I just returned from PyCon, and just like last year it was an inspiring and motivating event. The sheer size of the conference, with its three thousand plus attendees, paired with the ever expanding surface area of all the fields where Python is used means that you’re constantly overwhelmed. Continue reading...
When Max and Einat got married, we decided that instead of giving them a physical thing as a gift, we’d give them an experience that they could enjoy with us. This trip to NYC was it. Continue reading...
Coordination problems are fascinating, and this review of Posner and Weyl’s new book gives a taste of their ideas on political economy, with a nice dose of crypto. From immigration, to private property, to the gig economy, the authors seem to have controversial but well founded solutions on how to tackle some of the hardest coordination problems facing society today. Game theory and technology are a great pairing, and one that I don’t think has been explored enough. I can’t wait for this EconTalk episode to come out.
Having read Seeing Like a State earlier this year, this article resonated a lot with me. It takes James C. Scott’s message of being wary of authoratitave states bearing technology and good intentions, and applies it to Silicon Valley companies’ good intentions. Some technology companies are governments in their own way, and many of these web services are larger and more powerful than many nation states, allowing them to deploy large projects that affect the lives of millions of people. Unlike sovereign states, which in theory have aligned incentives with their citizens, companies have a series of other constituents with their own interests beyond that of their users. Our lives are starting to be permeated by these services, and, much like nation states, their grip is starting to become inevitable as more and more aspects of our lives are mediated by them.
Every business decision is at its core an exchange of time for money and vice versa. This is not just the case of a founder picking a VC, or investors picking startups, but also of employees deciding where to work, and average Joe thinking of where to put their savings.
Gotta love when economic analysis beats bad arguments.
There were a bunch of things I love reading about here. There’s history, there’s complexity science (that’s how I got into CS in the first place), there’s python, there’s UX, and more. If you’re interested in the future of education and the spreading of scientific knowledge this is a must read.
Most traditional economic models are based on bad assumptions: full information, rational agents, etc. Dillow uses the Modigliani-Miller theorem as an example of how this quickly breaks down, and uses it to argue for better undergad education. I mostly agree that there should be a higher emphasis on empiricism.
And speaking of perfect information and economic assumptions that make no sense, here’s one that’s ususally ignored altogether at the undergraduate level: symmetric information. This post discusses the problems of infrequent transactions (think a company going public, or a person buying a house) and how one side of this market has a clear advantage over the other.
This is a wonderful post about how computer science concepts are abstracted over again and again, making it much easier to build things without knowing the details of what happens under the hood. It made me think of Alfred North Whitehead’s quip: “Civilization advances by extending the number of important operations which we can perform without thinking about them.” It also reminded me of an essay about constraints in programming and how over time we’ve decreased the surface of what a programmers can do, enabling them to do what they actually want to do more easily.
Everyone has been talking about Hanson for other reasons recently. Here he makes an interesting argument about modern celebrities and human nature.
If you believe in specialization, you should believe Krugman’s argument on trade wars. He basically says there’s two reasons for the stock market’s current drop re the potential trade war: 1) Efficiency loss imposed by artificial costs on international trade, which undo specialization (tariffs). 2) Stranded assets, which are worth more on paper than on the market. I’d say there’s a third thing in play here: behavioral overshooting. In any case, the stranded asset thing is interesting. It isn’t just the soon-to-be unproductive factories in China that lost value, but also their counterparts in the US, like real estate owned by foreigners.
Sadly, Szabo has not yet released the second part of this series yet. He argues that we’ve forgotten that money used to exist outside of the State’s purview, and it is implied in his argument that cryptocurrencies are just a reversion to older models of money without state intervention.
Obviously Gurley is biased about this company, but he makes some really great points about the gig economy, and Uber’s role in kickstarting it. If there’s a single thing that we should thank Kalanick and the company he’s built for, it’s the availabilty of flexible work options.
EconTalk has been killing it lately with their interviews. I usually don’t agree with most conservative arguments, but Goldberg makes some really interesting points in this episode. Especially interesting was his point about celebrity (related to Hanson’s from above) and how tribal thinking has overtaken the republican party.
An hour long conversation of what essentially is Goodhart’s law. Clearly the focus is on policy and social cooperation, but a lot of this conversation can be applied to software engineering and entrepreneurship.
A great conversation on the economic tradeoffs of healthcare. I recently discussed this with a good friend who’s dad is going through sever medical issues and is in need of a transplant. It is easy to discuss these things in the abstract, but when you put a face to the problem then discussing whether society should cover the cost of medication that could extend someone’s life for another year or two gets way harder. It’s a tough conversation, but one worth listening to.
I have never understood the American model of car sales.
This interview with Ray Dalio was great. His management philosophies are very controversial, but his success running Bridgewater speaks for itself. Describing the reasoning behind his book, Dalio discusses how he tries to boil everything down to pattern matching. Since most things have some similar precedent in the past, you can look at previous instances and act according to what history tells you will be the successful decision. Essentially, it’s real life Duck Typing. If you define ahead of time what your principles are, and how you’d behave in a certain situation based on your knowledge of previous experiences, then nothing is surprising. Importantly, you have to be a student of history for this strategy to work.
This conversation tries to showcase philosophies and practices from the software engineering world to a less technical audience, but does end up going in the weeds. The main idea behind dev-ops is one I subscribe to: empowering engineers to not only be in charge of designing and building a new feature, but also of deploying it, measuring its reliability, and owning its delivery throughout its life. By virtue of having one group of people responsible for the whole life cycle, there is an alignment of incentives that leads to more reliable, easier to deploy software.
Dick Costolo is an unusual guy for Silicon Valley. This interview gives a window into his background in theatre and improv comedy, and how those experiences changed his management style for the better. About halfway through, he made a great remark: “As a leader, it’s not your job to prevent mistakes from happening, it’s your job to correct them when they happen.” I could not agree more. When you’re trying to prevent errors, everything slows down, and innovation stops. It’s the opposite of what you want.
Teaching requires empathy in order to figure out how to transfer knowledge that has already been synthesized in their teacher’s head to the student, who has a very different set of anchor knowledge than whatever the teacher already does. To put themselves in the shoes of the student, and make a new idea accessible to them in known terms is what makes teaching hard. Good museums, movies, and books, are built by people who recognize that fact. The first act, where a group of black kids are innocently brought to the movies to watch Schindler’s list was fascinating, precisely because the protagonist teacher had not considered that those kids would have no anchor to understand the movie.
Ah, the joy of Python
Here I am, keeping my promise, and posting less links, more often:
This is definitely targeted at someone, but I can’t figure out who. Over the years I’ve realized that VCs, are not really gatekeepers, and they follow the same patterns of competition as everyone else. This post by Walk is an example of that dynamic. I like the idea of the VC portfolio page as a lagging indicator of what the VC wants to invest in.
A thousand times yes. The average citizen knows nothing, is easily manipulated, and does not consider second and third order consequences. The 95 percentile citizen is not that different either. We outsource policy decisions and let politicians take care of things not because people are dumb, but because understanding the nuances of policy takes time and effort. Modern society believes in specialization and, by extension, if we let politicians be politicians while dentists take care of teeth and farmers take care of farms, everyone is better off.
Krugman discusses how running a deficit vs. a surplus is about math and accounting and not about winning or losing, but just plain accounting. Then he goes on to compare international trade to inter-state trade. Inter-state and inter-city interactions are super interesting. I’ve always wondered why there isn’t more research along these lines. Made me wonder whether there is a dataset that could be used to replicate the methodology of Dirk Brockmann’s research on mobility and effective communities, but capturing trade instead of commerce.
An extension of Albert’s well known idea of bot representation. For most people “bot representation” is too abstract, but his rephrasing to “any system with 1 million+ users should by law be required to issue users with personal API keys” is quite clear. Brings up questions about the boundary of one user’s data and the next user’s data (ie, MY post on YOUR profile), what constitutes enough APIs to meet said standard, and who gets to decide these terms. These systems do things on many layers, not all available via UI. They’re different per user and change over time. The data that the API consumes and produces today looks different than it will tomorrow. The software engineering side would be a nightmare. But that can be solved with incentives. The more interesting question is whether we get access to the derived data, or just the raw. Am I only allowed access to my data points, or also the aggregates computed over time and in relation to other users? It’d be awesome to see something like this implemented.
A discussion of how we consume information, and the cyclical nature of how we share things online. Evans’ comparison of the number of items in your news feed to the number of invitees to a party is a clever analogy (kinda like Big O!) but the essay is fraught with uncharachteristically bad assumptions. Stories are less so about “units of content” than about market segmentation. Stories are a news feed, even if in a different format, and don’t solve the oversharing problem - you still get N*M pieces of content.
This problem has existed for ages. Technologists and scientists work on things because they are cool, and because they want to push science forward, regardless of the implications. The debate about whether technology is inherently good or not will keep going.
When people in 2018 think of money, they think of it much like we think of the nation state - this is how things work in the world, we are all citizens of some country, and that country issues its state-sponsored currency. Most people don’t consider that history has followed different paths, and that we’re not at the end of history. Within our lifetimes these institutions will probably shift shapes.
I still use RSS daily. It aggregates the tiny blogs that produce content at a very slow rate, which I have no motive for visiting between posts, as well as socially ranked content from sites like Hacker News. The volume that I consume is relatively high, but the top of the funnel is limited on purpose, in ways that maybe it shouldn’t be. Going back to Evans’ post above, I think that having something with levers and knobs would be amazing, but I know I’m part of the fringe. The closest I’ve gotten is RSS.
This is one of those topics that confused me for a long time about Python. Ned’s post is more of a description of the lay of the land than an explanation of how and why things work the way they do in Python, but it has inspired me to go and dig through those bytecode talks he linked to, and learn more about Python internals.
If you went to Northwestern, Al’s was a staple. I walked in front of the shop daily, on my walk from home to school, and I probably had their sandwiches at least once a week while I lived in Evanston. This essay made me nostalgic about college life, and now I’m craving their croissant with baked ham and swiss.
An amazing procedural generation piece. The fact that it is both 3D, and animated, makes it all the more interesting.
I have been making an effort to carry my camera with me whenever I take public transit to work in the morning. I usually bike, but a couple of weeks ago, my bike got stolen. The result? An upset Avy. The other result? More city photos from downtown San Francisco. More...