Let’s ignore the whole life-after-death thing, any idea of continuity of existence. Even then, it could be convenient if there was some real but ineffable thing which humans could not produce other than through reproduction.

Why?

If cognition is a physical process occurring entirely in reality, and the brain is through some transformation Turing-complete, humans will eventually develop true general AI. They have to avoid extinction in the meantime, but it is not an area of research that will ever stop being of interest. Given that I assume those two conditions are true, it follows that I assume that eventually AI will be a practical reality.

When that happens, there will be people who recoil from the entire field of programming. Once we’ve written sentient programs or otherwise created sentient machines, the notion of writing any program (even a fully deterministic one) becomes just a little bit icky: it’s like enslaving the idiot-savants. I side with the programmers: it’s not really like that at all. That won’t prevent the arguments from being an enormous headache.

I bet it’ll be such a huge headache, in fact, that it’s almost reasonable to wish that there was such a thing as a soul.

What is the correct ethical option here?

I see several possibilities. You could turn over the hard drive to DARPA so they could keep running fully intelligent warbots, though these warbots are as intelligent as an average human and hate dying just as much. You could hide the source but run it yourself to get something like your friend the AI back. You could open-source the project and give it to the world. You could move out of the US, then attempt to sell the project to the highest bidder, on the assumption that full AI would be worth millions or more. You could tinker with the source to make it a little less smart, then give it back to DARPA. You could mortgage your house to buy a nice chassis for it, then set it free to make its way in the world (and hopefully pay you back for your house in time). You could destroy the source in the assumption that any human encounter with an AI will be an attempt by the human to enslave the AI, and annihilation is better than slavery.

I would probably open-source the thing. Upload it to sourceforge, put one link on digg and another on slashdot, and trust that enough people would download it before DARPA caught on that it’d be impossible to put the genie back into the bottle. DARPA would probably be annoyed, but I think I could gather enough public opinion on my side to prevent any real consequences from harming me. Most of the other options seem defensible if not optimal in my mind, except the last. That one seems both unduly pessimistic and short-sighted.

Naturally, the last one was the one chosen in the novel which prompted this post.

What if it had truly new features? Say a conversational AI which you could talk to using natural speech, to which it would reply with speech as natural? The sort of thing which would pass a traditional Turing test. Would you play it then?

There is a threshold at which people decide whether or not to play a game. I suspect that it is related to the product of features and prettiness. At what level of features would you accept a game which isn’t graphically pretty at all?

Your Personal Statement should be a brief but carefully written essay regarding: 1) your reasons you want to do graduate work in this particular field, 2) your specific interests and experiences in this field, 3) any special skill or experiences that may relate to an assistantship, and 4) your career plans.
—–

One of the most important lessons that I’ve learned during my time as an undergraduate student is that computer science is a continually evolving field. Specific languages and procedures that I use now might be rendered useless by new developments in theory or technique at any moment. Keeping abreast of current developments is a requirement just to survive in the job market.

However, I’d rather not just tread water, just keep afloat. I’d rather be part of the group pushing the envelope of what can be done with computers; I’d rather be writing the papers than reading them. Undergraduate studies are a step toward a career coding; they bring one up to speed with curent practice. Graduate studies are the next leap forward; they transform my role from that of a learner to that of an innovator.

WPI’s undergraduate computer science curriculum is project-based. Among other things, this means that before graduation, I complete two projects for class credit: the IQP is worth one class of credit, and the MQP is worth three. The former involved the development of a tool to aid people unfamiliar with computers with the process of setting up a complex database-backed web system. The latter, currently in progress, seeks to extend the Napkin Pluggable Look and Feel (napkinlaf.sourceforge.net) by creating a system which generates icons which look as if they had been hand-drawn, as the icons are called for.

The only common elements between those projects is that they both represented a significant amount of research, of delving into programming territory unfamiliar to me. I learned enough from the first project to qualify as a PHP/MySQL expert at the job I got subsequently; I expect to emerge from the MQP an authority on both Java PLAF technology and the specific graphical tricks involved in tricking people into thinking that a human drew something machine-generated. Some of my most interesting classes have been in AI and pattern recognition. I enjoy extending myself, branching into new areas and fields.

I don’t yet know what I will do after graduate school. I certainly wouldn’t mind doing pure research; holding a job learning things so new I’m the first person to get there could certainly be interesting. Alternately, I’ve had the idea for some time now that it might be fun developing intelligent opponents for computer games. For the moment, I intend to keep my options open.

One of the major reasons I look forward to graduate school is that I often read an article from one news source or another about a really interesting project being carried out by a team of graduate student researchers. Be it developing a control system for DARPA’s unmanned vehicle race, or a system which can beat most people at 20 questions, there are all sorts of interesting developments being made in the computing world today. Linus Torvalds was a graduate student when he wrote Linux. I want to make my mark too.

Any suggestions or comments are appreciated–I have a month before this is due…

My project partner contributed three or four pages. The rest was my own. I am kind of put out about this…

As long as I pass the course, I suppose it’s ok. Nevertheless, I think that choosing a good project partner with whom I work well is an essential component of classes like this. I’ve had tough, project-based courses in the past, and come out with a lot less stress, and an ‘A’. Notably, AI. I don’t think I’m going to do nearly as well in this course…

Of course, there were minor little adventures on the way, like prying the drive I wanted to transfer out of its housing with pliers because a screw holding it in place was stripped and wouldn’t come out, and the frantic race to install drivers that the Windows reinstall had deleted before the lack of those drivers caused a system crash… but it seems to be stable now. At the very minimum, it boots faster than the old machine used to.

So, now that this is done, after dinner I get to go work on AI for another several hours.

Over the last three days, I’ve spent (with my project partner) a total of about 55 man-hours working on the AI project. Last night I was dreaming about code to implement decision trees. Finally, though, we have just over 40k of code which seems to mostly work. There’s still some debugging to do, as one of the routines generates an infinite loop (one which checks for loops in the node structure, in fact), but it’s close enough to being done that I’m not nearly as worried about it as I was before. Add this to the fact that the professor just gave everyone a day’s extension, and I feel like I can spare the time to work on my other project.

My other project is to set up the new computer. I picked it up yesterday, and started working on it today. Unfortunately, I have to do a bit more than just string an ethernet cable between the two machines and copy the files I want; there are a few things which just didn’t come with the new machine, like a 3.5″ drive. The plan I’m going to try is to stuff everything I want to keep onto the second hard drive, and migrate that entirely onto the new machine.

After I get the hardware set up, my understanding is that I’ll have to reinstall XP (which came with the new machine, and doesn’t let you easily switch to, say, 2K) because the hardware profile will have changed significantly. Luckily, the school has a site license for XP Pro, so I have access to an access code and a less privacy-invading version of the software. Then, I’m going to be doing a bunch of simple configuration, installing programs that I like to use, etc.

At any rate, it’s been a while since I’ve poked about within hardware, so it’s going to be interesting fixing all this stuff up. Mostly, though, it’s just very nice to have a working machine.

On the positive side, I hope to be able to pick up my computer today. Two different people have indicated that they would be willing to give me the ride necessary, and as long as they don’t both renege, I’ll be in a happy place by some time this evening.

It’s been snowing now for a good 40 hours, and the campus is beautiful. I’ll take the beauty of the snowdrifts and slides over the convenience of having cleared paths to walk any day. Unless I really need to get somewhere in a hurry. But that doesn’t really happen that often, and by tomorrow morning I’m sure every path on campus will be cleared meticulously, because the staff at this school are really quite competent.

I spent the major part of yesterday working on an AI project. The game-playing program was fun, but relatively simple. Now, we’re working on constructing efficient plans to move from an initial condition to a goal condition, and even though the algorithm is really easy to understand, it just takes a lot of code to implement. I’m going to continue working on it today after I get back from picking up the computer. After that’s done, I get to work on the Databases project.

Basically what I’m saying is that not a lot is exciting right now.

I have an AI project due tomorrow. It’s pretty simple: we have to produce something that plays tic-tac-toe in three dimensions on a board four spaces across. So my project partner and I started working on it this afternoon.

We took maybe two hours to do the algorithms, the data representation, all the tough parts. Test programs showed that it was working fine.

Then we spent four hours making the IO work, because the thing has to talk to a referee and for whatever reason just didn’t want to parse the data that was being passed to it about the state of the board.

We are using C++ for efficiency, because there are a lot of crazy optimizations that work in that language that you can’t do in many higher-level ones. We need efficiency because there is a tournament, and each player gets only one minute of realtime to make their move. It’s a really good language when you need to do efficient fast manipulation of strange data structures, deep recursive functions, experimentation and problem solving. It’s a really bad language when you just want to parse input sent to you over standard input.

Our player works now, and has been demonstrated to win a game played against itself. Tomorrow’s work will involve implementing an idea I had which, if it works, should make it have snap decisions about trees which are obviously good or bad, and explore ambiguous ones much more deeply. Also we will be working on implementing a timer system in a separate thread so that it can go explore trees without having to worry too much about losing for going overtime. I just wish that we hadn’t been required to spend as much time as we did on the IO, so that we could have worked on some of that tonight.

I’ll figure out a post with some content later, after I’ve finished all the current project/crises. And probably a while after that, because apparantly this weekend is anime weekend and the SFS is showing a bunch of stuff on a big screen for 72 hours continuously.