Is AGI already here? Well... it depends on what you mean by "intelligence." If intelligence is indistinguishable from the articulation of meaningful responses, then we're already close (or maybe there).
But for me, the critical boundary between human and artificial intelligence is the possession of tacit knowledge. Tacit knowledge, succinctly, is that knowledge that comes only from experience--qualia as philosophers dub it. You can know what salt tastes like only by tasting it for yourself.
Can AI ever possess tacit knowledge? I'm quite sure it doesn't have it yet and I'm pretty skeptical that it will ever have it. Of course, we don't actually know how or why we experience qualia, which David Chalmers famously dubbed "the hard problem of philosophy," so we can't really predict if AI will ever suddenly begin to have experience or not.
The instinctive retort, of course, is that it probably doesn't matter, since it will inevitably *seem* as if they have experience. It can already talk as if it has experience.
But here's why it matters. A TON of our intuition and judgment comes from tacit knowledge. What will I like on the menu? How must they be feeling right now? Will other people pay for this new product idea?
Of course, the behavioral sciences have long argued that human judgments are highly error-prone and so mechanizing judgment algorithmically may improve it, reducing common biases. But this is a fundamental misunderstanding of judgment, I think. The most *human* aspects of judgment manifest most clearly where there is, and can be, no "right" answer. This sort of judgment is made primarily from tacit knowledge, or what is often called "intuition" or sometimes "system 1."
AI will be able to give us answers, but I expect we'll learn pretty quickly that AI doesn't *feel* like we do and so never quite gets us like other people do. And that will be a massive and persistent barrier that will leave plenty for humans to still do.
I'm calling it. AGI is already here – it's just not evenly distributed yet.