All of cjb's Comments + Replies

2011 has been a huge year for Artificial Intelligence. With the IBM computer Watson defeating two top Jeopardy! champions in February, it’s clear that the field is making steady progress.

Do people here generally think that this is true? I don't see much of an intersection between Watson and AI; it seems like a few machine learning algorithms that approach Jeopardy problems in an extremely artificial way, much like chess engines approach playing chess. (Are chess engines artificial intelligence too?)

I actually do think it's a big deal, as well as being flashy, though not an extremely big deal. Something along the lines of the best narrow AI accomplishment of any given year and the flashiest of any given 3-5 year period.

4brazil84
Further to my previous comment, I found the second final Jeopardy puzzle to be instructive. The category was "US Cities" and the clue was this: A reasonably smart human will come up with an algorithm on the fly for solving this, which is to start thinking of major US cities (likely to have 2 or more airports); remember the names of their airports, and think about whether any of the names sound like a battle or a war hero. The three obvious cities to try are Los Angeles, New York, and Chicago. And "Midway" definitely sounds like the name of a battle. But Watson was totally clueless. Even though it had the necessary information, it had to rely on pre-programmed algorithms to access that information. It was apparently unable to come up with a new algorithm on the fly. Probably Watson relies heavily on statistical word associations. If the puzzle has "Charles Shulz" and "This Dog" in it, it will probably guess "Snoopy" without really parsing the puzzle. I'm just speculating here, but my impression is that AI has a long way to go.
3brazil84
I found Watson to be pretty disappointing. For one thing, it's big advantage was inhuman button-pushing speed since an actuator is much faster than a human finger. Now, one might argue that pushing the button is part of the game, but to that I would respond that reading the puzzles off of the right screen is part of the game too and Watson didn't have to do that -- the puzzles were inputted in the form of a text file. Also, travelling to Los Angeles is part of the game and Watson didn't have to do that either. If the game had been played in Los Angeles instead of New York, then all of Watson's responses would have been delayed by a few hundredths of a second. Another problem is that a lot of the puzzles on Jeopardy don't actually require much intelligence to solve particularly if you can write a specialized program for each puzzle category. For example, I would guess a competent computer science grad student could pretty easily write a program that did reasonably well in "state capitals" And of the puzzles which do require some intelligence, the two human champions will split the points. I'm not saying that Watson wasn't impressive, just that it's win was not convincing.

Thanks. I guess the followup questions are:

  • Is the legal action still pending, or can the situation be talked about openly now?
  • Has SIAI been able to recover the money?
  • Was it a mistake to trust a contractor with access to >$100k of funds? Do they still do that?

My understanding is that the case is ongoing in criminal court, at least as of a few weeks ago, and that the money has largely not yet been recovered. As far as I know, only that one contractor had the relevant financial access, which was required for the job, but obviously the financial controls on that access were not sufficient. I think that currently only the President and COO have the relevant access to the accounts (though others, including the majority-donor board, have limited access to monitor the accounts).

  • In 2009 the SIAI reported $118,802 to theft - "Misappropriation of assets, by a contractor [...]" This is a significant amount when compared to annual revenue or liquid assets. The year's surplus appears to have been eaten up by the theft. No details are provided, other than the fact that suit has been filed to seek restitution.

I'm surprised that no-one's mentioned this -- it's hard to imagine how someone can steal that much money. Can someone at SIAI tell us whether they're allowed to talk about what happened; and if you can't right now, do you have any idea when you might be able to?

The theft must have been discovered to be more extensive than thought, because one early report says

Embezzlement report: Alicia Isaac, 37, of Sunnyvale arrested on embezzlement, larceny and conspiracy charges in connection with $51,000 loss, Singularity Institute for Artificial Intelligence in 1400 block of Adams Drive, Dec. 10.

Which is significantly less than $120k.

Michael Vassar sent out an email with more information back in Dec 2009 (shortly after they discovered the theft?). I'm not sure if it was just to donors or also included newsletter subscribers. It basically said, 'we trusted this person and they took advantage of that trust.' It also states that since legal action is still pending, they have to "limit what [they] say", but that you can send further inquiries to Michael.

I see, thanks. I was thinking of "to" as just being the other direction of "by", but you can interpret it as more like "towards" and then it's all good.

Oh, I bet it's because the previous sentence was "When someone draws Mohammed, it is considered offensive to Muslims.", and that one seemed like a straightforward "no-one except Muslims is being offended by this" mapping, which was then extended to cover sexism.

When someone writes a story where all the sympathetic and interesting characters are male, it is considered offensive to women.

I'm sure you don't actually have any confusion here, but I feel compelled to point out that you kind of did that thing where you only expect a member of Minority X to be offended by *ism against Minority X, where in fact everyone should join in sharing the offense caused by it, because that's just part of being a decent person.

(I probably wouldn't have mentioned this but for the fact that we're having a meta-discussion about how offense works!)

0[anonymous]
Yvain didn't write "considered offensive by women". He wrote "to", which only very weakly implies the idea you criticize and can be charitably read as saying no such thing.
1Alicorn
Yvain didn't write "considered offensive by women". He wrote "to", which only very weakly implies the idea you criticize and can be charitably read as saying no such thing.

I work opposite that Cosi branch; it's rarely crowded when I'm there. I think it'll be fine.

Interesting article!

I think you're overstating the difficulty of Go to computers. The latest wave of Monte Carlo programs -- when run on fast multicore machines -- are able to beat professionals with a modicum of handicap on 19x19, or at even games on 9x9; they're certainly now better than the average club player.

3MatthewW
Seven stones is a large handicap. Perhaps they're better than the average club player in English-speaking countries, but I think the average Korean club player is stronger than Zen.

It would cost less to place someone in cryonic suspension than to execute him, and in so doing we would provide a chance, however small, that a wrongful conviction could be reversed in the future.

Hm, I don't think that works -- the extra cost is from the stronger degree of evidence and exhaustive appeals process required before the inmate is killed, right? If you want to suspend the inmate before those appeals then you've curtailed their right to put together a strong defence against being killed, and if you want to suspend the inmate after those appeals then you haven't actually saved any of that money.

.. or did I miss something?

4Morendil
Some of it is from more expensive incarceration, but you're right. This has one detailed breakdown: * Extra defense costs for capital cases in trial phase $13,180,385 * Extra payments to jurors $224,640 * Capital post-conviction costs $7,473,556 * Resentencing hearings $594,216 * Prison system $169,617 However, we're assuming that with cryonics as an option the entire process would stay the same. That needn't be the case.

That's true. I didn't spend my own money on them (I grew up in the UK), and they didn't cost very much in comparison, but I agree that it's a good example of a medical long shot.

1Morendil
Yep, the cost and especially the administrative hassles are, in comparison to the probability considerations, closer to the true reason I (for instance) am not signed up yet, in spite of seeing it as my best shot of insuring long life. To be fair, vaccination is also a long shot in terms of frequency, but definitely proven to work with close to certainty on any given patient. Cryonics is a long shot intrisically. But it might not be if more was invested in researching it, and more might be invested if cryonics was already used on a precautionary basis in situations where it would also save money (e.g. death row inmates and terminal patients) and risk nothing of significance (since no better outcome than death can be expected). In that sense it seems obviously rational to advocate cryonics as a method of assisted suicide, and only the "weirdness factor", religious-moralistic hangups and legislative inertia can explain the reluctance to adopt it more broadly.

Cryonics is comparable to CPR or other emergency medical care

.. at a probability of (for the sake of argument) one in a million.

Do I participate in other examples of medical care that might save my life with probability one in a million (even if they don't cost any money)? No, not that I can think of.

0Morendil
Did you ever get any vaccination shots? Some of these are for diseases that have become quite rare.

Ah, I see. So when you spend money on yourself, it's just to motivate yourself for more charitable labor. But when those weird cryonauts spend money on themselves, they're being selfish!

No, I'm arguing that it would be selfish for me to spend money on myself, if that money was on cryonics, where selfishness is defined as (a) spending an amount of money that could relieve a great amount of suffering, (b) on something that doesn't relate to retaining my ability to get a paycheck.

One weakness in this argument is that there could be a person who is so fearf... (read more)

3Blueberry
Cryonics is comparable to CPR or other emergency medical care, in that it gives you extra life after you might otherwise die. Of course it's selfish, in the sense that you're taking care of yourself first, to spend money on your medical care, but cryonics does relate to your ability to get a paycheck (after your revival). To be consistent, are you reducing your medical expenses in other ways?

It's not really an argument against those other things, although I do indeed try to avoid some luxuries, or to match the amount I spend on them with a donation to an effective aid organization.

What I think you've missed is that many of the items you mention are essential for me to continue having and being motivated in a job that pays me well -- well enough to make donations to aid organizations that accomplish far more than I could if I just took a plane to a place of extreme poverty and attempted to help using my own skills directly.

If there's a better way to help alleviate poverty than donating a percentage of my developed-world salary to effective charities every year, I haven't found it yet.

5knb
Ah, I see. So when you spend money on yourself, it's just to motivate yourself for more charitable labor. But when those weird cryonauts spend money on themselves, they're being selfish! How wonderful to be you.

Thanks for the reply.

One can expect to live a life at least 100-1000 times longer than those other poor people

.. when you say "can expect to", what do you mean? Do you mean "it is extremely likely that.."? That's the problem. If it was a sure deal, it would be logical to spend the money on it -- but in fact it's extremely uncertain, whereas the $50 being asked for by a group like Aravind Eye Hospital to directly fund a cataract operation is (close to) relieving significant suffering with a probability of 1.

Well, sure, but I asked how it could be moral, not how you can evade the question by deciding that you don't have any responsibilities to anyone.

0nazgulnarsil
what are morals? I have preferences. sometimes they coincide with other people's preferences and sometimes they conflict. when they conflict In socially unacceptable ways I seek ways to hide or downplay them.

Hi, I'm pretty new here too. I hope I'm not repeating an old argument, but suspect I am; feel free to answer with a pointer instead of a direct rebuttal.

I'm surprised that no-one's mentioned the cost of cryonics in relation to the reduction in net human suffering that could come from spending the money on poverty relief instead. For (say) USD $50k, I could save around 100 lives ($500/life is a current rough estimate at lifesaving aid for people in extreme poverty), or could dramatically increase the quality of life of 1000 people (for example, cataract o... (read more)

2Will_Newsome
One can expect to live a life at least 100-1000 times longer than those other poor people, or live a life that has at least 100-1000 times as much positive utility, as well as the points in the other comments. Although this argument is a decent one for some people, it's much more often the product of motivated cognition than carefully looking at the issues, so I did not include it in the post.
6knb
This is also an argument against going to movies, buying coffee, owning a car, or having a child. In fact, this is an argument against doing anything beyond living at the absolute minimum threshold of life, while donating the rest of your income to charity. How can you say it's moral to value your own comfort as being worth more than 100-1000 other humans? They just did worse at the birth lottery, right?
4nazgulnarsil
like this: I value my subjective experience more than even hundreds of thousands of other similar-but-not-me subjective experiences. additionally, your argument applies to generic goods you choose over saving people, not just cryonics.

Hi Hal. I'm sorry to hear of your diagnosis.

I spent two years as the maintainer of Dasher, and would be happy to answer questions on it. It's able to use any single analog muscle for control, as a worst case (and a two-axis precise device like a mouse as a best case). There's a video of using Dasher with one axis here -- breath control, as measured by diaphragm circumference:

http://www.inference.phy.cam.ac.uk/dasher/movies/BreathDasher.mpg

and there are videos using other muscles (head tracking, eye tracking) here:

http://www.inference.phy.cam.ac.uk/dashe... (read more)