Here’s links with commentary from recent reading.
1. Notes on the College admissions scandal. The New York Times has a good explainer. The primary ways people cheated on college admissions where: 1) paying someone else to take SAT tests, and 2) bribing coaches to fake sports resumes. The context here is elite colleges face unrelenting pressure to raise their rankings. In practice this makes them lure in max applicants but only accept a flat cap, because lower acceptance rates equate to greater eliteness. But somewhat surprisingly, this pressure also leads to easy grading, with students rarely flunking out. Why? Because if particular demographics of students flunk out, people will suspect unqualified students must have snuck in. This dynamic makes cheating to get into an elite school work. Because once you sneak in, the university doesn’t want you to flunk out, especially if you’re underqualified. Everyone needs to get an A.
In terms of commentary, Arnold Kling argues the “fundamental scandal is that elite colleges are a positional good for parents. The whole process is built around that.” That is to say, kids weren’t cheating. Parents were cheating so they could brag to their friends. Often not even telling their kids. Thus the people most outraged are rich parents with elite degrees, who have been grinding their kids down playing by the rules. Kevin Carey points out that statistically elite college degrees tend not to matter very much for the wealthy or those with high test scores, as they’ll get fine salaries either way. It makes the biggest difference for the underprivileged, as an entree into upper class culture, networking and jobs. This is true. But downplays the social status allure of the elite college credential, which drives of the problem. Parents cheat for status, not money.
Ross Douthat say many “misunderstand the real function of meritocracy, which is not the facilitation of upward mobility but the legitimation of a ruling class.” That is to say, Harvard’s purpose is not merely to select applicants, but to justify the legitimacy of those who govern. What we’re seeing now in US politics is a breakdown of trust in leadership of our institutions. This breakdown encourages a nihilistic populism.
One proposed response to the scandal is to eliminate test scores on applications. Daniel Friedman walks through why this will makes things worse. The easy way to see this is by asking whether it would be easier or harder for wealthy parents to rig the system if they didn’t have to provide test scores. One less thing to lie about.
Having read a lot of commentary, I suspect nothing will change. The dynamics behind it have been growing for decades, with no end in sight. So if you want to relax bit about getting your kids into an elite college, I’d suggest Bryan Caplan’s book. 🙂
2. Waymo’s self driving blunder. An excellent piece by Timothy B Lee on how Google Waymo is repeating one of silicon valley’s most classic mistakes, Xerox PARC and PCs. In the 1970s, Xerox PARC famously created the Alto computer, with a mouse, copy/paste, windowing, and all the other things that would start the microcomputer revolution. But they failed to commercialize it. Why? Because the Xerox PARC “workstation started at $16,595—more than $45,000 in 2019 dollars.” Instead Apple released the Apple I for $666, the Apple II for $1298. As a side note, I got an Apple II+ as my first computer in high school. It was great! But felt super expensive at the time. Apple did not release a personal computer with all the Alto features until the Lisa in 1983, a flop at $9995. But then the first Mac in 1984 for $2495 was a hit. Lee argues that Google Waymo is making the same mistake by trying to jump directly to fully self driving cars (jumping to the Mac instead of releasing the Apple I). Waymo should go to market with a lessor solution and iterate. If you are into technology and disruption theory, a great read. link
3. Level 2 self driving is pretty silly. On a related note, the levels getting to self driving cars are defined as: level 0 – manual, level 1 – cruise control, level 2 – partial automation (hands on, may have to take back control), level 3 – eyes off (car can stop if needs help), level 4 – full (doesn’t need help), level 5 – no steering wheel. This is ok except for level 2, which is silly. Why? Because with level 2 you’re supposed to let the car drive but still pay attention and keep your hands on the wheel at all times. But human natures being what it is, when it was tried in 2012 by Google “In-car cameras showed users napping, putting on makeup and fiddling with their phones.” Pretty much. Level 2 shouldn’t exist.
4. Survey of political divide in US. Unsurprising voter survey results, but still makes me sad:
How about: “Do you ever think: ‘we’d be better off as a country if large numbers of the opposing party in the public today just died’?”
Some 20 percent of Democrats (that translates to 12.6 million voters) and 16 percent of Republicans (or 7.9 million voters) do think on occasion that the country would be better off if large numbers of the opposition died.
We’re not finished: “What if the opposing party wins the 2020 presidential election. How much do you feel violence would be justified then?” 18.3 percent of Democrats and 13.8 percent of Republicans said violence would be justified on a scale ranging from “a little” to “a lot.”
Since we’re doing voter surveys. Maybe part of this is the demographics of age. See below. If age is the driver, things will get worse. From America’s Defining Divide Isn’t Left vs. Right. It’s Old vs. Young.
5. New paper on Wigner’s friend quantum mechanics. Ok. If you’re not into foundations of quantum mechanics, skip this and jump to #6 below about Zebra stripes and flies! But if curious, plow ahead.
Wigner’s friend is a thought experiment in quantum mechanics. Observer a measures the outcome of a quantum experiment, which collapses the wave function, so observer a records a particular result. Now. If there’s an observer b who watches observer a, what observer b sees is the quantum superposition from the experiment is contagious, in the sense that observer b sees observer a is still in superposition. At least until after observer b does their measurement. This is related to the famous Schrödinger’s cat thought experiment.
At a technical level the contradiction between observer a and b manifests as a violation of the Bell inequality. The experiment in the paper did an extended version of Wigner’s friend, with an outside observer c.
From the abstract “In a state-of-the-art 6-photon experiment, we here realise this extended Wigner’s friend scenario, experimentally violating the associated Bell-type inequality by 5 standard deviations. This result lends considerable strength to interpretations of quantum theory already set in an observer-dependent framework and demands for revision of those which are not.” The last sentence says the observer is part of the quantum system. So the naive Copenhagen interpretation of wave collapse is no good. This paper won’t change any minds. But to me is another small push towards many worlds being reasonable. Your mileage may vary. link to paper
6. Zebra stripes discourage biting flies. One longstanding hypothesis (among many) for why Zebras have stripes is to discourage flies. A new paper supports this idea. Apparently flies can see the Zebras, but the stripe pattern makes them struggle to land. Excellent write up by Ed Yong. Plus lovely pic below, dressing up a horse with stripes to test the theory.
And that’s all for now. Thanks for reading.
The copenhagen interpretation works fine with the wigner experiment. You have to remember the wave state can’t collapse unless you have an infinitely large apparatus at an infinite distance from the experiment. Modern physics though is almost always that sort of approximation we measure things very far away with very large apparatus. Now if you use the Wigner experiment it is using finite sized measuring devices, and thus you get these weird effects where to one “observer” the wave state collapses but to the other “observer” it doesn’t. What is really happening is you have to model the “observer” using a quantum mechanical model and then the copenhagen interpretation works fine. There’s a lecture by Suskin that describes this sort of thing, where you really need to do a full model of both the experiment and the observer, and then to observe the result you need an observer of the experiment and observer so on. the reason the shortcut of the copenhagen interpretation works at all most of the time is that we are so incredibly large compared to the qauntum effects we measure that we can get away with this approximation of wave state collapse, but that’s all it is an approximation, since we are not truly infinite.
One more thing. There’s often a misconception about Schrodinger’s Cat. There’s a thought that the Copenhagen interpretation demands a dead/alive cat. However that’s not true. All the air molecules in the box are “observers” all the cells of the cat are “observers” so the cat will collapse into either dead or alive very quickly. You can have a real Schrodinger’s Cat only if the “Cat” is a single atom isolated in a vaccum box. Then it does work and you have a dead/alive atom.
I’ll put there here since why not here. Consider the schrodginger’s cat from a multiverse perspective. The random event goes off, then the universe splits into two. One with an alive cat one with a dead cat. Now just think about that. Isn’t that odd? Shouldn’t it be the random event happens, the atoms of the device hit each other one after the other, each atom entangles with the next, each quantum mechanical atom and electron interacting with the next, due to quantum decoherence, the micro events of quantum mechanics eventually create a macroscopic event, and by the time the poison has reached the cat, so many atoms have been involved that the wave state has collapsed. The cat is not dead/alive because the system of elements has become massive enough that you’ve crossed the boundary where a macroscopic approximation of wave state collapse is possible. How simple the multi universe idea is. You have this dead cat in one multi verse. You have an alive cat in another multi verse. Instead of this incredibly complicated wave state where all the quantum mechanical effects of every atom of air, every atom of the poison every atom of every cell of the cat, all has a quantum mechanical state involved, and there’s never a chance of a cat being both dead in one universe and alive in another. That’s the problem with the multiverse model it simplifies how quantum decoherence works in to a model that’s easy for us to understand, instead of having to work out every single interaction. And of course the multiverse can’t do probabilities, it only works on 50% quantum effect, there’s a lecture by Sean Carrol where he admits that the multiverse model can’t do anything except splitting universes into two, so if there’s a quantum mechanical effect with 33% odds, the multiverse has no idea what happens. I’m not saying the multiverse is wrong, just that you can’t use it to simplify quantum effects so that our brains find them easier than to have to work out each interaction and figure out quantum decoherence. In the real multiverse each atom touching the poison of schrodinger’s cat spawns a new universe, so there’s not just one dead cat and one alive cat there are infinite alive cats, infinite dead cats and not only that there are infinite number of alive cats exactly alike and infinite number of dead cats exactly alike, so how do you calculate the probability that the universe chooses any particular path?
oh I should be more clear about calculating the probabilities of the universes and path that our universe takes. Of course you can use quantum theory to calculate the probabilities of any event happening in our universe. What I mean is if you want to use the multiuniverse model to calculate antropogenic events, you know like how likely it is to have the higgs scalar the value that it is, well you look at all the multiverses and then calculate the probability that we are in our current universe. I don’t see how you can use the antropogenic effect and theory of multiverses to calculate anything because of the infinities involved in the multiverse.