outsourced to AI. Hence, only humans with superior intellect will attract market bids. Since there’ll be fewer of them, they’ll command higher rates. Before, humans who operated below their limits could improve output through persistence. Because every human will be getting closer to their theoretical capability through the leverage of free intelligence, demand for superhumans will increase. This is, of course, a temporary process until the ASI.