Goodbye to intelligence. What remains for us is character
Before AI, intelligence had a clear working definition, even if the scientists disagreed on how to phrase it. We called a person smart when they remembered a great deal and recalled it fast, understood from half an explanation, connected distant things, and answered under pressure with no time to look anything up. Schools and companies built their tests on these signs from the early twentieth century onward.
These signs were honest for one reason: they were expensive. Memory, speed, fluency, and technical skill take years for a person to acquire, so whoever had them had proven they did the work and that their mind worked.
What broke?
AI made these signs free. It remembers more than any person, answers faster, writes fluently, and codes today at a level beyond most programmers. Calling a person smart for knowing a lot or coding fast has become like calling a man fast because he arrived before everyone else in a car. The measurement lost its meaning.
That is why many people now ask for a new definition of intelligence. I think the sharper question is: when capability is free, what is a person still for?
Three things remain
In my view the answer is three things, and AI can do none of them on your behalf:
- Wanting: knowing what you want. AI answers any question, but it does not choose the question. Choosing the question, and refusing the ninety-nine others that were available, is the scarce act now. AI may suggest better goals than the ones you had in mind, but they stay suggestions until someone who loses something if they fail adopts them. Only then do they become a decision with weight and a price.
- Owning: standing behind the decision. AI pays no price for being wrong. The person who signs the decision, pays for it, and adjusts when reality disagrees is doing something no capability replaces. The answer has become cheap. Standing behind it is what costs.
- Taste: telling the good from the merely convincing. When answers are free and unlimited, the skill moves to telling the right answer from the one that only looks convincing, and the answer that is enough from the one that sounds logical but is wrong. We used to value whoever produced more. Today we need whoever rejects more, because producing is no longer the problem.
What if artificial general intelligence (AGI) arrives?
The ready objection: general AI, the kind that surpasses us at everything, is coming soon, and then all three fall. I think two of them hold because they are a role a person carries, not merely a skill a person owns. Wanting and owning mean that someone loses if the decision is wrong. If AI runs your company entirely, it does so on your behalf, but the name, the money, and the reputation stay yours. General AI takes execution from you and leaves you the choosing. Choosing then becomes your only job, and it is the hardest part.
Taste AI will partly absorb, because general AI will recognise its own good answers. But taste is finally a preference, and the preference belongs to whoever the work is for. AI gives you the best five ways to run your company, and you alone know which life you want, because that is a question with no correct answer.
Suppose the fastest forecasts are right and all capability now sits with the machine. Then "smart" stops being a useful word for ranking people, the way "strong" stopped after the engine. Nobody was diminished by that. Strength became a hobby and a measure of health, and stopped being a status. Intelligence will go the same way: a personal quality its owner enjoys. And when intelligence stops separating people, what decides who leads? The same criteria: who knows what they want, who stands behind the result, and who has the taste to guide it. So the three survive even in the most extreme case. They only lose their old name. We used to call them intelligence. We will call them character.
What I see from where I stand
When I started working, the difference between one company and another was plain: who knew more, and who executed faster. Today execution is available to everyone at nearly the same quality: the code gets written, the plan gets drafted, the reference gets found, all in minutes. Competition has moved to the two edges of the work. Before execution: who decides what gets built in the first place. After it: who notices that the result is wrong even though it looks right. Whoever works at those two edges looks smarter today than ever, and whoever's whole value sat in the middle suddenly looks ordinary.
The definition I work with: a smart person is one who asks the question worth asking, tells a good answer from one that only looks convincing, and takes responsibility for acting on it. The old definition held a piece of this under the heading "adapting to new situations", buried beneath the items on memory and speed. AI simply brushed the dust off that meaning.