nice, nice, very nice

so many people in the same device ...

    • Edit
    • Delete
    • Tags
    • Autopost

    unanswered letters (#2)

    Note: I've written a number of letters/emails over the years many of which have gone unanswered leaving me to wonder whether or not the intended recipients ever received them and, if so, what their responses were. To be honest, the vast majority of those are best forgotten, but a few are ones I sincerely hoped would succeed in being actual communications — the beginnings of conversations. The following is one of those.


    While the following was addressed to the NY Times Editor (and, due to its length, stood no actual chance of being published) it was also cc'd to Google where I'd hoped it might be appreciated.

    June 21, 2004.

    Dear Editor,

    I cannot decide whether the title for Saul Hansell's article about Google's GMail service ("The Internet Ad You Are About to See Has Already Read Your E-Mail" - June 21, 2004) is subtly brilliant or just very badly worded. How can an advertisement read anything? Isn't it Google that's reading "your e-mail?" Well, no, that can't be right either. How can a "Google" read anything? So, maybe it's Google's software that's doing the reading? That sounds a little more plausible and one might be excused for thinking that software, with all the information processing and artificial intelligence the word would seem to imply, is certainly able to read. And yet, in truth, that is also certainly not right.

    In "A Mathematical Theory of Communication," Claude Shannon's seminal paper that launched us into the so-called Information Age, he begins with the mostly forgotten statement that, and I'm paraphrasing here, although he's chosen to use the word information, of course he isn't really talking about the communication of actual information as that would imply a whole host of semantic issues his theory doesn't need to deal with. It was an unfortunate choice of words, one I've often wondered about. I think that he probably never imagined how willingly people would accept and build upon a concept of information without semantics. To me, what he presents in his paper equates to a theory of the communication of "meaningless information," a phrase that sounds pretty oxymoronic.

    It would have been far better to have used the word data, a word that more accurately describes what flits about on wire and wave both begotten and consumed by brainless electronics. Information, as hinted at by its -ation suffix, is really more verb than noun. It isn't some thing that can be isolated, stored or shipped about. It's a process, one requiring an enormous amount of knowledge and intelligence to carry out. People do information so easily that it seems like something given, something that is just out there in the world for the taking. We rarely notice how much of what we see isn't really out there at all.

    Consider this series of 5 numbers and letters: T 11 C W 9. This is probably about as uninformative to you as it would be to a computer. But pay attention to what happens when they're presented in a slightly different manner: WTC 9 11. To a computer it's just another arrangement of bytes, but to you this is vastly different. What's so easily missed, though, is where that difference resides. It's a phenomenon of our minds that gives us the impression that the difference is outside of us, that it's in this arrangement of characters, but it's really inside, inextricably bound up with the process of information. And that process, itself, is bound up with who we are, what we know and what we've experienced. The meaning of WTC 9 11 is different for each of us and it changes as we change. For that to be true—for the same character sequence to result in so many different meanings—the meaning, and therefore the information, can't be out there in the character sequence at all.

    In his article, Mr. Hansell touches upon this miscomprehension of information with examples of how ads bundled with a particular Gmail message are sometimes odd or inappropriate, and he gets closer to the heart of the issue when he discusses how the software behind Gmail can't really tell if "polish" is a nationality, an action or a quality. What he fails to state, though, is that Google's software, as sophisticated as it is, doesn't read in any sense of the word approaching what we mean when we say that we read e-mail. It mixes and matches and manipulates symbols and rules but it doesn't comprehend—it doesn't do information, all it does is process data.

    What bothers people about Gmail isn't the targeted ads but the misconception that Google must be figuring out what the messages mean (and, as it's the only way we're familiar with, it must be getting the same kind of meaning from them as we do). The truth of the matter is that today's state-of-the-art software-intelligence is still hung up on the self-inflicted conundrum of how to find meaning in "meaningless information." There is certainly progress being made but, in the end, it simply takes what it takes to be smart enough to do information as well as we do and what it takes is enormously more complex than anything Google's working on (not to mention enormously different).

    Software may eventually get there (and probably will). When it does it will appreciate the subtle humor in the title Mr. Hansell has (unwittingly or not) given his article. As nonsensical as it is literally, people aren't going to get stuck on the impossibility of "ads that read" and miss the intended message. We can just skip right past this kind of mistake and still manage to understand what the author was getting at. In doing so in this case, we might notice that to give an article about software's supposed ability to read our writing a title that's a perfect example of what software can't read is kind of clever and funny. How will we know when Google's software is getting smart enough to worry about? Maybe when Google can giggle. Until then, I think there are more significant and realistic things to be concerned with.

    Sincerely,
    Paul Pomeroy

    • 12 March 2010
    • Views
    • 0 Comments
    • Permalink
    • Tweet
    • 0 responses
    • Like
    • Comment
  • Paul Pomeroy's Space

    Oh a sleeping drunkard up in Central Park / Or the lion hunter In the jungle dark

    Or the Chinese dentist or the British Queen / They all fit together In the same machine

    Nice, nice, very nice
    Nice, nice, very nice
    So many people in the same device

    Oh a whirling dervish and a dancing bear / Or a Ginger Rogers and a Fred Astaire

    Or a teenage rocker or the girls in France / Yes, we all are partners in this cosmic dance

    Nice, nice, very nice
    Nice, nice, very nice
    So many people in the same device
    _________________
    From "Cat's Cradle" (Kurt Vonnegut, Jr.)

  • About Paul Pomeroy

    Oh a sleeping drunkard up in Central Park / Or the lion hunter In the jungle dark

    Or the Chinese dentist or the British Queen / They all fit together In the same machine

    Nice, nice, very nice
    Nice, nice, very nice
    So many people in the same device

    Oh a whirling dervish and a dancing bear / Or a Ginger Rogers and a Fred Astaire

    Or a teenage rocker or the girls in France / Yes, we all are partners in this cosmic dance

    Nice, nice, very nice
    Nice, nice, very nice
    So many people in the same device
    _________________
    From "Cat's Cradle" (Kurt Vonnegut, Jr.)

  • Subscribe via RSS
  • Follow Me

Theme created for Posterous by Obox