Month: October 2019


Encyclosphere


Larry Sanger, the first “editor” of Wikipedia (speaking loosely), is launching a new project to define technical sharing and interopability standards for online encyclopedias.

The aim is to create an “encyclopshere” of online encylopedias, based on the example of the “blogosphere” of online comment, and thereby route around the fairly obvious flaws with Wikipedia today.

I wouldn’t exactly buy into this project, but of course I don’t need to. Even if it is not likely to succeed, there is still a chance of it producing something valuable, either in the form of a product, or in the form of a lesson about what makes collaborative projects and search for truth work.

I’m not going to dig into it all today. It’s a large question. This post is a collection of resources, some from Sanger’s project, some from elsewhere, that I think are relevant.

Text of speech by Sanger announcing the project:

2-part post by Sanger on Slashdot in 2005 about the early history of Wikipedia:

David Chapman on subculture evolution, and Venkatesh Rao on identity

My own earliest memory of an online community becoming poisoned and dying is the Eternal September of 1993.

Missing resources

These are things that should be listed here but I can’t find

  • Moldbug posted an idea for a site where you could basically attach commentary to articles? A sort of cross between Wikipedia and Gab’s Dissenter product (though it was long before Gab, of course). Someone actually implemented a first cut, and I used it.
  • Stack Overflow set out to be an online reference of solutions to programming problems. Its founders thought quite deeply about what it would take to build the site and its community of contributors, and they were extremely successful in achieving their aims. I’m pretty sure i’ve seen a good longform account of this somewhere, probably by Atwood. This little piece gives a flavour of the way they think about things. https://stackoverflow.blog/2010/01/04/stack-overflow-where-we-hate-fun/

A few of my own pieces I think are relevant:

Brief thoughts on culture & community

Each of these needs fleshing out. Some of them are discussed in the links above.

  • A community has to stand for something, and have a way of making sure it will continue to stand for what it stands for. (Oh, that’s Goal-Content Integrity)
  • If it’s too restrictive about what is allowed in, a few early adopters with a lot of energy will turn it into their own private club
  • Conquest’s Laws
  • If you let in trolls, you will end up with a community of that subset of people who are willing to put up with the trolls
  • A system that works well when the real-world stakes are low will immediately fail catastrophically when the real-world stakes get high (Mtgox, arguably wikipedia)
  • It would be ideal to just not have a community, just a project, but that’s not possible (one of the points from Sanger’s slashdot history)
  • All the real value comes from the best contributions. But if you don’t have the mediocre contributions, you don’t have anything. Most of your project only exists in order to be the venue where the rare valuable stuff happens.
  • Rules are meaningless independent of the people who follow and enforce them. Identical rules will succeed in one place and fail in another. But rules still matter.

Requirements for Encyclopedia Protocol

This isn’t remotely definitive, but I have actually tried to produce encyclopedia content, and run into obstacles I didn’t expect, so here’s some stuff to throw on the whiteboard. (“Requirements” in the technical sense of what needs the solution is trying to satisfy, not in the sense these are all definitely 100% necessary)

  • Obvious stuff: rich text, embedded pictures and diagrams.
  • Internal linkage. Each entry/article is a first-class sharable entity. It has a single title. It can be referenced by that title, plus source and version
  • Rendering. Debatable, but personally I strongly want to be able easily to turn an encylopedia into a genuine printable format, as well as it being an easy-to-access web resource
  • External links. You have to be able easily to reference external resources in a way that is compatible with standard bibiliography/citation techniques. I’ve found this frustratingly difficult.
  • Revision control. You need to know who changed what, when.

Aside on Software Bloat

October 17, 2019

Main

Comments Off on Aside on Software Bloat


I fell yesterday into the sad bitching about how big and slow software has become. This is a very old complaint — the EMACS editor used to be mocked as “eight megabytes and constantly swapping” back when eight megabytes was a huge amount of memory, but that rounds down to zero pixels on a graph of memory utilisation on a modern laptop.

I retailed the usual whines about electron and so on, but really any disagreements are at the margin: the real underlying reasons for software bloat are, unfortunately, good reasons.

Here’s a more interesting illustration: I recently watched this video. It’s a one-hour presentation by a Microsoft developer explaining MS’s implementation of the new C++ charconv header.

This is a library for converting numbers to and from decimal format. Computers internally work with fractions or large numbers in a binary floating-point format, so you have to be able to convert that format to and from a string of decimal digits.

All computers have to do that. My ZX81 did it 27 years ago (though its predecessor the ZX80 couldn’t — it worked only with whole numbers). It was part of the 8K of software built into the machine, along with the full floating-point mathematics support in software.

The new charconv library the Microsoft guy was presenting contains 5300 lines of C++, taking 221K of code and another 400K of data tables.

And — to make it clear — it’s awesome. I was glued to the one-hour video on what they’ve done. The clever bit is getting the right number of decimal digits.

The technical problem is that a fractional decimal number usually doesn’t convert exactly to a binary number. Therefore when you convert from decimal to binary — to do any calculations with the number — you’re getting a slightly different number. That’s OK. But then when you convert back from binary to decimal, you can get an exact decimal representation of the binary approximation of the original decimal number, so it’s a bit different to what you started with. That’s quite annoying. It can even cause program bugs.

The current C++ language standard says the new functions to convert binary to decimal should be able to round to the shortest decimal representation that will exactly convert back to the same binary value. That’s difficult to work out, and really really difficult to work out quickly. In fact a new method of doing it was produced by a guy called Ulf Adams at Google just in 2018, and the Microsoft team have implemented that algorithm for their standard library.

This is all very cool. But the relevance to my point is that when I, in a C++ program, decide to output a floating point number in a decimal form, maybe to save into a database or communicate to another program, and I use this standard to_chars function, I’m invoking all this mass of ingenious code to do the conversion. I may or may not notice that the rounding is now perfect in a way it never was before from 1982 to 2018. I probably won’t notice the 600K of library code that’s being used by my program. If I hadn’t happened to see this video, I would never have had any idea about any of this.

That’s for printing a number! It seems close to the simplest thing a computer program can do. Everything else in my program, dealing with text, or graphics, or networking, or anything has gone through this kind of improvement, often many times. Sometimes your program is getting real benefit from the improvements. Sometimes it’s getting the effect of the improvement, but they don’t make any useful difference for you. Sometimes you aren’t using the new functionality at all, but it still gets included when your program runs. That’s slightly unfortunate, but simplicity is valuable, and grabbing big chunks of functionality is simpler than finely selecting them.

The bottom line is that everything has a cost, even slimming down software, and if you insist on using a low-end 6-year-old computer like I do then it’s not worth most developers’ time to cater to you. I do think there is too much bloat, but it’s about tradeoffs at the margin; there will always be bloat, and that’s OK.


Sunk Moral Costs

October 15, 2019

Main

Comments Off on Sunk Moral Costs


I don’t understand Syria, and I’m not going to, and I’m OK with that. Trump’s pullout may be bad for America for all I know.

The concrete harmful impact of Russia having a lot of influence in Syria (as it did in the 1980s) isn’t spelled out, instead we just get innuendo.

I tweeted that Kurds will always be allies in destabilising, and always be enemies of peace, because of their situation as a stateless cross-border group. That’s simplistic, but if it’s not true someone needs to explain how. Peace in any of the countries in which they have large populations has to include either (a) they give up their claim to statehood, or (b) they achieve their own state, and I have never heard anyone suggest that (b) is a realistic possibility. There is a chance in any one country that you could get an autonomy-based settlement short of statehood which is beneficial for them, but while the other countries in which they have large populations are unstable, that can’t be a peaceful settlement, because they will still be fighting in the others. As I tweeted, none of this is their fault — it seems they were completely screwed in the 20th Century but this is the position today.

If there’s any coherent view coming from the US establishment, it’s anti-Iran. They may have a good reason for that, but I don’t know what it is. The reason probably has a lot to do with either Israel or Saudi or both, but I don’t expect to ever find an answer I can be sure is true.

Syria has been a bloodbath since the beginning of the Arab Spring attempt to depose Assad. Anyone suddenly upset about the humanitarian impact this week can be dismissed out of hand.

“Kurds were our allies”. How is that, exactly? I asked on twitter, sarcastically, for links to the announcements of and debates of this policy. It was made ad-hoc by the military and civil service. The president never talked to the electorate about it. Quite possibly the president (Obama) never even knew about it. Which is perfectly OK. But there is sleight of hand here. The line we are getting is: “We allied with the Kurds and relied on them, now we need to stand up for them”. The two “we” in there are two different groups. The opaque Washington foreign-policy establishment allied with the Kurds, without input from or notification of the general public. Now the voters are being asked by the media to stand by some implied commitment they played no part in making.

1) So much context has been lost and recent history revised in the coverage of this growing crisis between Turkey and Syria. US always assured Ankara that their support for the YPG was ‘temporary, tactical and transactional’ – a US diplomat quoted here in my new book on Erdogan

@hannahluci https://twitter.com/hannahluci/status/1184012129562775552

From around 14th October, the Kurds have made some kind of arrangement with the Syrian Government, and the narrative has switched from “it’s terrible to abandon the Kurds” to “Now the Russians are winning”. This is utterly disgraceful. It entirely proves that the complaints about the fate of the Kurds the previous days were insincere. Had the concern really been for the Kurds, then Monday would have been a day of rejoicing at their safety. Instead, the opposition to the withdrawl policy stays the same but the reasons change.

It is because of this sort of thing that I automatically disregard all foreign policy arguments that are made on humanitarian grounds. I don’t even consider the possibility that they might be well-founded. The concept of intervening internationally to protect civilians is 100% discredited in my eyes.

Around 500,000 human beings were killed in Syria while Barack Obama was president and leading for a “political settlement” to that civil war Media has been more outraged in the last 72 hours over our Syria policy than they were at any point during 7 years of slaughter
Ask why

@BuckSexton https://twitter.com/BuckSexton/status/1183812563261382656

Kinda telling that the intensity of Online Outrage expressed by Smart People today over the Kingsman-meme isn’t any perceptibly different than the Online Outrage they were emoting yesterday or the day before over, like, The Kurds being slaughtered
it’s all a video game

@soncharm https://twitter.com/soncharm/status/1183750875321438208

Trump, though I find him amusing, I consider no more trustworthy than the rest of them. I am not able to judge whether his policies are good or bad, but he is the only person who makes arguments for his Syria policy which make sense. The arguments against are always obviously dishonest (like the ABC gun show footage), insincere, or rest on vague unstated assumptions (such as that nothing that Russia wants can be allowed).

The FSA leader who John McCain took a picture with is now part of the invasion of Northern Syria, which the hawks are insisting we must oppose.

@j_arthur_bloom https://twitter.com/j_arthur_bloom/status/1183364011708080128

There’s another related point, more subtle but much more general. Modern thought does not admit of a distinction between crimes of commission and crimes of omission. To a naive rationalist, causing harm and allowing harm to happen are equivalent. But like so many arguments you hear today, the equivalence rests on an entirely unrealistic level of certainty towards the assumptions that are being made about the results of action or inaction. The potential for very large unexpected harmful effects is very much greater in military action than it is in inaction, and the expected benefits of action have to be large enough to outweigh that category of risk. That is equally true whether the harms and benefits in question are political, financial or humanitarian.

Tweet links:

  • https://twitter.com/anomalyuk/status/1183128988803371009
  • https://twitter.com/anomalyuk/status/1183135846226108416
  • https://twitter.com/anomalyuk/status/1183450270585540609
  • https://twitter.com/anomalyuk/status/1184063105669709824


Double Standards Again

October 11, 2019

Main

Comments Off on Double Standards Again


A few years back, some American sportsmen made a big deal about pushing their politics — basically BLM — during events and interviews and things.

On the right, this was quite widely seen as unpleasant. For example, “Lion of the Blogosphere” wrote two years ago on his blog:

The average NFL player is paid $1.9 million/year to entertain prole whites who love the American flag, and part of the show is that they are supposed act patriotic when the National Anthem is playing.

If I inflicted my political opinions on my employer’s customers I’d be fired, and I get paid a lot less than $1.9 million/year.

https://lionoftheblogosphere.wordpress.com/2017/09/24/boycottnfl-part-2/

I happen to have picked @LionBlogosphere — I follow a whole lot of weirdos on twitter, but he is much closer to a mainstream American Conservative.

Anyway, not that it’s really any of my business, I agree with him entirely; I think that’s a totally reasonable position for him to take. And, while the whole issue has dragged on, not 100% resolved, his side has at any rate not definitively lost. I think it has come out slightly ahead, and the leagues and teams have mostly taken the view that their players should not insult their spectators.

But now, of course, we have the China thing. NBA basketball is huge in China, and there’s some kind of protest movement in Hong Kong that I don’t know much about, and some basketball people made sympathetic noises about the HK protesters, and the Chinese government was very upset.

I’m trying hard to remain ignorant of the Hong Kong thing. A new extradition law was brought in, or something, and that’s maybe against what the Chinese government had previously promised in terms of HK’s autonomy, and there have been protests going on for a month or three, which have been getting the whole unquestioning popular support in the West that I hate so much. So my reflex view is on the side of the CPC, but of course they may really be doing something bad for all I know.

Either way, right or wrong, the Chinese government are not going to be friendly to foreigners who take sides against them. They are not going to allow them a public platform in their country, any more than they allow their enemies within the country a public platform. And if you are a basketball team that’s going to do tours and broadcast games in China, the Chinese government is your customer. And, as @LionBlogosphere said, “If I inflicted my political opinions on my employer’s customers I’d be fired”.

So I think we’re all agreed.

Except, of course, that @LionBlogosphere today retweeted Ted Cruz saying,

It is outrageous that the Chinese Communist Party is using its economic power to suppress the speech of Americans inside the United States:

https://www.tedcruz.org/news/the-hill-ocasio-cortez-ted-cruz-join-colleagues-blasting-nba-for-outrageous-response-to-china/

Again, I’m not especially objecting to Lion, I just follow him and not, say, Ted Cruz, who in September 2016 said

Here’s a peaceful protest: never buy another shoe, shirt, or jersey of rich spoiled athletes who dishonor our flag. https://t.co/GrGPYX8HCh

If you are putting economic pressure on sports teams, and you expect them to respond, well, so can their other customers, and maybe it’s reasonable for the teams and leagues to respond to that too.

Surely, you can draw a distinction between #BLM and “Free Hong Kong”. The teams in question are all American, and you can demand of them a loyalty to America while they have no equivalent duty of loyalty to China. But they would no doubt claim they were being loyal to America by seeking to change it in the way they sought — the real objection is they were offending their customers. Another distinction is that it wasn’t the US government putting pressure on the teams to censor themselves, but the Chinese government is doing so. Well, the relationship between citizens and government is different in China than it is in the US. International sport has long depended on not bringing one country’s politics into another. Most Americans probably think that China should be a democracy with free speech. But it isn’t. If Ted Cruz thinks that means the NBA shouldn’t do business with China, that’s a coherent position. But if he doesn’t think that, then obviously the NBA will take steps to make their product marketable there, if there is commercial reason to. And if he thinks that this is a matter of the principle of free speech, which should outweigh that commercial reason — then why did he think the opposite in 2016?

Now, you can point to a contradiction between the relative willingness of the entertainment industry to allow opposition to the US government, and their very rapid arse-covering with respect to the Chinese government. That reflects both their own political biases, and the relative power of the US and Chinese governments over their ability to do business. But if the other side is contradicting itself, it doesn’t help to contradict yourself, even worse, in the opposite direction. Ted Cruz was right in 2016, he (and @LionBlogosphere) could very justifiably spend these weeks banging on about the inconsistency in sportsmen demanding the right to insult symbols of the American nation while being careful to avoid insulting the Chinese nation. Instead they just destroy their own previous arguments.

Finally, the 2016 argument was the more important of the two. By flipping now, they are putting themselves in the wrong for next time. You said that political protests at sporting events was a matter of free speech.


Decentralised Monopolies

October 11, 2019

Main

Comments Off on Decentralised Monopolies


[context: I never actually made a decision to step away from blogging and twitter, I just had more interesting things to do]

Ten to twenty years ago, one of the big buzzwords was decentralisation. New communication technology means that things that used to have to be organised by a central body can now be done spontaneously between users.

By decentralising an activity, you remove the bottleneck of the central coordinator. You reduce the status war of having a “leader” (who leads Extinction Rebellion?) You can evade countermeasures.

Also, it was generally assumed, you create choice. If you can have one decentralised network, you can have two, or ten, or a million.

Some of that was hype, some of it is true. But the issue of choice has turned out to be the most interesting. Replacing a hierarchical organisation (of authority, or of communication framework) with a decentralised network does do away with the nominal leader or controller. But it doesn’t do away with network effects. Indeed, by removing some of the barriers to scale it can greatly increase them.

The result of decentralisation plus network effects is the decentralised monopoly.

Outside of one special case I’ll come back to, I think the best examples of decentralised monopolies are open-source software projects. The essence of these projects is that there is no control; anyone can take the source code, change it, build it, and distribute it. That goes even for large widely-used things like the Linux kernel or the Apache webserver. But, most of the time, there’s no good reason to do so. The result is a voluntary centralisation.

The recent trend has been to recentralisation: the old centralised Television and newspapers give way to decentralised blogs and podcasts, which gives way to recentralised Facebook and Youtube. But that is still voluntary. The newspapers and television stations had control because they owned the actual infrastructure. Facebook and Google own some infrastructure, but in comparison to their actual business that’s negligible. Alphabet market cap is apparently USD 825bn, their balance sheet lists plant & equipment at USD 60bn.

That’s not really counterintuitive. It’s just network effects, and/or Schelling points. I’d love to see that recentralisation reversed, but I don’t think it’s possible — if there’s going to be a monopoly because of network effects, then a business that can pay to market its network is bound to outcompete a network that doesn’t have a central owner.

Where you have something that is made of decentralised contributions, the network effects get so much stronger than they do for pure consumers. You have to put real work into contributing, and the return on doing that work depends on the contributions of others.

Wikipedia is a prime example of this. Now it isn’t 100% decentralized, as somebody owns the servers and the domain name, so while anyone can contribute, there is an actual hierarchy with a root of sorts. But if you could design out that root authority, I don’t think it would change much. There is still just one Wikipedia, and all the work that is contributed to it can’t be contributed anywhere else instead without severely reducing its value. The decentralised selection of content (selection is of course the main work of an encyclopedia) still has to be done by those people who show up to do it, coordinated in some way that enough of them can put up with. If you fork it to produce some rival, as has been done a few times, your rival has none of the value.

That’s not to deny Wikipedia’s many flaws. There are many areas where it is systematically bad. But I think Wikipedia is what it necessarily must be. That’s the real point of decentralised monopoly — if it is not under some central control, then there is nobody who can make it other than what it is. This echoes, somewhat, the repost about political parties: how can one political party, open to anyone and run by its members, be different from any other political party open to anyone and run by its members? In a sense, complete openness is the most unyielding authority of all.

The one special case of decentralised monopoly is, of course, the bitcoin blockchain. It’s special because being decentralised and a monopoly are not incidental attributes, but the central aims of its design. As such, it bears the same relationship to a study of decentralised monopoly as dog breeding does to natural selection.

I explained the essence of what makes bitcoin before: it’s a voting system where you are fined for voting on the losing side. That rule is guaranteed to produce a consensus, and the consensus is likely to be “correct” from the point of view of the contributors. That’s the design aim, but see how similar it is to making a contribution to any other open collaborative project. Work that is put into maintaining a Myspace page, or keeping a presence on Gab, is nearly as wasted as the work spent mining an orphan bitcoin block. Either I’m contributing to the project that everyone else is using, or I’m shouting in the wilderness.

The most important aspect that drives this authority is probably not decentralisation as such, or even openness to contributions, it’s being public. Ed West tweeted yesterday that he wished he could maintain different “flavours” of his twitter stream — a toned-down one for the normies, and a more hard-hitting one for the fans. I’ve tried to do that sort of thing numerous times, but it never worked; it failed on the same point: if I wasn’t blogging or tweeting as AnomalyUK, I was losing most of my audience. Moldbug outed himself because he couldn’t resist discussing his technical work on UR. He didn’t have to do that — it is possible to be two people online, but it’s not possible to be one person with two public faces. They automatically become one under the pressure of being public.

That basically is the same point as I made about the decline of conspiracy: the political mode we are now in is the one that you get when nobody conspires to prevent it. As such, you would expect it to be the historical norm, unless it destroys itself. And it isn’t the historical norm. (This is just a restatement of Jim on left singularities).




Recent Comments