Microsoft and OpenAI's close partnership shows signs of fraying

The question is how rigorously defined is AGI in their contract? Given how much AGI is a nebulous concept of smartness and reasoning ability and thinking, how are they going to declare when it has or hasn't been achieved. What stops Microsoft from weaseling out of the contract by saying they never reach it.

It’s almost like a contractual stipulation of requiring proof that one party is not a philosophical zombie.

OpenAI's short definition of AGI is:

A highly autonomous system that outperform humans at most economically valuable work.

Will AGI be able to stock a grocery store shelf?

Some of those works would need a tight integration of AI and top notch robotic hardware, and would be next to impossible today at acceptable price. Folding shirts comes to mind; The principle would be dead simple for an AI, but the robot that could do that would cost a lot more than a person paid to do that, especially if one expects it to also be non specialized, thus usable for other tasks.

Sometimes it is more narrowly scoped as “… economically valuable knowledge work”.

But sure, if you have an un-embodied super-human AGI you should assume that it can figure out a super-human shelf-stocking robot shortly thereafter. We have Atlas already.

Which is funny, because what they’ve created so far can write shitty poetry but is basically useless for any kind of detail-oriented work - so, you know, a bachelors in communications, which isn’t really the definition of “economically viable”

No, but it might be able to organize a fleet of humans to stock a grocery store shelf.

Physical embodied (generally low-skill, low-wage) work like cleaning and carrying things is likely to be some of the last work to be automated, because humans are likely to be cheaper than generally capable robots for a while.

I think I saw the following insight on Arvind Narayanan's Twitter, don't have a specific cite:

The biggest problem with this definition is that work ceases to be economically valuable once a machine is able to do it, while human capacity will expand to do new work that wouldn't be possible without the machines. In developed countries machines are doing most of the economically valuable work once done by medieval peasants, without any relation to AGI whatsoever. Many 1950s accounting and secretarial tasks could be done by a cheap computer in the 1990s. So what exactly is the cutoff point here for "economically valuable work"?

The second biggest problem is that "most" is awfully slippery, and seems designed to prematurely declare victory via mathiness. If by some accounting a simple majority of tasks for a given role can be done with no real cognition beyond rote memorization, with the remaining cognitively-demanding tasks being shunted into "manager" or "prompt engineer" roles, then they can unfurl the Mission Accomplished banner and say they automated that role.