Seriously, it’s not. What Google or OpenAI may claim, and with some justification, this year, the next, or in 5, may be a type of AGI. But it’s not all we need for the Artificial Intelligence journey to be over, by far.
Big Tech making AGI is like a developer saying “it works on my computer.” It’s not a bad sign, but true scientific and academic achievement of AGI requires much deeper standards. Starting with: we know how it works, we know why it works, we know how to make it again, and we can explain it to others so that they can work on a better version.
An invention without theory is not science. For understanding AGI—its workings, construction, confirmation, reproducibility, and philosophy—extensive work, nearly orthogonal to the looming invention, is needed.
We probably need to adapt the ways we think about engineering, science, research, and computers, not to mention intelligence itself.
We could shift to an open standard for ethical, collaborative computer design and a more cooperative academia-industry environment for this, but the current systems do not really support this move. This is not a next-quarter earnings move.
The way we do science evolves slowly, and almost unnoticeable by any single person, except the visionaries. The goal is not just a mind in a cloud, but a mind integrated into, and understood by society. This is not coming within the next decade at least. It seems that we’re headed more to a chaotic state where we don’t know if AI is really helping us individually, if we are using AI for any particular project indirectly, or whether we should, or what AI even is.
This is hard to argue with. The more AI develops, the less we understand it in the current paradigm. And that’s a sad state of affairs, considering that capitalism invests ever more resources into its development. We’re facing not existential risk, but constancy risk, and people are sticking to the old ways of distributing wealth, money and salaries, because they don’t know if there’s anything on the other side.
So, the market has decided that AGI is not three letters. It’s not even Human-Centric AGI, which is quite a bit more. What remains is for us to understand what the three letters we’re currently getting mean for individuals, for science, and for society.
— END POST