Here is an audio recording synchronized to slides of my talk a week and a half ago in Pittsburgh. I noticed some places where I said the wrong thing such as conflating neuron with synapse. I also did not explain the learning part very well. I should point out that we are not applying a control to the network. We train a set of weights so that given some initial condition, the neuron firing rates follow a specified target pattern. I also made a joke that implied that the Recursive Least Squares algorithm dates to 1972. That is not correct. It goes back much further back than that. I also take a pot shot at physicists. It was meant as a joke of course and describes many of my own papers.
I gave an invited plenary talk at the 2017 Society of Applied and Industrial Mathematics Annual Meeting in Pittsburgh yesterday. My slides are here. I talked about some very new work on chaos and learning in spiking neural networks. My fellow Chris Kim and I were producing graphs up to a half hour before my talk! I’m quite excited about this work and I hope to get it published soon.
During my talk, I made an offhand threat that my current Mac would be the last one I buy. I made the joke because it was the first time that I could not connect to a projector with my laptop since I started using Mac almost 20 years ago. I switched to Mac from Linux back then because it was a Unix environment where I didn’t need to be a systems administrator to print and project. However, Linux has made major headway in the past two decades while Mac is backsliding. So, I’m seriously thinking of following through. I’ve been slowly getting disenchanted with Apple products over the past three years but I am especially disappointed with my new MacBook Pro. I have the one with the silly touch screen bar. The first thing the peeves me is that the activate Siri key is right next to the delete key so I accidentally hit and then have to reject Siri every five minutes. What mostly ties me to Mac right now is the Keynote presentation software, which I like(d) because it is easy to embed formulas and PDF files into. It is much harder to do the same in PowerPoint and I haven’t found an open source version that is as easy to use. However, Keynote keeps hanging on my new machine. I also get this situation where my embedded equations will randomly disappear and then reappear. Luckily I did a quick run through just before my talk and noticed that the vanished equations reappeared and I could delete them. Thus, the Keynote appeal has definitely diminished. Now, if someone would like to start an open source Keynote project with me… Finally, the new Mac does not seem any faster than my old Mac (it still takes forever to boot up) and Bard Ermentrout told me that his dynamical systems software tool XPP runs five times slower. So, any suggestions for a new machine?
Here is a 1975 performance of Chopin’s Piano Concerto No. 2 by Artur Rubenstein and the London Symphony Orchestra with Andre Previn conducting.
When I was a child, the toy Lego was the ultimate creative experience. There were basically 4 or so different kinds of blocks from which you could connect together into whatever you could think of. Now, most Lego sets consists of a fixed number of pieces that are to be assembled into a specific object, like a fire truck. Instead of just making whatever you can think of, you now must precisely follow an instruction book with the major anxiety that a crucial piece will be missing. Creativity has been replaced by the ability to follow detailed instructions. Maybe, this makes kids look for creative outlets elsewhere, like pretend play. Maybe, they become more compliant employees. Or maybe, the world is just becoming less a imaginative and duller place.
The conventional wisdom in economics is that trade is mutually beneficial to all parties and the freer the trade the better. However, as David Autor and collaborators have empirically shown, the benefits of trade can be unevenly distributed. A simple way to think about this is to consider a simple model of a nation’s income () as a function of socio-economic status (), . Here, can be distributed in anyway but has zero mean. The mean income of the nation is while is a measure of inequality (i.e. proportional to standard deviation). Generally, it was presumed that trade increases . However, as Autor finds, trade can also increase and then it becomes a quantitative game as to whether you personally will do better or worse with trade. Your change in income will be . Thus, if you are above the mean then trade is always beneficial and increasing helps you even more. However, where the mean is with respect to the median is strongly dependent on the tails of the distribution of . So if people with high are very far away from the median, then the mean could also be high with respect to the median. If you are below the mean then gains from are offset by decreases in and if you’re is more negative than then you will do worse in absolute terms. This could explain what has been happening in the US. The nation benefits from trade by having cheaper goods but some sectors like manufacturing and textiles are greatly hurt and the cheaper goods cannot make up for the decrease in income. Those above the mean are benefitting from a mean shift in income due to trade as well as any increases in inequality. Those below the mean are getting smaller gains and in some cases doing worse as a result of trade. Thus, it may not be surprising that there are divergent views on the benefits of trade.
Near the end of the twentieth century, there was a battle between small bookstores and the big chains like Barnes and Noble and Borders, typified in the film You’ve Got Mail. The chains won because they had lower prices, larger stocks, and served as mini-community centers where people liked to hang out. It was sad to see the independent bookstores die but the replacement was actually a nice addition to the neighborhood. The Barnes and Noble business model was to create attractive places to spend time, with play areas for children, a cafe with ample seating, and racks and racks of magazines. The idea was that the more time you spent there the more money you would spend and it worked for at least ten years. Yet, at the height of their dominance, the seeds of their destruction could be plainly seen. Amazon was growing even faster and a new shopping model was invented. People would spend time and browse in B and N and then go home to order the books on Amazon. The advent of the smartphone only quickened the demise because people could order directly from the store. The large and welcoming B and N store was a free sample service for Amazon. Borders is already gone and Barnes and Noble is on its last legs. The one I frequent will be closing this summer.
The loss of B and N will be a blow to many communities. It’s a particular favorite locale for retirees to congregate. I think this is a perfect example of a market failure. There is a clear demand for the product but no viable way to monetize it. However, there already is a model for providing the same service as B and N that has worked for a century and that is called a library. Libraries are still extremely popular and provide essential services to people, and particularly low income people. The Enoch Pratt Free Library in Baltimore has a line every morning before it opens for people scrambling to use the computers and access the internet. While libraries have been rapidly modernizing, with a relaxation of behavior rules and adding cafes, they still have short hours and do not provide the comforting atmosphere of B and N.
I see multiple paths forward. The first is that B and N goes under and maybe someone invents a new private model to replace it. Amazon may create book stores in its place that act more like showrooms for their products rather than profit making entities. The second is that a philanthropist will buy it and endow it as a nonprofit entity for the community much like Carnegie and other robber barons of the nineteenth century did with libraries. The third is that communities will start to take over the spaces and create a new type of library that is subsidized by tax payers and has the same hours and ambience of B and N.
In any introductory economics class, one is introduced to the concept of supply and demand. The price of a product is expressed as a function of the number of products that suppliers would produce and buyers would purchase at that price, respectively. Supply curves have positive slope, meaning that the higher the price the more suppliers will produce and vice versa for demand curves. If a market is perfectly competitive, then the supply curve is determined by the marginal cost of production, which is the incremental cost of production for making one additional unit. Firms will keep producing more goods until the price falls below the marginal cost.
Increases in productivity lead to decreases in marginal cost, and since the advent of the industrial revolution, technology has been increasing productivity. In some cases, like software or recorded music, the marginal cost is already zero. The cost for Microsoft to make one more copy of Office is miniscule. However, if the marginal cost is zero then according to classical microeconomic theory firms would produce goods and give it away for free. Public intellectual Jeremy Rifkin has been writing about a zero marginal cost society for several years now, (e.g. see here and here), and has proposed that ubiquitous zero marginal cost will lead to a communitarian revolution where capitalism is overturned and people will collaborate and share goods along the lines of the open software model, which has produced the likes of Wikipedia, Linux, Python, and Julia.
I’m not so sanguine. There are two rational strategies for firms to pursue to increase profit. The first is to lower costs and the second is to create monopolies. In completely unregulated markets, like drug trafficking, it seems like suppliers spend much of their time and efforts pursuing monopolies by literally killing their competition. In the absence of the violence option, firms can gain monopolies by buying or merging with competitors and through regulatory capture to create barriers to entry. There are also industries where size and success create virtual monopolies. This is what happens for tech companies where a single behemoth like Microsoft, Google, Facebook, or Amazon, completely dominates a domain. Being large has a huge advantage in finance and banking. Entertainment seems to breed random monopoly status where a single artist will garner most of the attention even though objectively there may not be much difference between the top and the 100th best selling artist. As costs continue to decrease, there will be even more incentive to create monopolies. Instead of a sharing collaborative egalitarian world, a more likely scenario is a world with a small number of entrenched monopolists controlling most of the wealth.