Peter Thiel: Artificial General Intelligence Isn’t HappeningThat whole transhumanist movement is slowing down, he told COSM 2021. But, he adds, What IS happening should sober us up a lot
In his talk yesterday at COSM 2021, venture capitalist and philanthropist Peter Thiel — the ultimate Silicon Valley insider, prophet, and sometimes needed gadfly — offered a cold shower for transhumanism, The Singularity, the computers we will supposedly merge with by 2030, and all that.
Those things, he thinks, are uncertain. We should worry about what’s happening now in everyday time, to which, in his view, too few are paying heed: The growth of total AI-based surveillance and the disappearance of privacy.
Thiel considers arguments about whether computers that think like people will ever be developed to be “above his pay grade.” Given that he is reputed to be worth $3.7B dollars, that’s a polite way of saying that such arguments are a pleasant waste of time.
Always skeptical of buzzwords, he went on to say he hates the word “AI” (artificial intelligence):
AI is the worst buzzword out there. And one of the reasons it’s such a terrible buzzword is it can mean just about anything. So it can mean something about futuristic computer technology, the next generation of computers, the last generation of computers, and anything in between.
Or whatever. He pointed instead to today’s concerns: Specifically, surveillance AI, which involves “fairly dumb computers” whose real potency is that they “monitor us in all places and all times.” That will shift the balance of society away from civil liberties toward centralization, he warned:
You can think of the crackdown in China or Hong Kong, where you have facial recognition software. This is not futuristic, super, super smart AI, but it is sort of a next generation technology.
He told the gathering that he had been involved in the debate over artificial general intelligence for 20 years in Silicon Valley and “One of the things that I’m struck by [when] we talk about AGI in 2021 is how much less intense it’s become than it was six or seven years ago.”
He offered an anecdote: His venture capital firm had invested in DeepMind, best known for Deep Blue and AlphaGo, whose founder was Demis Hassabis. He asked Hassabis to speak at an investors’ meeting in 2012:
And it was, it was a very powerful speech. And, you know, one of the, one of the investors came up to me afterwards and said, “Wow, that was really incredible. I thought I had to shoot the speaker. This was the last chance to save the human race.
Waste of time. Not too long afterward, DeepMind was acquired by Google. Whether or not Google is running the world has been offered for debate but the machines aren’t. Or, for sure, not all by themselves.
Thiel offered another anecdote, this time featuring Elon Musk in, maybe, the fall of 2013:
One of my colleagues introduced Elon and Demis to each other and the conversation went something like this:
Demis: I’m working on the most important problem in the world. I’m working on superhuman intelligence.
Elon: Well, I’m working on the most important problem in the world. I’m working, making our species interplanetary. I’m going to Mars!
Then, Demis responded, Well, you know, my, my AI will be able to follow you to Mars.
And, and at that point, Elon was a little quiet.
Thiel thinks the acquisition of DeepMind by Google helped Google see itself as a futuristic company building the future. But apart from conquering games, he noted, “The energy has, you know, deintensified a great deal, where, you know, Elon doesn’t talk about it very much anymore”:
Elon doesn’t talk much about it anymore? That matters, in Thiel’s view:
And I think of him as — I don’t know, a hundred million or however many Twitter followers he has — like a weathervane for the zeitgeist. And he knows nobody cares. And so he doesn’t talk about it anymore.
And Larry Page?
You know, who was driving it at Google, sort of retired to Fiji or something like that at the end of 2019. And, as far as we can tell, doesn’t seem to care about it that much anymore.
He summed up: “Almost all the paths that lead to AGI, as it is currently conceived, go through giant organizations with giant databases, looking at people, modeling people, doing machine learning on people to build the AGI. You need this sort of surveillance to get to the AGI. And of course the surveillance a AI has a sort of a creepy totalitarian undercurrent.”
Indeed. Before the computer does our thinking for us, if ever, we may have long since become afraid to think for ourselves.
You may also wish to read: Facebook (Meta) is strengthening, not dumping, facial recognition. They’re getting rid of the annoying parts but read the fine print. We are to be informed in due course what DeepFace is going to be used for in Zuckerberg’s new, admittedly sci-fi, Metaverse.