I’ve been looking recently at the e-Democracy sites funded by the Sunlight foundation. There’s some pretty impressive stuff there — I think that MyOpenCongress is going to be a wonderful tool. I was thinking of trying to coordinate something similar for Canada.
While I was doing my research, I drew up some pictures to help me remember the value chain of the different sites:
Attended Blitzweekend this weekend, and the vibe was crackling. On a personal level, it turned out to be a bizarre confluence of things I’ve been interested in lately, including:
I know the Resipricate team got knocked at the end for not actually working on a prototype, but give these guys a break — they came all the way from Toronto and Kingston expecting a bunch of eager ruby programmers to jump on their project. Instead, they got me. After a few hours of Tas making a valiant effort to show me ruby, we decided instead to spend the weekend working on their business plan and pitch slides, and talking about the things they’re going to have to deal with when starting a company and getting angel/VC money. Trust me, not all of the outcomes of the weekend were visible in the presentation.
Speaking of government accountability, I came across a fascinating review, here, on ways of using the web to promote transparency. It’s a fun read, with useful examples from groups like the Sunlight Foundation.
Google owns the future and everything in it. I’m not sure why people don’t realize that. There Google is, like a wide-eyed toddler with a magic wand, willing vast chunks of infrastructure into existence. Seriously, these pieces: maps, local search, checkout, calendar, library, open-social, are going to come together as the platform of future consumer business. Current web 2.0 startups exist in the realm of ‘things Google hasn’t gotten around to doing yet’.
I have to laugh when people get huffy over a Google new release, like the re-tooled JotSpot. Dude, it’s the first release. Is Google out of money? No? Then I think it’s going to get better.
Of course, I have my own complaints about one of the new apps — the My Library section of Google Book Search. Interesting, and a good start, but it could have been exponentially, mind-blowingly better. This is clearly the product of a culture fixated on ’search’ rather than ’share’.
I meant to post earlier re: a book I enjoyed over the holidays: Why Most Things Fail, by Paul Omerod. This is, apparently, where Nassim Taleb got most of his ammunition for his recent hit, the Black Swan.
In the Black Swan, I was delighted to find out that the Black-Scholes model for derrivatives — which always seemed bizarre to me — is not accurate, and in fact systematically under-predicts large price changes. I was disappointed, however, as I read further and discovered that Taleb was tearing down the bell-curve model only to replace it with a different one: the scalar. Why price changes should follow a stochastic distribution at all and not be just, well, random was left as a mystery of the universe.
Omerod takes a more satisfying run at the the problem…. (Read on …)
A few days into traveling with my dad in Cuba this Christmas, Dad asked me if the Cuban dual-price system bothered me. (In Cuba, foreigners are only allowed to spend Cuban Convertible pesos. Anything bought with convertibles — restaurant meals, taxi fare, etc. — automatically costs from 10 - 20 times what a Cuban would pay.) I replied that, no, it was only formalizing something that existed anyway in India, where foreigners pay more for everything because a) they don’t know the local cost, and b) they prefer to pay more for specialized restaurants, hotels etc. that keep out the hoi polloi with their locally unfordable prices.
Perhaps my favorite result of reading the Taleb’s the Black Swan was that it dragged back to top-of-mind a literature review I had done in my MBA regarding decision making in venture capital. In broad strokes:
- VCs like to gather lots of information
- Usually, this information only serves to support initial biases
- More information does not lead to a more ‘accurate’ estimation of a company’s success… however, it does lead to increased confidence on the part of the investors
Taleb advises VCs (and individuals) to acknowledge the degree of randomness involved in prediction – and to maximize returns by making lots of small bets – to ‘expose yourself to random upsides’, in his memorable phrasing.
I think this is bang on. VCs spend too much time and energy gathering (and requiring companies to generate) useless minutia. A better system would combine a broad stroke analysis with minimum time investment – spreading investments and risk over hundreds of companies. I’m deeply and profoundly jealous of the company
Fluid Innovation (Virtual Ventures)
For coming up with a risk free way of prototyping this sort of system. Virtual Ventures is a prediction market (kisses to James Suroweicki) for fortune 500 companies to determine the value of their moldering IP. Once the concept is demonstrated in this arena, it’s only a small, inevitable step to use it in VC.
As I mentioned before, I was blown away by Nassim Taleb’s the Black Swan. I love this man for confirming what I suspected in my MBA finance classes: the stock analysis techniques taught are not, in fact, particularly useful. Chapter 10, The Scandal or Prediction talks of a paper scrutinizing 2000 predictions by security analysts:
‘What it showed was the these brokerage houses predicted nothing – a naive forecast made by someone who takes figures from one period as a predictor of the next would not do markedly worse.’ (p 150)
Very nice. Similarly, a review of papers on the predictive value of economic projections:
“collectively show no convincing evidence that economists as a community have an ability to predict, and, if they have some ability, their predictions are at best just slightly better than random ones – not good enough to help with serious decisions.” (p. 154)
Taleb advises stock analysts and economists to give up the pretense of being useful contributors to society, and to look for other jobs.
Amazon hates me. I give up a certain amount of my monthly income to this entity (a tithe, really). In return, I get mis-delivered packages, inscrutable customer support emails, and general abuse. Witness: here is a book I would like to read.
Unfortunately, the book will not be available until 2012. Note how Amazon cheerfully provides a link to pre-order a book that will not be released for five years.
Odds are that this is a data entry mistake. However, I know from past experience that the link to try to complain is about 8 clicks away. Further, the first response will be canned, and the second one will be from someone with limited English, answering an entirely different question.
We covered the Black-Scholes options formula in a portfolio management class in my second year of the MBA. We had started using the formula in my derivatives class, and I didn’t get the premise there either. Black-Scholes is a formula borrowed from physics that calculates the probability that a random variable that moves according to a bell curve distribution will have a particular value in the future. For some reason, this formula is used by stock analysts to determine option prices. I put up my hand.
‘How many stocks follow a bell curve distribution?’
My professor, who somehow always seemed three sheets to the wind, even when he wasn’t, replied:
‘None!’
The other day at lunch we were complaining about government inefficiency and bureaucracy.
It’s frustrating that as a service consumer (and funder), you have no information. About six weeks ago, I sent off my taxes, with no result. Granted, I was late in filing– they owed me money, and I wasn’t in a rush — but still, six weeks, and no information. I wonder: Where is my file right now? Where does it have to go next? What is the average throughput of that stage? At a higher level: is the system designed as efficiently as possible? How does it compare to other similar systems, other governments?
I was thinking that what the government ought to do is expose all their systems in process diagrams on the web. Like the ManyEyes project, they could take advantage of 1000’s of generally concerned (or just bored) people analyzing the data for anomalies. Why trust the opinion just one or two process analysts when you could get the opinion of hundreds — and generate better results via the Wisdom of Crowds.
For throughput statistics, you could hook your process diagram up to an RFID system that tracks individual paper files, or put some tracking points into your electronic file system. What could be easier, I thought?
A quick search on the web, and lo, it exists — the RFID part, anyway.
RFID file tracking as a pilot project in a hospital:
http://www.hin.com/sw/hospital_HSmanagement.html
The US govt has an RFID program in its tax courts:
http://www.rfidupdate.com/articles/index.php?id=1154
In 5 years, I want to know where my tax file is within a 3 meter radius.