wgyn

If you really want to do linear regression by hand, check out Chapter 1 of Stephen Stigler's History of Statistics: https://www.amazon.com/History-Statistics-Measurement-Uncert.... You can do least squares on astronomical data the way Legendre did it.

Parse.ly recently wrote a fairly insightful piece about their reasons for taking less funding than was on the table: https://blog.parse.ly/post/6282/why-we-said-no-vc-money/

Mission Bit is an organization doing this on a local scale in SF: https://www.missionbit.com/. It looks like the main difference is that Mission Bit works directly with SF public schools. Personally, it feels like there is plenty more space for non-profits reaching out to underserved communities.

Actually such a calculation is possible (of consumer surplus), though perhaps some would argue against one or another part of the methodology.

See: http://freakonomics.com/podcast/uber-economists-dream/ or http://www.nber.org/papers/w22627.

> So we found that in 2015, if you extrapolate to the whole U.S., we found that the overall consumer surplus added up to almost $7 billion. So people spent about $4 billion on Ubers, but they actually would have been willing to spend about $11 billion. So for every dollar people spent on Uber, they got about $1.50 worth of extra joy that they would have been willing to pay on average above and beyond what they did pay. And to put it in a context just how big this consumer surplus is, the total amount of money that went to what Uber calls the driver partners, the people who are driving the cars, was something like $2.5 billion. So the consumers got more than twice the benefit of all the money that went to the drivers. And Uber itself only got to keep about $1 billion.

Of course, none of this makes Uber immune to criticism for the way they operate as a company. And there are sufficiently many alternatives that, even without Uber, society is likely to continue reaping benefits from ride-sharing as a technology.

> “There is one and only one social responsibility of business,” the economist Milton Friedman famously wrote in 1962. And that is “to use its resources and engage in activities designed to increase its profits.”

It's strange to me that Friedman is viewed (at least by economists and "business" folk) as a paragon of logical reasoning and empiricism, all the while his views reek so strongly of ideology. There are tons of reasons why the quoted statement might be false whether you're a policy maker (negative externalities) or even the most cold-hearted capitalist (short-term incentives).

Perhaps relatedly, it's actually not true that corporate responsibility to profit is codified in the law. See e.g. https://en.wikipedia.org/wiki/Shlensky_v._Wrigley (discussed further in https://www.amazon.co.uk/Shareholder-Value-Myth-Shareholders...)

You're overlooking the counterfactual here. The fact that ARPANET was a Department of Defense project doesn't mean that defense spending itself caused it to come to fruition, and similarly it doesn't mean that without war (and defense spending), something like the ARPANET would never have been invented.

Separately, it seems far more significant that the early ARPANET was composed of networked computers based at universities.

Cool stuff. But it's kind of surprising that you can't link to a tagged chart, only download a raw base-64 encoded SVG.

> but at least it's 100% your own infrastructure

This is probably a good thing for security. But for the use case of development against third party services, it seems like an unnecessary constraint.

Here's an existence proof that congresspeople pay attention to calls from their constituents: https://twitter.com/RepJudyChu/status/845044275427659776.

What's perhaps striking is the low number of total calls regarding AHCA (439 support, 6 against) compared to the total population of her district (~700k). Unless I'm misinterpreting, this seems like pretty strong evidence that a call to your representative can have a massively disproportionate impact?

I think that just "networking" would be the wrong takeaway, though. The author[0] is pretty active on Twitter / the blogosphere and, even though I don't follow him personally, his writing has popped up on my feed a number of times. He's likely meeting people and having interesting conversations ("networking") because he has interesting things to say.

[0] https://twitter.com/tdhopper

The Oroville Dam is really a crazy undertaking when you consider: 1) it primarily serves to move water all the way from north of Sacramento to San Joaquin Valley / Southern California, 2) despite generating some amount of electricity, it actually consumes 3/4 of that energy just in transporting the water (over mountain ranges). More generally, the State Water Project (of which the dam is a part) is the largest electricity consumer in the state of California.

Source: https://www.amazon.com/Introduction-California-Natural-Histo... (kind of "dry", but extremely informative)

Because it signals that Uber's employees and customers have more ability to impact the company (via it's bottom line) than a presidential policy or statement.

This is what I find so silly. Companies have all this information about you that is actually relevant to their customer support. They could get so much out of making e.g. the phone support system talk to their account system.

Don't forget finagle and scalding.

At the company I work at, various of the above projects are integral to one system or another (though to be fair, there are quite a number of ex-Twitter folks floating around). I'm not a Scala expert, but as far as I can tell Twitter also has a large footprint in that community.

> My experience of ramping up in the Facebook codebase is probably pretty similar to what most new engineers here go through. I was consistently impressed by how well organized our code is, and how easy it was to find what you're looking for -- whether it's related to face recognition, speech recognition, the Messenger Bot Framework [messenger.com/platform] or iOS development. The open source Nuclide [github.com/facebook/nuclide] packages we've built to work with GitHub's Atom make development much easier. The Buck [buckbuild.com] build system we've developed to build large projects quickly also saved me a lot of time. Our open source FastText [github.com/facebookresearch/fastTex...] AI text classification tool is also a good one to check out, and if you're interested in AI development, the whole Facebook Research [github.com/facebookresearch] GitHub repo is worth taking a look at.

Face recognition and speech recognition are the most obvious pieces that look to be proprietary, but there is lots of other useful open stuff that he explicitly links. FastText looks like it would be really helpful for at least the natural language piece of Jarvis-like applications.

Brian Krebs pegged a company called XiongMai: https://krebsonsecurity.com/2016/10/hacked-cameras-dvrs-powe...

At the end of the article, Football Manager (an extremely souped-up version of FIFA's manager mode) is briefly mentioned as having influenced data analysts and scouting. I wish they had dug into it more. The connection between video games and scouting seems more direct than the one to on-field play (until we get really good VR).

That's not quite true. E.g. Google's email spam models are trained without access to the raw text. In general the field of differential privacy is developing rather quickly.

Here's a simple illustration of why diversification is considered "good" in general. If you think of your portfolio as a weighted average of returns, the variance of the portfolio decreases as you add uncorrelated assets (i.e. diversify)[0]. In the simplest case, imagine two assets with the same expected return whose returns are uncorrelated. If you don't believe in stock-picking ability, diversifying between the two is a good idea. When you try to be more sophisticated about this, you get modern portfolio theory.

[0] https://en.wikipedia.org/wiki/Variance#Sum_of_uncorrelated_v...

I find it kind of strange that there's no mention of the university in Urbana/Champaign.