One step further, our current generative AI is not discrete. “AGI checking other AGI” doesn’t really make technical sense. Currently, every interaction between two generative AI units is treated as agglomerative. We talk about “Claude”, not the 70 subagents that Claude spun up to achieve a task.
In other words, as soon as two generative AIs interact, they become one. Our current definition of AI (generative AI operating in feedback loops) is dependent on that.
Edit: this is somewhat an epiphany to me. Our current generation of “AI” isn’t an “entity”, it’s a “process”. I guess it’s hard to define formally, but I would compare it to how law and the pursuit of justice is a process, not an entity.
And I think that’s a fundamental limitation to achieving artificial general intelligence.
Yes, and that is why you can't say "its as smart as a single human" and think its as smart as humans, because humans can get smarter by cooperating, if the AI can't do that to create a smarter AI it is not as smart as humans, meaning an AI that is not as smart as a team of humans is essentially not as smart as a single human either.