How Algorithms Shape Our World, Fascinating TED Talk by Kevin Slavin

[Video Link] Via Laughing Squid, this TED talk by game developer Kevin Slavin:

Kevin Slavin argues that we’re living in a world designed for — and increasingly controlled by — algorithms. In this riveting talk from TEDGlobal, he shows how these complex computer programs determine: espionage tactics, stock prices, movie scripts, and architecture. And he warns that we are writing code we can’t understand, with implications we can’t control.

How Algorithms Shape Our World, Fascinating TED Talk by Kevin Slavin.


  1. It is so true… I work with various decision-making algorithms in my research.  The complexity is such that I can’t fully account for all possible behaviors of a single algorithm.  When you have these things making decisions based on the decisions of countless similar algorithms, the results are way beyond comprehension, much less prediction. 

  2. That’s got to be one of the most fascinating TED talks yet. We’ve always imagined the rise of the machines to resemble ourselves, or at least to be embodied, but really they’re algorithms designed to compete against each other (and us), and right now we (mankind) are installing them into the hearts of civilization in places like New York city.  

  3. Well, we’ve been living in an environment of algorithms we don’t understand and can’t control for hundreds of thousands of years. I guess the danger is if we assume we understand/control these new ones we’re writing and as a result expose ourselves to serious risk.

    Anyone with a computer can probably see that we don’t really understand or control them though :)

  4. In my profession creating algorithms the programmer doesn’t understand isn’t a side-effect, it’s the objective. Even without GAs, though, control theory of complex systems has become a whole range of specialties in its own right. For a speculative fictive take on the matter, I recommend Daniel Suarez’s excellent duology Daemon and Freedom TM.

  5. Fascinating. I remember watching the FC of 2:45 in real time. What a wtf moment. Great stuff, BB.

  6. My current research involves another system that we created but don’t really understand: the Internet. We assume that it’s stable and resilient, but numerous phenomena resulting from the protocols and implementations make it much less so.

  7. I found the part about trade algorithms to be somewhat uniformed scare mongering.  Working in this industry, I can tell you that the trading algorithms can be unpredictable but the folks trading them have strong financial incentives to put tight stops in so as not to incur huge losses if they start to go wrong.  What is more interesting is the idea of algorithms driving social behavior. We see some of this with “the filter bubble”; algorithms guiding our search results, what music or TV is recommended to us and so forth.  The idea that algorithms will increasingly drive our perception of the world is correct, I think.  Already it is fairly common to see folks walking around with their cell phones or music out of touch with what’s going on around them.  Machines will likely continue to insert themselves to provide narrative, context, opinions on everything we see moving us away from directly experiencing our world.

    1. Even if the people overseeing these systems have “strong financial incentives”, well — so have all other traders throughout history; yet those same traders do stupid things like play chicken, follow tactics and plans that don’t take into account all the forces and variables, take lunch breaks, fail to comprehend what’s happening, and design in “features” that let trading continue because the failsafes are too good and stifle unsafe but lucrative trades. Every trader that’s lost millions or billions or squillions of dollars has had very strong incentives, financial and otherwise. Their behaviour is motivated by personal greed, inside knowledge, and a failure to give a damn about anything else.

      1. It’s easy to see high frequency trading algorithms creating short term price spikes or exaggerating existing market trends. It’s more difficult to see moves created by these algorithms persisting over a longer time frame since at that point human traders and judgement comes into play.

      1. I tried it with DownloadHelper (plugin for Firefox); it doesn’t do the trick, although it works on a lot of other sites.  Do you have any suggestions?

  8. Whoa, this is the guy who turned me on to Black Flag in high school. I realize this comment is of no interest to anyone other than myself but that’s the internet for ya.

  9. I’m a software engineer and I find the stock market part of this disgusting. It’s trading wildly based on crazy algorithms and the speed of the network – It’s based on virtually nothing. The stock market shouldn’t be allowed to work this way – it should be based on tangible goods and services.

    1. In many ways I agree with your sentiment. At some time (around the ’70’s) trade started to shift from production and steady/stable growth towards speculation. The ideal here being that instead of real economic good being used/produced/traded/etc. the system started to resemble gambling. So the numbers looked more like 70% spec. 30% real trade where before it was basically the opposite. Of course the problem with this is that while more and more money is being used for speculation there is less and less money available for real investment aimed at long term, stable growth. What people (rich people) wanted was quick gains with little work. What a stable economy(and incidentally the majority of normal/working class people) needs is basically just the opposite (good planning, hard work and steady growth). So far the only thing I have come across that seems a decent solution to this growing mess is what is called the ‘Tobin Tax’ which is a tiny percentage applied to all speculative trade on all global markets. In a nutshell, this tax would drive down the interest in speculative trade and also help offset some of the financial problems associated with this gambling, er, I mean trade.

      More OT, very interesting talk. Makes me wonder about the wisdom of blind (semi-blind) algorithms in the way that cyberbadger means but it also makes me wonder about the sort of things we have yet to learn about ourselves and maybe (just maybe) our machines can teach us. On a more pragmatic note, the destruction of, or rearrangement of, the natural environment to accommodate the facilitation of faster transmission for faster ‘fingers’ on the buttons disturbs me greatly. The main example of financial transactions only makes this sort of endeavour seem more trivial and useless. And therefore more worrisome.

      FWIW I love me some Black Flag so there is that too. Cheers.

Comments are closed.