Artificial intrusion?

Let's think for a minute about some things a good AI system would be able to do. I'm not talking about having "human emotions" or any junk like that, I mean being useful to real, practical human tasks. Take, for instance, an away message a friend of mine had posted: it said something like "my new email address is the same as my old but with indiana.edu at the end." A good AI helper background task would notice that, know that AIM contact corresponds to a certain person in my address book, and update the entry (probably without getting rid of the old one, but making the new one the default).

This kind of thing would be incredibly useful. (And, of course hard... and a fascinating problem to study :-) ) However, how did you feel reading that? Did you feel like it might be a bit invasive, to have a program running with that kind of knowledge about you and the real-world correlations in your electronic data (or at least enough equivalence classes in the electronic world), and access to make a change like that?

(An aside: in part I find this example interesting because it might not bother us for a human to do this - if you had a secretary doing things like that, it might not be so odd. Of course, the setting where secretaries work on things for you usually has stuff that you're not emotionally protective of the way you would be of personal information. The electronic agent doing this task might not be troubling in a business setting either.)

The problem with good artificial intelligence is not only that it's hard to do, but that we're really very bothered by it, even without it supposedly having feelings or intentions or anything like free will. I mean, heck, we're bothered by some algorithm at Google scanning our Gmail to automatically put in ads (see earlier commentary). It doesn't matter that the program can actually take no action which would cause us any conceivable harm; we anthropomorphize it unconsciously.

Maybe this will be a mentality shift of the next few generations (and I mean a couple of decades; not long enough for a few generations in straight succession). As it is, we're bothered my machines doing things for which we'd be bothered if a human did them or even some things we wouldn't. Maybe we'll come to realize the difference on a more visceral level, and stop caring if a bot reads our diaries. I think in part it depends on the direction technology goes - if you buy or subscribe to some kind of electronic journal which makes comments on your writings ("you like Jimmy? What do you see in him?"), that's really not going to help the case of AI not being invasive or weird.

Somebody on Google's side

Good to see at least Wired is siding with Google on the Print lawsuit:

http://www.wired.com/news/politics/0,1283,68939,00.html?tw=rss.POL

The line that gets me:

"Copyright is not designed to afford consumer protection or convenience but, rather, to protect the copyright holders' property interests," U.S. District Court Judge Jed S. Rakoff wrote at the time in UMG Recordings Inc. v. MP3.com.

Funny, wouldn't the content creators be better served if their content were better known and easier to find? What is the point of the copyright laws, if not to allow content creators to make the most profit they can from their works? Remember, even Alexander Pope said "Rules were made but to promote their End" (Essay on Criticism).

Immortality... sort of

"I don't want to achieve immortality through my work. I want to achieve it through not dying."
--Woody Allen

A bit more to the point... I'm slashdotted!

http://slashdot.org/articles/05/09/19/1116238.shtml?tid=126&tid=133&tid=99
(okay, okay, it doesn't talk about me directly, but I hung out with Marc for a good chunk of time while I was there and there are pics of me in the report)

This is up there with the Wired article about NerdCore:

http://www.wired.com/news/culture/0,1284,67970,00.html
The interesting thing about this one is that the MC Plus+ so prominently featured is Armand, whom I used to work with, and he wrote a song about Tomek, who I was working with at the time I read the article... just another guy I played volleyball and pool with (and even beat at 8-ball once), and then I found out he won TopCoder and ACM... such was life over the summer.

The Existence of Words

It shouldn't be too hard to believe language influences the way we think; the question is, how much? If you subscribe to the Whorfian hypotheses, what we can think about is entirely determined by what language we have available to us.

I don't go quite that far, but I definitely think it's important to study the tie between how we think and the language we have. In particular, I was thinking about why it is we have words for some things. I postulate (after a couple of linguistic classes and a fair amount of thinking) that we have a word for something when there is a need to refer to it frequently, and thus compactly. In this case, it worries me that we have certain words. "Anti-Semitism", for example (this was the one that got me thinking about the problem). "Rape", "torture", "murder", "theft", and many others come to mind. "Pirate", even (we do realize that they were bad people, despite the cool image now, right?) That we actually needed, in human history, to refer to these things early on (old Anglo-Saxon words, mostly, I think), and with such short words... well, that worries me.

I'm not a fan of just griping about things; I like to come up with something to do about them. But what can be done about something like this? I guess this is where the whole "politically correct" thing came from (although I think they were going for something a bit different). Do you try to get the words dropped from the language? "Forced sexual intercourse and destruction of property accompanied by seizure of property belonging to others" doesn't sound as cool as "rape and pillage", but I doubt that will catch on. Getting people to say "challenged" instead of "crippled" was hard enough, and in most cases, all it seemed to do was lend a bad association with a word that used to be good (this is what I mean when I say PC was going for something different). Maybe I'm trying to go the other way - let's stop using words with cool connotations for things that are, in real life, very bad, and not to be taken casually.

Addendum 8:43pm 9/18:
I think it's worth noting the line of thought that got me to the first word. I was thinking that there really ought to be more of some people in the world - like Ron. And Sergey. And Steve. And then I noticed that they're all Jewish. I guess I'm pro-semitic. What kind of messed up people wouldn't like those guys? What kind of messed up world comes up with a word that specifically means not liking those guys?

One more great Micro$oft quote....

Gotta love 'em:

To Ballmer's Chagrin, some of his up-and-coming programmers have left for Google. He was apoplectic about Lucovsky's departure, according to documents made public during the Lee trial. Lucovsky said in a sworn statement that after he told Ballmer about his plans to move to Google, the beefy CEO threw a chair and cursed Google's chief executive. "F__ing Eric Schmidt is a f__ing pussy. I'm going to f__ing bury that guy.... I'm going to f__ing kill Google," Ballmer said, according to Lucovsky. In a statement, Ballmer calls Lucovsky's account "a gross exaggeration of what actually took place."

(from Businessweek)

No! Not Eric! We like Eric!

And what, Ballmer? Did you only say F__ three times?

Hacking as a social problem

For a long time, I objected to the whole area of security because, as I saw it, it's something we wouldn't need if we could just cure the underlying social problems that cause people to commit crimes in general (a true xNFp statement). At some point in the past few months, I realized that it goes beyond that, that it is in the blood of a good coder to want to break things, and so no matter what the situation, security would always be needed, because, to a coder, breaking the system is another interesting, challenging puzzle to solve, whether they intend to do anything bad or not.

I was reminded of this just now when I came across this article on slashdot (scroll down to point #4).

There are a couple issues to separate out here: one, getting the system to do something it was not intended to do, and two, doing something harmful with the results of that bug. In a security talk I once heard, the speaker told a story about how he discovered that the prices for items on one company's web ordering system were stored in cookies on his local machine, so he changed the prices to a tiny fraction of what they had been, and submitted an order that would have gotten through successfully, if he hadn't called them to cancel and tell them their system really easy to hack. Had fun breaking the system, but used the knowledge to help the people running the system he broke. That's an action I don't have a problem with, and his response upon finding the bug I certainly can admire.

Noting that distinction, it seems that both my earlier observations have some part of truth: good coders will always want to see if they can make systems do things that weren't intended, but the intent to harm is a social problem that might be largely treated.

What about harmful results that come of accidentally making the system do something harmful, whether you were trying to break it or not? Well, maybe a lot of the things we consider harmful now are only so because of some social problems. Information release, for example. It doesn't matter if your credit card number gets out if no one's going to buy anything with it they shouldn't. It doesn't even matter much if someone does accidentally charge something to an account that isn't there, so long as they realize it and report it and fix it (only damage is a bit of time to correct the error). What if a bunch of data gets deleted? Well, ideally it should all be backed up well in real time, so that shouldn't matter either.

There are far more examples I should go into, but I think I'll save that for a post (or, more likely, series) about my idea of utopia and how it could actually work.

Academic disillusionment

It may, in the end, be a good thing that I didn't have an internship until this past summer. I know, for all I kicked myself for not getting experience earlier, not seeing a broader part of industry... I think this may well be one of those things that just happened to work out for the best in the big picture.

Why would I say this? Because, after one summer at Google, I'm painfully aware of how much the process and setup of academia bothers me. I had always been a bit bothered, but now I have concrete experience and observation to contrast it against.

I'd been saying for a while that I don't want to be in research because I want to do something which has immediate, positive impact on the world. I want to do something real. School is the opposite of real. Everything in it is designed to be artificial. Not, note, without purpose - but deliberately removed from the messiness of the real world. Each assignment is crafted to be a neat little tidbit capturing and exercising the understanding of an idea. Each exercise from a section of a book is solved using ideas from that section. This is a good way to understand certain concepts, but really all you're doing is solving a heavily restricted problem. Even for the "tricky" word problems, you already know something about the solution, since it will involve some techniques from the section. Comprehensive exams do a bit better, since you have to draw from the material of all or most of a course, but it's still restricted to just that course, and you're still solving "neat", artificial problems. This is hardly "problem solving".

On top of this, there is the demeaning status of a student. You're the bottom of the food change. You can't argue with anything. It's one thing to be in this status when you're a child (and I have enough complaints about that; must post about those some time), but for someone 18, 21, 24... it's really frustrating. The instructor and his/her whims are entirely responsible for a mark that will be on your record for the rest of your life, and will affect what you can do for at least the next few years. Now, you have many instructors over time, but each one of them has this power. The system is rigid and beaurocratic. A few nice administrators or instructors might bend the rules, but on the whole, it's set for the masses, and there's very little you can do to escape jumping through many pointless hoops.

Now, it's not to say that I don't think many companies have many problems (including those of schools), but I think there could be much better ways of running the system. What we need is a holistic approach to taking individual from childhood to a happy, productive member of society. When I say holistic, I don't mean some weird hippy thing; I'm being entirely practical. Let's take it from the perspective of greatest total, long-term economic benefit, maybe something like lifetime per-capita GDP.

I have heard from multiple sources in multiple fields that the optimal degree for money is a master's. Let's assume this takes until age 24, and that you work until age 70 (probably more reasonable for my generation, and a lot of people take up second careers after their first retirement anyway). That's 46 years of productivity for 24 years of training. Let's also assume that under a certain age, say, 12, you're not going to do anything really productive (most of the tasks a 12 year old could do can be automated, or should be soon), so it's really only 12 years difference. Now, if the 12 years of training makes you at least 26% (12/41) more productive, it's good to spend those 12 years training.

That sounds fine, but why can't the training be productive, and in the process, make one even better prepared for the rest of one's life? From my observations of the early school system, most of the work is repetitive and the students don't care about it. Now, this alone is a problem that could be remedied without affecting the rest of the system, but I'm going for the holistic best solution. My suggestion for this is apprenticeship. Maybe this isn't practical at early ages (although I still think more individualized work should be done), but starting around late middle school/early high school, I think it would be more beneficial to set each student up with a mentor to apprentice with, for as long as each arrangement seems to work (no sense artificially dividing the time into semesters or years if it's already individualized). This will not only allow the apprentice to do something productive in the real world, but see the applicability of the things s/he is learning.

In terms of overall productivity, though, one must consider that the mentor is losing time too: time spent training the apprentice that could be used on their own pursuits. Maybe this is an hour per day, per student - I don't think it would take much more, since part of the point of this is also self-reliance and self-direction, and small-group instruction could be held with several apprentices at once. This is weighed against the professional teachers we have now, who may have 20 students each (or 1/6 of each of 120 students; each has 6 periods a day), and doesn't do anything but teach them. This means, say 1/8 of a mentor's time used vs only 1/20 for a traditional teacher. Does the increased productivity of the apprentice, including both the apprenticeship and over their lifetime, make up for the 3/40ths of an adult professional's time? I'd say yes, but then I'm a fan of this system for other reasons as well. If a 20-year-old working 40 hours can't make up for 3 hours out of someone else's week, even if they are spending time on things like book exercises, I'd be really surprised.

Finally, I'd like to argue for the long-term benefit of this over traditional school, in a slightly less economic sense (although it still has economic implications). When a student goes from a purely school environment to the professional world (as I did this past summer), they're unused to anything other than the idea of getting closed-form assignments and doing them individually (maybe in a group, but this is less common, and usually considered cheating if not assigned). This is a great contrast to open-form assignments ("make this better"), and having social networking being a big component of getting things done. If your assigment is "make this better", and you figure you might need a particular algorithm implemented to do this, it's all well and nice to look through text documentation and explanation, but it's much better in many respects to just go ask somebody who has knowledge of that. For one, they might have your needed functionality already implemented in a library. For another, they can point you to other related things that might be useful, and even better, to other people who have ideas about the subject.

My ideas of an apprenticeship system obviously has potential drawbacks which need to be addressed. For one, it requires a lot of mentors who are reasonably good teachers. Otherwise, we get another problem that the public school system has right now, measuring by standards which are not necessarily reflective of quality - class size, in this case. Sure, this system has a class size of probably a handful, but most of us would agree that it's better to have a class with 40 students with a really good instructor than a class of 20 with a poor instructor (hence another idea I've heard, to fire the bottom half of teachers and pay the remaining ones twice as much - which I think would result in an improvement, since it would draw even better teachers).

Maybe the best approach would be some kind of hybrid. There are some pursuites where one does just need to cover a whole lot of specialized material, basically in bulk. Under a strictly apprenticeship program, it would be hard to get people like math researchers. I imagive you'd have some studentents who'd apprentice with, say, an engineer, who'd wind up really liking the math part, and devoting big chunks of time to learning it mostly on their own, and after a while, switch to being the apprentice of a math researcher. Well, maybe it wouldn't be so hard then. I appreciate any comments for problems of this type and suggestions of how they be investigated or treated.

As a side note, think about what it means that I complained about the particular things that I did... these are things I had as contrast against working at Google.

Oh, and if you're bothered by my occasional use of 'they' as a gender-neutral singular... it's intentional. It was accepted in the time of Shakespeare (who did use it), and there is linguistic backing for its use as such. So get off your prescriptive grammar and look at the purpose and functionality of the construct.

A place for miscellaneous thoughts (and some quotes reflecting company culture)

I wanted to separate my blog into events and thoughts. My Mountain View blog will be one for a series of events, organized by chronology and location. I may make another about some other contiguous bit of time (maybe about Burning Man, but I doubt I'll post much of that publicly). This will be just for my thoughts about life and the world.

My random thought for the moment was spurred by a few quotes I came across:

I have 100 billion dollars... You realize I could spend 3 million dollars a day, every day, for the next 100 years? And that's if I don't make another dime. Tell you what-I'll buy your right arm for a million dollars. I give you a million bucks, and I get to sever your arm right here.
--Bill Gates

You see, antiquated ideas of kindness and generosity are simply bugs that must be programmed out of our world. And these cold, unfeeling machines will show us the way.
--Bill Gates

Obviously everyone wants to be successful, but I want to be looked back on as being very innovative, very trusted and ethical and ultimately making a big difference in the world.
--Sergey Brin

(all from http://www.brainyquote.com/)

Now, maybe Mr. Bill isn't serious, and these are probably extremes, but... does anybody wonder why we're rooting for Google? And why I'm so in love with the company? And why Sergey is my hero?