Q* + Tigris + EQ + 4D + The Power of a Haka
Please find the link here to a piece I wrote for Unherd here on the possible role China is playing in the Open AI drama. It is a simpler follow-up to my last article here (Q* + Tigris = ?) and is designed for a general readership.
In short, China and the US both want Sam Altman’s skill sets and vision, but they don’t want him to achieve them independently of themselves, let alone with their opponents/competitors. But, the superpowers are still assuming that nation states and borders still matter. Their thinking assumes we can tell where AI resides when we can’t. It assumes we can identify a company as “American” or Chinese, but can we really, in a world where the investors come from everywhere? All this assumes that borders have meaning in the modern tech world when the tech is eroding the meaning of borders and nationality. The public is concerned about the immigration of humans but clueless about the immigration of algorithms and databases. We want to know “where is my data”? The answer is that it’s already all over the net and accessible from everywhere. We think removing our name from the dataset protects us when, in fact, it’s easier to identify an individual just from their data history. Everyone’s data is unique. Your digital twin gives away your physical identity. So, we can try to fight this (follow the Palantir NHS story), but the reality is that we are all digital emigres now. Your digital twin is spread around like a thin layer of you that spans the entire globe. Your data interacts with the news flows, which guides the algorithms to choose what news feeds you see and which ones you don’t. It precedes you and defines you before you even enter a space. It’s like a 4D version of you.
This is all about the ceding of power and control not just to a state but to a new hyper object called AI that is becoming smarter than humans at an incredible pace. Ray Kurzweil predicted that computers would become “smarter” than humans by 2023. He seems to have been spot on. Scientific American’s article on the IQ of Chat GPT is worth reading. Eka Roivenan wrote, “I gave Chat GPT an IQ Test: Here’s What I Discovered.” The Wechsler Adult Intelligence Scale (WAIS) was applied. The result was that “the Verbal IQ of the ChatGPT was 155, superior to 99.9 percent of the test takers”. Some say that Chat GPT will be between 280 and 1000. The average IQ of a human is 85-115. What nobody is testing for is EQ. We all know superbright people who can’t navigate through a grocery store because they lack EQ.
Now, this is the problem that AI engineers are NOT trying to solve. Instead, they are trying to deal with human stupidity. They want a higher IQ to decide the right course of action. So, it is very interesting to find that Larry Summers, the Former Treasury Secretary and newfound board member at Open AI, was interested in the use of Q Tables to generate reinforcement learning and automation of public policy decades