Category: Geek stuff

If it involves physics, math, computer science, electronics or has general geek cachet, this is the place for it.

iTunes has lost its mind

Not only is Apple’s iTunes application constantly crashing, stealing focus, or crashing my entire computer – it just erased all my podcasts and iTunes U files for no clear reason. This is especially maddening for someone who is careful to catalog their media well.

Given that I was about to go on a long bus journey, this is unbelievably annoying. All of a sudden, university courses I was in the middle of listening to are just gone. Podcast collections that I had assembled with some care are deleted. All my audiobooks have vanished as though they never were. Much of the content I had in there cannot simply be re-downloaded.

It is all extremely frustrating.

Rating the raters

Nate Silver of FiveThirtyEight has an interesting post about sovereign public debt default ratings from Standard & Poor’s. He argues that their ratings from five years ago did not reflect the risks that arose with the financial crisis.

Despite their potentially misleading character, the financial system relies on such ratings to be a proxy for probability of default. A bondholder with a good rating should be less likely to default than one with a poor rating, and a highly rated security should be a safe investment, If the ratings produced by rating agencies are not a good proxy for risk, it may be a mistake to continue to give them such an important role within the financial system and financial regulation.

Of course, that raises the question of what to use as a superior indicator for risk.

Fischer random chains

Fischer random chess seems to have much to recommend it. It removes the advantage of memorizing large numbers of opening book moves, and arguably injects more creativity and dynamism into the game.

One major problem is that it heightens the asymmetry in chess between the player who moves first and the player who moves second. Some randomly generated positions will strongly favour white.

An easy way to counterbalance that is to play each random setup twice, with the players alternating who plays as white. The problem with this is that it reverses the original advantage by too much. The first game will provide all sorts of chances to think about the unique characteristics of the random opening position. The person who plays white the second time has the advantage of all that information.

To be more fair, then, perhaps players should play four games based on two positions:

  • Game 1: Random position one, player A as white
  • Game 2: Random position one, player B as white
  • Game 3: Random position two, player B as white
  • Game 4: Random position two, player A as white

Even with a ‘chain’ of games like this, the balance between the players still isn’t perfectly fair. One of the starting positions might offer more of an advantage to white than the other one does. For instance, position one might randomly be the ordinary chess setup, while position two might be one that allows white to gain a strong early advantage.

The cost associated with playing Fischer Random / Chess 960 is that it allows some degree of luck to enter into chess – a game that it otherwise devoid of any element of random chance. That being acknowledged, it seems like the amount of luck introduced can be minimized through pairings like the ones above, and that the remaining fraction could be worth accepting as the cost of achieving the other benefits of Fischer Random chess.

I would be up for trying a four-game mini-tournament with someone, perhaps with 15 minute per side sudden death time controls.

Why I left Facebook

I have been worried about Facebook for years. I worry about how personal information on users is their most valuable asset, and the ways in which they may seek to profit from it. More generally, I worry about the unintended consequences of creating massive searchable databases on social interactions.

What actually prompted me to ‘deactivate’ (not ‘delete’ yet) was two things.

Excessive time demands

First, Facebook is too time-demanding. People expect me to keep up to speed on their many postings, despite how there are hundreds or even thousands of status updates that appear every day. If you advertise your event on Facebook and I miss it completely, it is probably because I was trying to get some reading done, or enjoying a walk and a cup of coffee, or dealing with my neverending flood of unanswered email and so I missed the status update message or invitation for a few days.

If you really want me to know about something, you must at least send me a text or an email. Putting notice on Facebook (or Twitter, or your own website) is not a sufficiently attention-grabbing action to ensure that I will see it.

As I am writing this, I am ignoring a sizeable collection of projects that are in need of attention. I should be working on finding an apartment in Toronto, packing up my current apartment, and making plans for how to move. I should be researching possible doctoral programs, working on my research proposal, and corresponding with possible references and supervisors. I should also be reading various books from various stacks of semi-read tomes, refining my low carbon mutual fund idea, improving my chess, getting exercise, exploring some elements of Ottawa that are still unknown to me, planning a trip to Washington D.C., planning a trip to New Orleans, writing articles and letters to editors, processing and uploading photos, going out and taking new photos. I should be taking university courses, learning practical skills, responding to letters, searching for photographic gigs, learning to drive, joining clubs, going camping, and improving my data backup regime.

All of those tasks are better uses of time than Facebook.

Privacy

Second, I am worried about facial recognition. The only barrier to it becoming absolutely ubiquitous seems to be the availability of data on our faces. The cameras are already out there, and the software and the computing power to turn pixels representing faces into names are coming inevitably.

Someone with a lot of determination can dig around the internet and probably find dozens of photos of me to feed into a facial recognition algorithm. While I was on Facebook, however, this process was simplified to the point of easy automation. In thousands of photos, I had been specifically identified and even had the region of the photo containing my face marked.

Still not too isolated

So far, I have been glad to be off that particular grid. Anyone who actually wants to contact me has a wide variety of ways to do so. My email address and cellphone number are both on the ‘contact me’ page of my blog, and my blog comes up immediately when you Google my name. If that is too much work for a person to go through, it seems fair to say that they didn’t really want to contact me in the first place.

I don’t want to delete my Facebook account completely because it does have some value to me as an archive. Nearly all my photos from Oxford are in there, with tags and comments affixed. If Facebook provided a way to download all that as an elegant, accessible archive that can be used offline I would be happy to do so. I doubt, however, that they will ever provide such a tool. All their plans hinge on attracting people to the site and making them visit as often as possible. Helping them untangle themselves and walk away with whatever data they find valuable runs completely counter to that. Facebook actually lets you download your photos easily in quite a good archive format.

I will miss the chance to see what distant friends are up to easily, and to have the occasional fortuitous bit of contact with them. I will try to remember to send an out-of-the-blue email every once in a while.

P.S. I left LinkedIn too, but who cares about LinkedIn?

This American Life on patents

A recent episode of the This American Life podcast centres around technology patents, with emphasis on the so-called ‘patent trolls’ who harass legitimate companies using dubious patent claims, in hopes of getting cash settlements.

Designing an ideal patent system is an interesting question from a utilitarian perspective. It seems beneficial to encourage innovation and protect small companies with novel ideas from giant companies that might steal them. At the same time, patents can be used by big companies to bully small ones, and when obvious ideas are given patents it can prevent useful technologies from becoming widely available.

Automated facial recognition

As processing power becomes cheaper and smarter software is produced, it seems inevitable that more and more people and organizations will begin to identify people automatically by recognizing their faces with surveillance cameras.

London’s Heathrow airport is planning to install such a system, and Facebook may be the ultimate database to let freelancers do it themselves.

To me, it is all rather worrisome. At a basic level, life becomes more paranoid and less creative and interesting when you are being watched at all times and all of your actions are being archived forever. It’s only a matter of time before photos from every fun party ever are being combed through by investigative journalists hoping to catch someone who has become famous in an embarrassing-looking situation. Facial recognition allows for the creation of databases that can be used for truly evil purposes, from suppression of political dissent to stalking and blackmail.

Like nerve gas, facial recognition technology is probably one of those things that it would be better if we could un-invent.