There is a lot of talk about when we will reach AGI. The problem is that there still isn’t a standard definition of what AGI actually is. Depending on how we define it, we may already be there, getting close, or still be a long way away.
Literacy is a simple example. Being able to read and write was once a significant capability. As literacy became more widespread, it stopped being a meaningful measure of distinction. We started placing more value on higher education and specialized knowledge.
In technology, a computer beating a human at chess was once seen as an important milestone for artificial intelligence. Once computers became better than the best human players, chess didn’t stop being difficult or interesting. It just became a much less useful way of measuring general intelligence.
The Turing test may be heading in the same direction. If AI can routinely hold a convincing conversation, whether a machine can fool someone into thinking it is human doesn’t tell us as much as it once did.
We often set a boundary around something we consider difficult or uniquely human. Technology eventually crosses it. Once that happens, the boundary starts to lose its significance and we look for something else.
Think about a calculator. A human prodigy multiplying huge numbers in their head is seen as a genius. A calculator doing it is just arithmetic. The difference isn’t the capability—it’s whether the mechanism is visible.
Once we understand how a capability works well enough to mechanize it, it stops feeling like intelligence and starts feeling like computation. Mystery does a lot of the work in what we call intelligent.
We might define AGI around a particular set of capabilities, only to see AI reach them and then decide those capabilities weren’t really what we meant by general intelligence. This doesn’t make the progress any less real. It just makes the finish line harder to pin down.
At some point, the technical label of AGI won’t matter. We will keep debating whether we’ve reached it, while already living in a world shaped by systems doing things we once thought were impossible.