Monthly Archives: March 2013

Crowdfunding vs angels vs VCs

ESR has some good analysis of how crowdfunding will interact with the existing angel-investor/VC process.

Armed & Dangerous :: Eric S Raymond :: What does crowdfunding replace or displace?

I think he's on the right track, but I wonder how willing will people be to contribute to crowdfunded projects by established firms? It seems like people are more willing to contribute to the (imagined?) plucky tinkerer in his garage than to an existing small-medium firm who is just looking for lower cost of capital.

In my view Kickstarter is just busking, especially now that they've moved to discourage people from using it as a platform for prepayment. Are you really going to be willing to keep chipping in money to fund projects from existing companies with a track record of product releases?

Maybe there are simply too many Scots in my ancestry for me to get so loose with the purse strings, or maybe I've just been living on a grad student's budget for too long, but crowdfunding doesn't have much appeal to me either way. (Unless it's just a pre-order system.) I'm glad it exists, and I'm glad other people are getting utility from it, but it's not for me. Not until I can get some equity out of the deal.

PS Sort of related — The Economist: Babbage :: After the Crowd Leaves

Posted in Business / Economics | 2 Comments


Moneyness :: JP Koning :: Ripple, or Bills of Exchange 2.0

Here's some interesting news. Ripple is finally being implemented.

What is Ripple? Ripple is an open source P2P credit system dreamt up by Ryan Fugger in 2004. Its mission is to provide a non-banking payments alternative by decentralizing the process of creating and circulating highly liquid IOUs. Put differently, Ripple offers an environment in which individuals can be their own credit-issuing and credit-accepting banks. Ripple has always remained conceptual. But now a team of developers lead by Jed McCaleb, founder of MtGox, the world's largest bitcoin exchange, are implementing a living breathing Ripple network.

Ripple might seem to be unprecedented, but the decentralized credit system it envisions existed centuries ago in the form of the historical bills of exchange system. ...

This is a fascinating system. Koning's post has a good intro to Ripple and some good background on Renaissance-era banking.

I'm a little concerned about one thing with Ripple though: who am I supposed to endorse? The people I trust the most are very close friends and family members, so they'd be the obvious choice. But there's a reason it's a bad idea to go into business with friends and family: any monetary disputes have an extra layer of difficulty.

It makes most sense for me to endorse a Ripple of a close friend, but in a way those are the people for whom it is hardest to collect on an IOU from. This may sound cynical, but in every clique of friends I've ever seen there is a set of perpetually out-standing small debts. You know you can be delinquent in returning that drill, or suitcase, or $40 from from the time you forgot to hit the ATM before dinner and Dave spotted you your share, because there's a whole history of positive associations to balance out the other side of the ledger. Dave is not going to scuttle your entire relationship because you still haven't paid your share of the gas from the road trip you too took back in 2010. In limited quantities perhaps this perpetually-outstanding debt thing is actually a strength of a system like Ripple, but I'm not putting money into a system unless I have some high confidence I'll actually be able to settle up my accounts at some point of my choosing.

Bill of Exchange, 1779
Bill of Exchange, 1779

Here's my other concern: how public will it be who you've endorsed? Specifically, if a deadbeat cousin or flaky co-worker endorses your Ripples, will you face social opprobrium for not reciprocating? Because that's the dynamic most social networks have evolved to. Will someone with the upper hand in a combined business/social setting, like your boss or landlord, be able to turn that to their advantage in terms of pressuring people into Ripple endorsements?

Maybe I'm thinking about this on too personal of a basis. Does this kind of system make more sense if you're using it like the original bills of exchange, i.e. endorsing the debts of your clients, suppliers, business partners, etc? Koning's recent post on Bitcoin has at least convinced me that Ripple makes a lot more sense than Bitcoin does.

Posted in Business / Economics | Leave a comment

An Untold Stories Tax?

Carnivale - Season 2HBO canceled Carnivàle, a serial show with deep mythology, after two of a planned six seasons. The remainder of Creator Daniel Knauf's story was never told.

The AV Club :: Daniel Knauf tells us his plan for the end of Carnivàle

AVC: Have you ever considered trying to do it as a novel or a comic book [of the remaining story line]?

DK: Constantly. Yeah. Marvel, we had it all set up. At one point, they wanted to go forward and do a series of graphic novels, and they just couldn’t turn the corner with HBO. Since then, yeah, I’ve considered it. But one of the things that makes me a little crazy about Hollywood is, they’re idiots when it comes to their contractual stuff. If I write a novel, it’s like Random House publishes the novel, copyrights it, but when you do business in Hollywood, they say, “Everything in this thing, in all forms, in all potential forms invented and uninvented…” The language is draconian! “…throughout the universe. We own everything in your head. We own everything.” And it’s like, “If you own everything, at least exploit those rights, please. Could you please exploit the rights? And if you’re not going to exploit the rights, can I at least have them back, so I can exploit them?” It’s just a silly way of doing business. [...]

It didn’t make sense to spend $3.5 million an episode. So let’s do a graphic novel. Let’s tell the story!” But they’re on to other toys now. It’s like doing business with that kid down the street whose parents give him really bitchin’ toys, and he’d just leave them broken in the backyard. It makes me crazy, Hollywood.

I think you could make an analogy to Georgist taxation here, or perhaps more generally Gobry's argument in favor of the French wealth tax.

If property ultimately derives from mixing your labor with things, it's not unreasonable to suggest that people have an ongoing responsibility to continue doing so. If you hold some property, most especially land, you may have a responsibility to society to put it to productive use. (We're talking about theory here, not practice. The arbitration of what counts as responsible, productive use is nearly impossible in practice and so even if you had such a responsibility in theory it is likely best if that responsibility is never legislated into reality.) Gobry's argument is, briefly, that capital gains taxes discourage people to put their resources to use, while wealth taxes do the opposite. In essence, capital gains taxes makes it more expensive to put your resources to work, so people do less of that. OTOH if you're going to loose %1 of your accumulated resources anyway to a wealth tax that gives you reason to put your resources out in the world to try to get them to grow more than the amount you'll lose.

If a studio owns the rights to further adaptations in other media, do they have a responsibility to society to actually use those rights? Land may be a special case of property, because people aren't making any more of it. Or so I gather the Georgists, the Diggers, etc. would say. But people aren't making any more Carnivàle either. That idea can only be invented once, only to be owned by one person, just like a particular acre of land. Does that put an extra responsibility on HBO to do something with it? If an owner of arable land has a onus to see it cultivated, does the owner of fecund IP have a similar onus to see it reified?

I have absolutely no idea how you would actually structure this as a policy. Doing so in a way that wouldn't put the actual tax burden on the creators rather than studios would be harder yet. Even so, I think it's an interesting way to look at the ethical responsibility of content owners, if not a way to structure their legal responsibility.

Posted in Business / Economics | Tagged , | Leave a comment


Google is shutting down the Google Reader service, as you may have heard.

I am not happy about this, but I'm also not throwing fits about it. I spent more time in my Reader tab than any other by a huge margin. I'm sure there's some other service out there that will fill the void. Google is also being pretty considerate about this: they're giving users several months warning and they've made tools available to export your subscription data.

RWCG has had a couple of posts taking the wind out the sails of all the people who are tearing out their hair and rending their garments and casting the evil eye towards Mountain View. Overall I think he's on the right track, but this post doesn't quite line up for me:

RWCG | Crimson Reach | Google Reader: The end of the ‘free’ era

How dare Google shut down a service I made heavy use of for years and years and paid nothing whatsoever for! [...] I wonder if this event marks the end of the ‘free’ era. You know the free era, it’s the era where this was the prevailing business philosophy:

1. Drive competing services out of business with a free service (subsidized by a profitable product).
2. Cancel free service.
3. ???

Actually no, that’s still snarky. It’s more like this:

1. Company gives away something for free.
2. People like it and use it.
3. This makes people think the company is cool. Their friend.
4. When it goes public, they buy its stock, cuz it’s so cool, and everyone they know uses and likes it. Surely that’s gotta mean something, stock-wise.
5. People continue to use the free thing and come to not only rely on it but expect it as their birthright.
6. ...
7. Profit?

According to this philosophy, giving away cool stuff for free was the wave of the future. It’s what all smart, cool companies did. Only backward knuckle-dragging idiots couldn’t figure out how this added up to a business model. Economic realities were no longer reality.

I think he's short-changing the business potential of a product ("product"?) like Google Reader. There's no direct line between "make Reader & give it away free" and "profit," but this approach still has some uses.

1. Yes, it makes people think you're cool and friendly and not-evil. Many firms do things for that reason alone. They spend billions on things way outside their core competencies just so people think they're cool and friendly. Isn't that the entire point of the "Corporate Social Responsibility" fad?

Google paying its employees to create Reader is in it's wheelhouse; it makes sense. Far more sense than, for example, Chrysler paying its employees to lay bathroom tile in poor neighborhoods.

2. Providing services like Reader makes Google look cool to people generally, but more importantly it makes them look cool to geeks. It's a punchline that a firm's number one asset is it's people, but that's pretty true about Google. They can do what they do because they get the pick of the litter of hackers.

I went to career fair my CS department sponsored a month or so ago. The line for the Google table was literally out the door. Most people in my graduate program are angling for academic jobs. Google is one of maybe four private companies that people will be impressed you're interviewing with.

3. Projects like Reader not only motivate applicants, they motivate employees.

Talent and productivity are extremely unevenly distributed in coders. The best are many orders of magnitude better than the median; the bottom decile (conservatively) have negative productivity. You usually don't get the best by offering them orders of magnitude more money, you get them by giving them cool problems to work on.

If you're excited about spending some time developing X, there's a good chance Google will let you do that. (At least in comparison to if you were working at Initech.) What's more, there's a chance Google will roll out X to millions of people, like they did with Reader. I can't stress enough how big of a motivator that can be.

4. Google's strategy for a while has been that anything they do to make people want to use the internet more is good for them, because they capture a dominant slice of the ad revenue online. More people spending more time online is better for Google, period. Reader fits into that. That's not a strategy that will work forever, or for many (any?) other companies. It can also be used to justify a lot of wasted "investments." But it's also true.

5. Google lives and dies off of data. Reader could have been generating that for them. I have no idea how much they actually learned from people's reading habits, if anything, but it had the potential to be a goldmine.

If you can predict people's age, race, religion, political party and drug use only using publicly available "likes" on Facebook, what could you do with my online reading habits? (Answer: sooooo much.)

I have no idea if it was a good idea or a bad one for Google to shut off Reader. I'm skeptical, but I realize I have none of the facts. I can't imagine it would cost that much to keep it running as is, especially compared to their other projects. I'm not sure what better use they have for the resources they're redeploying. I'm curious that they didn't even try to make it ad supported. Hell, I would have even paid directly to keep using it, and I pay for approximately zero online subscriptions.

Again, I don't know what they know. But I do know that "there's no revenue from this free product; let's shut it down" should not be the beginning and end of this decision making.

Posted in Business / Economics, CS / Science / Tech / Coding | Tagged , | Leave a comment

Computational Cartography & Gerrymandering

My brief recent dabbling with computational cartography (see Part 1 & Part 2) was inspired by Neil Freeman. His project was initiated as a (speculative) reform to the Electoral College.

Incidentally, this is a great idea. Every four years when whichever party looses the White House starts complaining about how silly the Electoral College is I feel compelled to point out that it's actually an excellent way to maximize the expected impact of any one voter's vote. The catch is that this requires states to have equally-weighted populations, which obviously we don't.

I just ran a quick Monte Carlo to check the numbers on this. Let's say you voted in 10 million elections in your life, and in each one there were 100 million voters. (This is obviously way too many elections, but I wanted something that didn't require me to report the results in scientific notation.) Let's further stipulate that every one of those races was close: a dead heat with a standard deviation of 1%. In this idealized case you could expect to be the tie-breaking vote — thus having your vote "matter" — in less than four of them. Four in ten million.

cnt = 10000000;
pop = 100000000;
for i=1:1000,
  results = round(random('normal',n/2,pop*.01,[cnt 1]));
mean(ties)     => ans = 3.9160


If, on the other hand, there were 51 states of 1,960,784 voters each, you would be the tie-breaker for your state 203 times. This would only matter if the other 50 states were tied 25/25. You can expect this to happen 11.2% of the time. Which means your chance of being the deciding vote goes from 4-in-10,000,000 to 23-in-10,000,000 which means your vote "matters" more than five times as much in the (modified) Electoral College system than it does in the winner-take-all system.

All of this is far too numerical to explain to someone in casual conversation, so my standard defense of the Electoral College is to agree that we should abolish it as soon as the winner of the World Series is decided by whichever teams scores the most runs in 63 innings rather than the team which wins four of seven games. The reasons to prefer both the Electoral College and best-of-seven champions are the same, but I'm already way too far off track to cover that now.

Okay, back to the point. Equally-sized states would be cool. But the chances of us redrawing state boundaries is zero.

Despite this, I think an automated system to draw boundaries for equally populated regions is far more than just an exercise. Why? Because while our state boundaries are effectively carved in stone by tradition — or in the case of my home of Maryland, literally carved in stone — there are political borders that are re-drawn with regularity: congressional districts.

That's how we end up with absurdity like this:

Maryland 3rd
Maryland's (totally reasonable, not at all farcical) 3rd Congressional District

Gerrymandering was one of the chief complaints than Glenn Reynolds made in his recent EconTalk appearance:

I'm also pretty much agnostic on term limits. I used to be opposed to them. And as I see how effective gerrymandering is, which I think is one of the reasons people hate Congress but love their Congressmen, or at least tolerate their Congressmen, because of gerrymandering, I think that's something that is difficult to deal with and probably wouldn't be helped that much by term limits. If you've got a gerrymandered district, especially with modern gerrymandering technology, the new guy who gets elected in it is likely to look an awful lot like the old guy because the district is so one-dimensional.

Taking the "especially with modern gerrymandering technology" as a given, how can we use technology to pre-empt this trend? It's usually best to address technical problems with technical solutions rather than social ones.

One option is to draw the boundaries the same way we do now, but reject them if their score on some measure falls below a certain acceptable level. For example, IDV Solutions has ranked districts by perimeter/sqrt(area). This is, by their own admission, a crude way to do things. To name just one problem, it favors districts that fall on state borders like the Mason-Dixon Line over those that fall on borders like the Potomac River.

Other measures are possible, obviously. One that I think would be useful is based on nearest neighbors. In a non-gerrymandered map the vast majority of addresses will have all of their k nearest neighbors in the same district. As districts become more complex, more people will be have large proportions of their neighbors in other districts. If you limit the neighbors to those within the state, which is reasonable, you minimize the Mason-Dixon/Potomac problem.

Another measure, which I've thought out less fully, would be to measure the Kolmogorov Complexity of the map. The descriptive length of the Maryland 3rd is... well "long" doesn't cut it. Compare that to the former definition of the Maryland 1st: "the Eastern Shore up to the Susquehanna River." The Kolmogorov Complexity itself is incomputable, but I think we have very good proxies in this case. Whoever draws the maps must be able to present in writing, with no graphics, an unambiguous definition of the districts in less than k words. Or alternately, must be able to present them orally to a judge, spoken and without visual aids, in c minutes.

There's a general principle here: If you can't explain it succinctly, it's a bad rule.

Another approach might be to measure the curvature or the fractal dimension of boundaries between districts. I've thought even less about this though.

The approach of capping some metric has the option of being simple, objective, and easily implementable. Whatever partisan committee which now draws the borders presents their map to an adjudicator, who passes it through an established, preferably open-source, program which calculates the metrics. If they're above some pre-established threshold the map is rejected. This is already an improvement over the subjective way we do things now.

A second option, suggested by David Friedman, is to have the partisan committee submit not a finished map, but a program to generate a map. The inputs to the program could contain population data, municipal boundaries, etc. but no information about party affiliation or past voting records, or any proxies for them. There would also be a cap on program length to keep anyone from submitting what is essentially just a precomputed lookup table.

Given the low numeracy I see in most legislators, I think it would be a hard sell to get them produce such a program, but the idea has tremendous appeal to me. The whole problem is to remove subjectivity and make the decision on the basis only of explicitly acceptable criteria, which is exactly what computers do: they're entirely dispassionate and the only know what you tell them.

We could take this one step further and simply codify the program to be used rather than having partisan committees writing their own after every census. I find a technocratic Bureau of Districting Algorithms much easier to imagine than local politicians wrestling with GIS algorithms.

Posted in CS / Science / Tech / Coding | Tagged , | Leave a comment

Roman Mars

The latest episode of Bullseye includes a great interview with radio producer and podcaster Roman Mars.

99percent-invisible-logoMars is the man behind the wonderful podcast 99% Invisible. (No relation to OWS.) 99% Invisible is about the design and architecture of both the extremely weird (Kowloon Walled City, Razzle Camoflage) to the entirely prosaic (broken Metro escalators, check cashing stores, culs-de-sac) to the outright awesome (The Feltron Annual Report, Trappist ales).

I had always thought Mars had a background in architecture. Turns out he actually went to grad school to study genetics. A lot of what he said about studying and learning and why he went to grad school really resonated with me, which is why I'm writing this post. (Besides wanting to evangelize 99% Invisible, which I couldn't recommend more.)

The only complaint I have with the interview came when Mars said that if you simply read a list of his podcast topics without listening to the show you'd think they were the most boring things in the world.

I couldn't disagree more. The topics he chooses are exactly the sort of thing that lead people like me to descend into hour-long Wikipedia spelunking expeditions. (Except his investigations of them have way higher production value and are told much more artfully than the Wikipedia writing-by-committee process produces.) Don't you want to learn about how Gallaudet University designed buildings suitable for the deaf? Or how audio engineers sound-design the Olympics? Yes. Yes you do.

Posted in Uncategorized | Tagged , | Leave a comment