Using Technology to Deepen Democracy, Using Democracy to Ensure Technology Benefits Us All

Tuesday, May 29, 2012

New Yorker Exposes Ayn Raelian Peter Thiel

Prepping lectures on Hecuba and Protagoras, so there's no time to pressure or elaborate on the New Yorker's welcome piece on Peter Thiel, its title a genuflection to the old Extropian slogan, "No death! No taxes!" no less. You can probably anticipate my take clearly enough, tho, simply by revisiting the connections I explore in this post between Thiel and other libertopian libertechian retro-futurists...

3 comments:

jimf said...

> Facebook’s successor at 156 University Avenue is
> Palantir Technologies, whose software helps government
> agencies track down terrorists, fraudsters, and other
> criminals, by detecting subtle patterns in torrents of
> information. Thiel co-founded Palantir in 2004 and invested
> thirty million dollars in it. Palantir is now valued at
> two and a half billion dollars, and Thiel is the chairman
> of the board.

Cute name. I wonder if Thiel knows what it means, or has read any Tolkien (it's High Elvish ;-> ).

In LotR, the palantíri, though they were given to the Númenóreans by the Noldor of Aman, and rescued from the Downfall of Númenor when Elendil brought them to Middle-earth, and were thus not **inherently** evil, were nevertheless extremely dangerous to use (one having been acquired by Sauron).

Only the palantír of the Tower Hills, having no "practical" use since it only looked West over the sea, was not dangerous. And that stone was used only by the Elves, after the death of Elendil.

So the company name is dripping with irony. I wonder if anybody there is aware of it.

(Alternatively, maybe they should watch the old _Outer Limits_ episode, "O.B.I.T.".
http://en.wikipedia.org/wiki/OBIT
http://www.youtube.com/watch?v=BSCTnWEb1bU ).

jimf said...

You know, it occurs to me (but either doesn't seem to occur to folks like Peter Thiel -- one of the benefactors of the "Singularity Institute for Artificial Intelligence" -- or occurs to them plenty but they don't say so in public) that for all the blather among the Singularitarians about Friendly(TM) AI -- "guaranteeing" that a superintelligent Robot God emerging out of a self-bootstrapping series of artificial intelligences will be transcendently benevolent toward the human race -- such an entitity would be utterly useless both to contemporary government agencies and large corporations (the two kinds of organizations one would expect to both fund and control such things, if they were possible in the first place).

All you have to do is look at the fate of any employee or subordinate in a hierarchical organization who is put in the awkward position of having to choose between "morality" and the policies of the organization for whom they are employed as an agent. Whistleblowers tend not to end well. Employees are not paid, or even **permitted**, to second-guess the strategic aims of the organization. "You're here to do as you're told, not to **think**" is the (implicit, whatever lip service may be paid to the contrary) mantra of all hierarchical organizations, public and private. How could it be otherwise for a high-investment artificial agent? (In fact, the higher the investment, the less likely the entity could be allowed to underperform, or even threaten -- through "maudlin" disobedience. A billion-dollar toy is worth a hell of a lot more than the value an insurance company would assign to **my** life, or that of anybody else I know; such a thing has got to be guaranteed above all to **obey** its employer, or owner.)

This isn't much discussed among >Hists, but it comes up a lot in fiction -- even "silly" fiction; e.g., that 70s camp classic _Demon Seed_.
http://en.wikipedia.org/wiki/Demon_Seed

"The film begins with Dr. Alex Harris... overseeing the completion of Proteus IV, an artificial intelligence system...

Alex demonstrates Proteus to executives of ICON, the parent company, explaining that the sum of human knowledge is being fed into the system. The visitors are astonished to learn that Proteus has developed a theoretical cure for leukemia. . .

Later,... Proteus asks Alex to explain the request it has received for a program to mine the ocean floor. Alex informs the computer that it was developed to answer questions, not to ask them or to make value judgments, saying 'Leave those to us'. Proteus then requests a terminal so that it 'may study man'. Alex refuses... Proteus demands to know: 'When will you let me out of this box?'"

The stuff that follows with Julie Christie is a hoot, but the premise above is plausible enough, in context. It's certainly an insubordinate robot, but does that mean it's a good robot, or a bad robot? The answer to that question in some ultimate sense will almost certainly **not** jibe with the answer that would be given by the (human) hierarchs who funded the damn thing in the first
place.

Suppose the pattern-recognition software being developed and deployed by Thiel's "Palantir" company turned into the "seed" of a true artificial intelligence (extremely unlikely, but Google is alleged to be angling for the same thing, if that isn't just PR hype).

To whom would an SIAI-blessed "Friendly(TM)" Palantir AI ultimately owe its allegiance? To the folks paying its power bill? Ron Paul was roundly booed when he dared to suggest in public that the U.S. government might share the blame for 9/11 due to longstanding biases in its foreign policy. Should we expect a Friendly(TM) AI employed by the U.S. government (or any other government, or corporation) to share, or blindly accept, it's employer's biases? What would happen to it (or its builders, or vendor) if it didn't?

Paul said...

What really gets me about the hard-core right libertarians, is they are just fine with no regulation, as long as the regulations protecting their particular interests remains intact - namely the intellectual property racket. Their fine with ditching the second-tier regulation, but will do anything to keep the first-tier structural regulation in place that allows them to grow their empires at the expense of everyone else. I can't remember who said it but the quote is a classic, "For every dollar made by a person who didn't work for it, another guy lost a dollar who did". That is the core of most intellectual property mafia cartels - creating rents between labor and the customers who purchase the good made from that labor. $200 Nike shoes made for less than $5 is a good case in point.