Why I'm Feeling the AGI

> I arrived at [these views] as a journalist who has spent a lot of time talking to the engineers building powerful A.I. systems, the investors funding it and the researchers studying its effects. And I’ve come to believe that what’s happening in A.I. right now is bigger than most people understand.

> In San Francisco, where I’m based

In other words, the man (a) is not an engineer or specialist, (b) gets his information secondhand (c) from the people most heavily invested, literally and figuratively, in these developments, and (d) even lives in the same city as many of these people. I don't normally like the term "echo chamber," because it's fine for example, for example, not to "teach both sides," but if these are his qualifications and his sources, he shouldn't be taken as seriously as even the marketing copy that these companies circulate.

> I believe that the right time to start preparing for A.G.I. is now.

Nobody, including the author, has satisfactorily explained to me what the hell this means. Even if the breathless hype is true, what are they saying I should do?

Should I just, like, think real hard about it, because Something Big This Way Comes? If so, please show me the evidence that people are to be trusted when they make prognostications about the future of technology, even in the near term. If people were good at it, more of us would have retired after making millions on Nvidia a year or two ago.

If instead I'm to believe there are practical consequences and measures that I should take, then the people saying this are, I think, making an incoherent, self-defeating argument:

Is their argument that AGI, when it's available, will bear a close resemblance to the current tools, that we'll interact with them similarly, and that they'll be similarly integrated into the things that current integrate A(non-G)I?

If so, then why do they think AGI will Change Everything, rather than just becoming another fucking customer-service chatbot?

In turn, if they think AGI will Change Everything(TM) and will be qualitatively different from anything that came before (Ooo, "paradigm shift"), why would I waste my time on the current tools, especially considering their (in my experience) mediocrity and worthlessness for work of any significance? If my best hope for survival in the new technological order is to "start preparing," what is the prescription? It seems, conveniently, to be to buy in and pay the companies making tools unlike, and inferior to, the ones I'm supposed to expect in the near future -- or, at a minimum, just to use the stuff they make. Funny how that works, isn't it?

No, thank you. I share Jaron Lanier's skepticism and profound boredom with these tools and his deep mistrust of the people selling and promoting them. They are autocomplete token predictors. They're statistical models on steroids. And they're becoming just another way for owners of capital to squeeze value for themselves from the labor of others without paying for it -- theft via matrices.

I say this in every thread about AI, so I'll say it here as well: to say that AGI is just around the corner because LLMs producing polished-looking output is no different from mistaking special effects for the real thing, or looking at a picture of a dog on a computer screen and believing there is a real, actual dog inside the computer -- or that dogs inside computers are just around the corner.

> Today, software engineers tell me that A.I. does most of the actual coding for them, and that they increasingly feel that their job is to supervise the A.I. systems.

Utter nonsense. And to the extent that it's true, it is at least as true that these tools are effectively stack overflow copy+paste cud, albeit with less work on the part of the developer, which means users learn less, understand less, and are worse at their jobs.