Who do we remember?

The comment section of MR can go to hell, it just erased what I wrote there because I hit the wrong keyboard button when I was about to post it. So I decided to post some of my thoughts here instead.

Tyler asked, or rather a reader of his asked him: “Who will still be famous in 10,000 years?

Tyler mentions “major religious leaders (Jesus, Buddha, etc.), Einstein, Turing, Watson and Crick, Hitler, the major classical music composers, Adam Smith, and Neil Armstrong.”

Only one of those I can agree a little with – the religious leaders. The question is of course badly posed because it doesn’t clarify what is meant by famous. I think basically nobody alive/known today has much of a chance of being famous in 10.000 years. Fame implies more than name-recognition in my mind, and I’m pretty sure that’s the most you can hope for. To take an example, Aristotle isn’t famous today; some people think they know who he was (a philosopher), a few people know a little more and that’s it. Most people have no clue who he was. How many of the 1 billion people living in Africa know more than the name, if they know that much? The Chinese? People from Brazil? And a) that’s a vastly shorter timeframe, b) when he was alive, there were only maybe 170 million people who could potentially become famous, now there are 7 billion (40 times as many) and the number keeps growing.

Let’s say 10.000 years is 300 generations. Of course it could be more, it could be less, depending on how things develop, but let’s just try out with that. Let’s say that you know all there is to know about famous guy X today and you tell your child. Your child tells everything he knows to his child, ect. for 300 generations. Now assume there’s an information loss of 1 percent per generation, i.e. that your child only gets 99 % of what you had told him about the famous guy, his/her child only gets 99% of that right and so on. After 300 generations, a little less than 5 percent of the information about the famous guy will be left even though there was very little information loss at each information exchange point. If it’s 2 percent of the information that survived that gets lost instead each generation, it’s 0,2 percent of the information that survives to the end. If it’s 5 percent, it’s 2*10^(-7) of the information that’s left [that’s about equal to 2 divided by ten million] at the end.

This assumes no Great Disasters, no increasing costs of information storage over time (something I think people tend to forget when looking at very long time frames; this is basically the same as saying that individuals don’t learn new things in the year 8000 that might be more useful to know at that point in time than what color tie Einstein wore in 1934, so there’s no new information crowding out old information. Now ask yourself i) how much you know about the dietary habits of Cleopatra, ii) why you’d want to know something like that and iii) what the growth rate of historical knowledge available generally looks like? Yes, you got it right, barring Great Disasters it is increasing over time at a very fast rate.), no overcut links at any point. This is with a society basically in stasis for 10.000 years with almost perfect information sharing over time.

To be remembered in 10.000 years, you need to be a God. But even they will have a very hard time over that kind of time frame; most humans have already forgotten all about more than 99,99 percent of all the gods we ever made up. Who remembers Dis Pater anyway? Right now I’m thinking about an SMBC-comic I’d like to have linked to illustrating this ‘religions are mortal too’-concept. It does it by showing a dad tell his son some very garbled version of a mix of current religions and the Egyptian sun gods or some such in order ‘to give the child a head start, because that’s what things will be like in the future anyway, all mixed up’ or along those lines, but I can’t remember the number, the comic has over 2000 strips and I can’t find the specific strip via google (please leave a link in the comments if you think you know which one I’m thinking about).

The time frame is the killer. Maybe Muhammed or Jesus will manage, but I severely doubt it. 10.000 years is a very long time. Maybe some combination of Superman, Muhammed and Sauron will still be around at that point. Memetic mutations happen all the time and religions will not be immune to such changes in the long run. Another thing is that to think that a religion has been around for 2000 years is a sign that that religion will keep being around for ever is probably not a good idea. Religions stay around as long as the religious people keep having babies and indoctrinating them. If they do that and don’t get killed or forced to drop their Gods by people who have other Gods, they can manage in the medium to long run. The Gods of christianity and islam have done that so far, but they killed a lot of gods in the process. Who’s to say they won’t share the fate of the gods they replaced in the very long run? Would not that be the most likely outcome? Why not?

Incidentally, when you have a child, the child shares half of your DNA, half of the other parent’s DNA. Over 300 generations, what’s left of your unique original ‘DNA-package’ is equal to 4,90909…*10^(-91). The divisor in that expression is (significantly) larger than the number of atoms in the universe. Getting children is a very bad way to try to live forever, to leave a ‘permanent imprint on the world’ or some such. You might just (think that you’ll) live on a little while longer, but that’s basically it.


February 28, 2011 - Posted by | rambling nonsense, Random stuff


  1. I think that your notion fame that implies more than mere name recognition is a bit strict. Who is then even famous today?

    Comment by Pollux | February 28, 2011 | Reply

  2. @Pollux:

    “Who is then even famous today?” – Lots of people. If I spent my time differently I would know who they are. Brad Pitt (used to be? Still is?), Michael Jackson (used to be), Julia Roberts (used to be?), ect., ect. A lot of ordinary people read about them (or whoever have replaced them in today’s media world) in magazines, and there’s often some additional information in those magazines besides their names and how they look – like who they are married to, if they’re getting divorced, if they’re having a child or a quarrel with the in-laws or something like that. Quite a few people who haven’t read the magazine have still seen them on the telly and know more than just the name.

    But in terms of the time frame involved here, it doesn’t really matter all that much – though of course it does matter some – if it only involves name-recognition or more than that. Let’s say that instead of having a quantitative measure of the information loss that happens over time, you have a measure of how likely a person who was famous in one generation is to still be famous in the next generation. The same kind of argument applies. If you have a famous guy today who has a 5 percent likelihood of no longer being famous (‘people know his name’-famous) in the next generation, the likelihood that the 300th generation still hasn’t forgotten his name is about one to five million.

    And yeah, there’s no way the likelihood that generation i forget your name is constant over time and that the relevant likelihood is 5 percent. I included both 1, 2 and 5 percent to illustrate how fast it goes when the information loss is not trivial and how even trivial information loss can really take its toll over time. That argument stands no matter which measure of fame you apply. As I said, 10.000 years is a very long time.

    The ‘scientist-idea’ (Pythagoras?) is worth a try, but I’m not sure that’d work. Science is a lot more subject to political/religious/cultural/other pressures than people like to think, and even the names of natural laws are unlikely to hold for 10.000 years.

    Comment by US | February 28, 2011 | Reply

  3. One thing I forgot to mention in the last comment: Naming conventions are functions of language. Languages evolve quite fast.

    To take an example, “the English proper noun sun developed from Old English sunne (around 725, attested in Beowulf)” [wikipedia] The current most well known (‘famous’) name of the main star of our solar system is younger than Jesus.

    You might argue that there’ll always be a global ‘lingua franca’ that will keep the scientific naming conventions alive, so that ‘Newton’s third law of motion’ will also be called that in, say, 10.000 years. There’s a lot of assumptions hidden there, but even if that’s true, that language was latin one thousand years ago and English today. a) A lingua franca is supposed to facilitate communication, so if a name is hard for a lot of people to pronounce, people using the ‘main language’ will likely switch out the name over time. b) What’s easy to pronounce varies over time. Also, if there are coordination problems (people can’t agree what the law should be called), competing names can easily cause one naming convention to be pushed out by others over time.

    The reason why I’m so sceptical is that basically nothing is static when we’re looking at a time frame that long and I think people tend to hold a lot of stuff constant that really isn’t when analyzing a problem like this, because they (we) don’t even know that we’re doing it.

    Comment by US | March 1, 2011 | Reply

  4. Jumping a little late into a very interesting discussion: I think your arguments are great – first and foremost, nothing can be kept constant over 10,000 years, and over such a long time, the signal/noise ratio (or the R^2) of any prediction is asymptotically approaching zero.

    To begin with, as you note, you need a very precise definition of “famous”. If “famous” is “talked about”, you need to ponder how people in 10,000 years will be communicating, or in a slightly broader context, what will “people” be? Tyler’s “no computer upload” assumption is widely unrealistic, IMHO. Like it or not, Google and Wikipedia are extensions of many people’s brains already, and it’s probably a matter of a few decades before many humans become functional androids (or gynoids), with hardware implanted or otherwise (permanently?) attached, or at least attachable to their brains. Human and machine, I think, are destined to merge in some fashion. If that is so, then any current concept of “famous” is terminally shot. Before I followed your link, I did not know about Dis Pater; with a Wikipedia-on-steroids-enabled brain, I know more about him than 99.999999% of people who never visited that page (the exceptions probably being a few professors of Roman religion/history who are not tech savvy). Most definitions of fame seem to revolve around “easily recognized by many” – but what if everyone recognition of Dis Pater is the same as mine? Is Dis Pater famous? The precipitous drop in information storage, search, and sharing costs simply renders celebrity/fame an obsolete concept.

    This directly spills into an argument about whether in 10,000 years there will be individuals, at least as we define them, at all. Will a computer simulation of my brain be an individual? What if it is linked to the simulations of everyone else, instantly sharing memories, experiences, ideas? Transhumanism cannot be easily dismissed, not over a 10,000-year time span, and forecasting beyond the Omega Point, whatever form it manifests itself in, is akin to trying to peak inside the event horizon of a black hole – only we have somewhat better tools at guessing about black holes.

    To sum up, you are absolutely correct that the definition/forecast of what “famous” will be then is critical to any answer, as is the definition/forecast of language, communication, individual, and even human.

    To channel Wolfgang Pauli, Tyler’s answers are not right, they are not even wrong.

    Comment by Plamus | March 2, 2011 | Reply

  5. I guess another way to think about it is this: When the total amount of information available (call it X) grows, the likelihood that you know about a specific piece of information, x, goes down. Fame is a specific piece (/set of pieces) of information about an individual.

    “The precipitous drop in information storage, search, and sharing costs simply renders celebrity/fame an obsolete concept.”

    Just to make clear, I don’t make the argument that nobody today will still be famous far into the future because the concept of fame breaks down when all information about x is available to everyone everywhere. Though it’s true, you could make that argument. My angle is a different one, I think in terms of information management issues making x irrelevant over time. In my mind it’s not a cost thing (everybody could look the guy up), it’s a quantity thing (the ‘famous’ guy have a lot more competitors than he used to, and the number of competitors will explode over time).

    There’s too much stuff to know, and we will over time know less and less of the total body of available knowledge. No individual’s fame/actions can survive that process of knowledge accumulation in the long run.

    I’m not quite sure where I stand on the information cost thing. Yes, nominal information costs have gone down and will decrease over time. The amount of knowledge available grows and grows. But every second individual i spend on issue x is a second individual i don’t/can’t spend on (X-x), which is growing much, much faster than x. The opportunity costs related to any specific knowledge accumulation process is increasing dramatically over time. That and heterogenous preferences of human individuals make it very difficult for an individual to leave any kind of permanent mark on anything in the long run.

    Comment by US | March 2, 2011 | Reply

  6. I am not sure I quite agree with the definition “[f]ame is a specific piece (/set of pieces) of information about an individual”. The distribution of the (pieces of) information seems to me to matter. There seems to be a network externality – if you apply Beckstrom’s theorem/model, fame disappears.

    Your last paragraph poses a legitimate question – will the growing opportunity cost of, for lack of a better wording, “paying attention” to a specific piece of information outweigh declining difficulty of its retrieval and learning. But here is where I am agnostic, as I cannot begin to define what information retrieval and learning will be like in 10,000 years. If someone had mentioned Dis Pater to me 20 years ago, I probably would have nodded and forgotten about it soon, since learing about it would have required going to a well-stocked library. Nowadays, I became an expert (by the definitions of 20 years ago) in about 5 minutes – clicking on the link and reading the info. In the future, that might require no conscious effort at all – all the information may be loaded in my brain upon receiving the “Dis Pater” string. I am envisioning a model where everyone is more or less equally famous, and the only requirement for instant, but fleeting boost in fame being being “talked about” at the moment. For illustration, imagine a world of only you and me, both with instant access to knowledge, and talking. You mention Dis Pater; I instantly load the information; Dis Pater achieves fame – everyone (both of us knows) about him. I mention Tengri/Tangra. Same thing. In the case with many more users, the most famous becomes the name that the biggest sub-networks “talk about”.

    On the flip side, I do agree that, in your example, the average x/(X-x) will keep going down. My claim is that virtually all x’s willcome with a cost of zero, and your claim is that X will be so large as to make the ratio indistinguishable from zero for all x’s. The two effects are not incompatible, and lead to the same outcome – fame becomes meaningless, akin to salience, evanescent, and non-sustainable.

    Comment by Plamus | March 3, 2011 | Reply

  7. [I erased the comment of ‘XX storage (whatever)’. The commenter did not bother to read the entire post before commenting and did not even make an argument, (/s?)he merely stated an opinion. The commenter did it no doubt because (/s?)he expected that the comment would lead to an increase in hits to the site of the company that employs the person. I consider such comments spam which contributes nothing to the discussion, even if they’re written by real people rather than computer algorithms. If people want to advertise their products online, they should do it at their own sites instead of polluting other people’s blogs. If the commenter were to remove the link to the site and choose a name or internet alias not obviously associated with a company employing the person, I’d let such a comment through, even if it doesn’t contribute much, if anything, to the discussion.]

    Comment by US | March 3, 2011 | Reply

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: