Links for January 2017

Wanderers – a short film by Erik Wernquist from Erik Wernquist on Vimeo.

…refs make more defensive pass interference calls on the offensive team’s sideline but more offensive holding calls on the defensive team’s sideline. What’s more, these differences aren’t uniform across the field — the effect only shows up on plays run, roughly, between the 32-yard lines, the same space where coaches and players are allowed to stand during play.

I do worry about publication bias, since they only gave charts for a few classes of penalties. It’s possible they simply selected those that behaved like they expected. (H/t Will Riedel.)

it should be clear by now that with more than half the kids in America already going to college, academic expansion is finished forever.

The average American professor in a research university turns out about 15 Ph.D students in the course of a career. In a stable, steady-state world of science, only one of those 15 can go on to become another professor in a research university.

I think this gives you license to ignore advice on getting a permanent job from anyone who landed one before about 1975…

The evolution of mammalian auditory ossicles is one of the most well-documented and important evolutionary events, demonstrating both numerous transitional forms as well as an excellent example of exaptation, the re-purposing of existing structures during evolution.

In reptiles, the eardrum is connected to the inner ear via a single bone, the columella, while the upper and lower jaws contain several bones not found in mammals. Over the course of the evolution of mammals, one lower and one upper jaw bone (the articular and quadrate) lost their purpose in the jaw joint and were put to new use in the middle ear, connecting to the existing stapes bone and forming a chain of three bones (collectively called the ossicles) which transmit sounds more efficiently and allow more acute hearing.

If you take this set of claims seriously…then the most fundamental problem that the Internet poses is not one of network advantage, increasing returns to scale and so on advantaging big players (since, with a non-supine anti-trust authority, these could in principle be addressed). It’s the problem of how radically cheaper communication makes new forms of implicit and explicit collusion possible at scale, squeezing consumers.

H/t Tyler Cowen. Also from him: Declassified CIA maps.

“It’s going to be massively important,” says Karl Iagnemma, co-founder and CEO of self-driving startup nuTonomy, which is developing a remote control system. Even cars that can handle just about anything will have the occasional failure, even if that’s being hit by another vehicle. And in that case, you want a human around to decide what to do. It’s like an elevator, Iagnemma says: You don’t need a human operator, but you’ve still got a button to call for help when you need it.

Google’s self-driving car outfit, Waymo, has studied the idea, a spokesperson says. Uber declined to comment on teleoperation, but in 2015 the company filed a patent for a system that would let an autonomous vehicle follow a human-driven car, or get help from a remote operator. Stealthy self-driving car startup Zoox has a patent for a “teleoperation system and method for trajectory modification of autonomous vehicles;” Toyota has one for “remote operation of autonomous vehicle in unexpected environment.”

Now, Nissan’s cubicle-based drivers aren’t emergency backups. If the car hits black ice, it’s in charge of staying on the road. There’s no feasible way to get the human into the loop in time to act. But they can help out when the car encounters conditions it’s unsure how to handle. If a Nissan happened upon the construction scene from Sierhuis’ photo, it would stop and ping its control center. A human operator would look around using the car’s cameras and other sensors and issue new instructions—direct control would pose latency issues. Like: When it’s safe, cross the double yellow and get back to the right side after 20 yards. Or a new instruction set could ensure packages and disabled passengers get dropped off in exactly the right spot, and help assess potentially dangerous situations on the road. But most of all, the teleoperator is there to make sure the car’s doesn’t just shut down when it’s too dumb to know what’s going on.

Relatedly, here is VC Frank Chen discussing 16 open questions driving the development of self-driving cars. At the end is a summary of predictions for when these will appear. And Nvidia and Audi claim Level 4 Autonomy by 2020. Lots of answers between 2020 and 2023, but I think correcting for over optimism makes me still feel OK about my bet with Paul.

The Columbia-class submarine…is a future United States Navy nuclear submarine class designed to replace the Trident missile-armed Ohio-class ballistic missile submarines. The first submarine is scheduled to begin construction in 2021 and enter service in 2031 (some 50 years after its immediate predecessor, the Ohio class, entered service). From there, the submarine class will serve through 2085.

A fleet of 12 was recently approved by the Pentagon.

Rail is 4 times more expensive than by sea. Air is at least 6 times more expensive than rail, but can be even 18 times more expensive if payload is heavy.

Cost of shipping 40′ HC container:

• by sea: $1,600
• by train: $6,800
• by air, light, so you pay for volume: 4.6 USD / kg, 40 ft got volume 67.7 m3 * 143 kg / m3 ~= 9 tons, $41,400
• by air, heavy, so you pay for real weight: 4.6 USD / kg, 40tf can be 27tons, $124,200

It’s unclear whether such devices are a net improvement. Interestingly, it’s also unclear whether or not people experiencing a cardiac arrest outside of a hospital should be immediately transported.

But for Dr Jonathan Benger (University of the West of England, Bristol, UK), the strategy of taking these sick patients to the hospital made sense when the hospital was the only place that had defibrillators.

“However, this is no longer the case, and hospitals have nothing to offer almost all such patients beyond the care that is provided by a well-trained and equipped ambulance service,” writes Benger. “Preparing patients for transport, moving them, and driving them to the hospital leads to pauses in CPR and suboptimal chest compressions, even with the most skilled and committed staff.”

In this essay I argue that heavy reliance on a small number of metrics is bad for science. Of course, many people have previously criticised metrics such as citation count or the h-index. Such criticisms tend to fall into one of two categories. In the first category are criticisms of the properties of particular metrics, for example, that they undervalue pioneer work, or that they unfairly disadvantage particular fields. In the second category are criticisms of the entire notion of quantitatively measuring science. My argument differs from both these types of arguments. I accept that metrics in some form are inevitable – after all, as I said above, every granting or hiring committee is effectively using a metric every time they make a decision. My argument instead is essentially an argument against homogeneity in the evaluation of science: it’s not the use of metrics I’m objecting to, per se, rather it’s the idea that a relatively small number of metrics may become broadly influential. I shall argue that it’s much better if the system is very diverse, with all sorts of different ways being used to evaluate science. Crucially, my argument is independent of the details of what metrics are being broadly adopted: no matter how well-designed a particular metric may be, we shall see that it would be better to use a more heterogeneous system.

See also CognitiveMedium.com for his more recent thoughts on other topics.