Why I want to go to Law School

This is a quick-and-dirty mind dump of why I am considering attending law school next year (and will serve as a boilerplate for the inevitable personal essay to write on every single law school application ever) As many of you who read this know, I grew up being a 'computer guy' and an all-out geek. I wanted to solve problems that had the one correct answer. Math and logic puzzles seemed like the perfect encapsulation of a problem: While the problem may be convoluted and in slightly verbose language with complicated inter-dependencies between constraints given in the puzzle, after enough raw thought and avoiding logical fallacies, the correct answer was demonstrably generated. One could verify the correctness of the steps taken to reach the conclusion, the correctness of the input, and the correctness of the output. It was neat, tidy, and pure.

Computer Science has provided many a problem with a definite, provable answer, such as the lower asymptotic bound for a comparison-based sort. Proving such a fact is more difficult than solving a sudoku problem, but it has been done. However, while reaching into the upper levels of undergraduate Computer Science: automata theory, metalogic, and foundations of mathematics, it became clear that on higher levels, the logical systems I depended on so much were based on interpretations of symbol systems artificially constructed on top of a hypothetical set theory. Definitely not the neat and pure system I had so naively thought them to be. It definitely was a case of "the more you know, you more you know you don't know."

At the same time, I was lucky enough to be accepted as an IT intern (and then staffer) for Senator Reid's office in Washington, DC. While my role there was entirely of a technical bent, I was able to witness firsthand the almost-antithesis of computable problems: policy making. It was a sudden shock to me to watch debate about important issues, watch two speeches on the same topic from opposite viewpoints, and still both be rationally sound. I appreciated the spirited debate, save for one instance: when basic facts were misrepresented. I believe that basic facts: observable evidence, rigorously-obtained statistical information, and the like are the cornerstone for progress. I believe that the best debate (and the most well-thought out resolutions to that debate) comes from 'true' information and sound inference rules. When arguments are based off of false data or specious logic, they tend to lead to false results. While argumentative styles vary far and wide within politics and debate about non-computable issues in general, I respect arguments that follow a consistent, logical flow.

I suppose I see law school as a way to apply logical argumentation to decide on issues outside the normal operating procedures of engineering disciplines. It would help me bring my expertise in the technology field to a legal system so sorely in need of technologically-apt disciples.

I still haven't decided if this is the best path for me. Perhaps I place law school on too high a pedestal. Perhaps there is a better way to effect change. It is an option I am considering, however.

What your computer does while you wait (and how to make it go faster)

http://duartes.org/gustavo/blog/post/what-your-computer-does-while-you-wait For the many people who wonder why the computer takes so long to respond even while it's not 'doing anything' (myself included). Of particular interest, while your computer may operate mainly within close caches and system memory, reading data or code from the hard drive can take up to approximately 164,000 times as long as reading from main memory, and about 13.6 million times as long as reading from the fastest cache available to processors. Of course, this is the initial seek time, and is much slower than what are called 'sustained reads', where a large block of data is read at the same time. Think of it as travelling across the world to read something out of a book. Reading the first word takes 12 hours, but reading the next word is much faster.

Much research has gone into reducing hard drive reads, as many software architecture and operating systems professors/books can attest, and the reality of the situation is much better than the bolded numbers might suggest. Compact, well-written applications coupled with large and nimble CPU caches and main RAM can reduce the number of hard drive reads and writes to a minimum, but it still happens. Those hard drive hits crush the performance of any application, and keeping data in line and ready to be read can really improve performance.

This performance hit is felt the most during system boot, when nothing is cached in memory. Your computer has to load the basic core of Windows/Linux/Mac, followed by drivers, services, and other essential processes that allow you to use the computer. All of these require many seeks to the hard drive to get the files into memory, then into the CPU, and it hurts. The same effect is exhibited when computers resume from sleep or hibernate, where some or all the memory has been written to the hard drive and must be pulled back into memory at the comparatively excruciatingly slow pace that hard drives run at. So when you unfold your laptop, the computer is unresponsive for a few seconds as all the data is read back into memory while processes are trying to restart where they left off.

To combat that evil seek time, some computers (such as the MacBook line) have Solid State Drive options. These work just like RAM (although quite a bit slower, but not near as bad as hard drives), in that there is very little seek time! So when your system boots, all the files it needs can be loaded quickly without seeking all over the hard drive, easily speeding up system performance!

Sadly, solid state drives are hard to come by for desktops, and are normally just laptop drives in desktop casings, so my desktop will sit with standard seeking hard drives for the foreseeable future.

The Genius in the Cloud

So Apple released a bunch of new stuff today (read most of it at http://apple.com/hotnews), but of most interest to me is iTunes 8, and more specifically, the Genius mode. The Genius is supposed to investigate the information in your library (playlists, track played counts, ratings, etc), upload it to the 'Genius in the iTunes Store' and receive some sort of information back. This information is used to match up songs you have that are supposed to go well together. For example, my laptop has a (somewhat) small library consisting of some speedy rock, some techno, some classic rock, some metal, a bit of Apocalyptica (a band that plays metal cellos), and various other songs. When I play one of the heavier songs and ask the Genius for recommendations, I get more metal, and some of the heavier versions of my speedy rock (and the heaviest songs from Apocalyptica). It's actually pretty impressive.

This idea is also used by radio services like Last.fm and Pandora to match like songs together. Last.fm has always been pretty good at recommending music I like, and while I haven't used Pandora enough to give it a fair shake, the people who use it swear by the accuracy of it. Upon watching the keynote speech, however, I realized just how powerful this iTunes Genius could be. With at least 63 million users/libraries, Apple has tapped into the largest collection of music ever amassed. Jobs made multiple references to the Genius getting smarter as time goes on, and as more and more people add their collective music information to the database, it will get better at recommending music.

It depends on how Apple has implemented their Genius algorithms, but this is probably one of the larger Artificial Intelligences with gobs and gobs of data to pick through. If AI's are getting this good, it's going to be an interesting future. Google's algorithms consistently find the pick of the Internet litter, and if iTunes Genius is going to do the same with music, it stands to reason that other forms of information could be as easily integrated, categorized, and processed to find the most relevant information for any one situation. This high level of semantic awareness is quite the stuff of Science Fiction.

Since the iTunes Genius is primarily built to sell songs on the store (which it does with aplomb), one doubts it will have the stereotypical self-awareness moment and strike against humanity. in 20 years from now though... it's definitely going to be interesting