I don't know what this emotion is. It is uniformly unpleasant. I say "uniformly" because I feel numbness all around my body with no rhyme or reason. Activities feel unpleasant. Writing this description for my own later reference is unpleasant. I have work to perform but concentrating feels like it would be unpleasant. I clutch my head as if I have a headache but it's completely painless. I feel twitchy but I don't think I'm actually twitching. I don't have a name for this. Maybe just "unpleasantness".

spoilers.

right, so, this was good.

EE 102 students! Tired of waiting for your synthesizer to finish routing? Check out this one weird trick to prove P = NP - and more - that the computer programmers down the hall DON’T want you to know!

We’re all used to the computer scientists whining. Oh, this problem is impossible even conceptually. The traveling salesman problem takes more time than there are ways to coat beach sand with all protons in the universe. My job is sooooo hard.

Now you can expose what we always knew were lies! This simple circuit, discovered by a tragically overworked student JUST LIKE YOU, solves the “NP complete” (ridiculous!) Boolean satisfiability problem not just in polynomial time, but INSTANTLY! Simply construct the instance out of gates, insert it into the circuit, and drive the input to logic high. Immediately, you can read off the necessary Boolean inputs below!

The simple construction, already obvious to most of you, relies only on basic principles: that the op amp does “whatever is necessary” to keep its differential input zero! This principle, already known to the designers of op amps, is reflected right in the name. An OPerational amp lets you compute any operation!

Order now and we’ll show you how to solve negated instances JUST AS EASILY!

…

Op amps are an active, nonlinear circuit component that are nonetheless often taught in introductory EE, because they are hella useful (but not this useful). Intro classes are often oriented around passive linear components - resistors, inductors, capacitors, and sometimes transformers - so very idealized models are used.

They were taught to me like this:

So they’re usually used to equate voltages, as in a basic buffer amplifier, or bla bla they’re always used with feedback. E.g., the buffer works by putting the input on one input terminal of the amp, and connecting the amp’s output to the other input; so now the output has to equal the input.

But this is obviously impossible in general, as evidenced by this stupid circuit. It can’t possibly let you invert arbitrary functions, as evidenced by this stupid drawing.

What actually happens, as far as I can discern, is that the amplifier indeed works as an amp, but before it can reach “infinite output” it passes through every voltage in between, which (if the feedback is set up correctly) includes the output that drives the feedback input appropriately.

“If the feedback is set up correctly” implies that the correct output is actually there. In this circuit, the logic high input (positive) is inverted and compared to the initially zero feedback input, so the amp output is driven ever more negative, which of course doesn’t solve a Boolean circuit.

i've now read about two space missions (EJSM/Laplace and the giant space laser triangle) that were originally proposed as NASA/ESA joint missions, but NASA dropped out for lack of budget

i'm not very gung ho on space compared to a lot of nerds but this sort of thing still makes me sad. these sort of things are great. pretty much just totally useless science. knowing gravitational waves exist or the composition of gandymede's crust isn't going to help anyone in the forseeable future. it's just kind of nice to know.

it's not even a "moonshot" since going to the moon was a fuck-you to the USSR with science tacked on, whereas nobody is politically going to care about the space laser triangle.

private space exploration is coming up in the US, but it's not going to be exploration, it's asteroid mining and cheap launch vehicles and that's all well and good but i don't think virgin galactic is going to fund boring into makemake.

bummer.

Stages of my education wrt engineering fields

1. Wow! 2. This is so elegant. I don't really understand but it's beautiful. The whole universe must be based on these principles. 3. Ugh, but the [construction] people actually use is so ugly. Why not just use beauty? 4. Beauty is hard to maintain. 5. I abhor these hacks, but I have to admire their cleverness. 6. Maybe hacks aren't so different from beauty. 7. Everything is built on hacks. 8. Well, whatever, it works. 9. Usually. Possibly there are more stages. I'm on 9 for software, 8.5 for physiology, 3 for hardware.

current mood

Is the brain an FPGA?

http://nytimes.com/2015/06/28/opinion/sunday/face-it-your-brain-is-a-computer.html I complain about brains being considered computers a lot, and I don't think this article is very convincing on that front. But the somewhat tangential comparison to FPGAs is interesting. Turing machines, or rather universal Turing machines, are programmable, but they're essentially oriented towards an input being turned into an output; that is to say, a process that continually takes inputs and continually outputs has to be somewhat massaged into the formalism. (Practically speaking, this means interrupts and stuff. Not great.) Brains, obviously, continually input and output. Now an FPGA is a more obscure thing. You don't get them at Best Buy. What an fpga is is sort of to a regular electronic circuit (synchronous digital integrated circuits, more specifically) as a universal Turing machine is to Turing machines in general. They are "Field Programmable", see. Abstractly, the obvious structure for this general circuit is a lookup table ("LUT") for logic (as a LUT can do arbitrary Boolean functions), along with a flip flop for time dependence/memory. This is actually formalized as a "Moore" or "Mealy" machine depending on details of the connections, and assuming only one clock. Early FPGAs, or CPLDs or whatever, actually resembled this structure -networks of LUTs and FFs. But FPGAs are more complicated now, cos people want to use them for specialized tasks. They have specialized blocks, like multipliers, clock managers, and a LOT of shit for I/O. I used to think of neurons similarly to the old CLPDs. Neurons are universal, sure, so just hook up a fuckload together and bam. Why bother with structure? But of course it's not really like that. You have ganglia, you have specialized cell types. Superficial

Nevermind.

on speciation through behavioral fragmentation

B.B. Fulton (1932)1 was the first to use cricket songs extensively in distinguishing species, and he found numerous instances of supposed species producing more than one calling song. For example, he recognized four songs of "Gryllus assimilis" in North Carolina and showed that populations producing the four songs had different seasonal life cycles, occurred in different (though overlapping) habitats, and were reproductively incompatible in laboratory crosses (Fulton 1952). Subsequent studies of "G. assimilis" in the eastern United States proved the existence of at least eight cryptic species - including two that have indistinguishable songs but are seasonally separated (Alexander and Bigelow 1960) and one that has no calling song (Walker 1974). (The latter two cases illustrate that acoustic studies do not identify all sibling species of crickets.) Once the species of Gryllus in the eastern United States were recognized on the basis of song or life cycle, most proved identifiable by conventional morphological features or by the number and spacing of teeth in the stridulatory file (Alexander 1957; Alexander and Walker 1962, Nickle and Walker 1974). The species had not been distinguished in earlier studies of museum specimens because distinguishing species within a group of specimens containing na unknown number of species is far more difficult than finding differences between groups of specimens already known to represent species. A similar series of events occurred with "Teleogryllus mitratus," which are Gryllus-like Japanese field crickets. Differences in calling songs led Ohmachi and Matsuura (1951)2 to distinguish three species that subsequently proved different in color pattern, ovipositor length, and karyotype.

Modern recording and analyzing devices have made it much easier to study cricket calls, and sibling species have been recognized by song in practically every group of crickets in which the acoustics have been studied - for example, field crickets in Australia (Leroy 1965)3, field crickets in Japan (Matsuura 1977, 1978a, 1978b)4, mole crickets in France (Bennet-Clark 1970a), and ground crickets and scaly crickets in the United States (Alexander and Thomas 1959, Love and Walker 19795).

Sibling species seem unusually numerous in crickets, but they may be no frequent than in other insects - just easier to detect (Walker 1964b)6. As Henry (1985) proposed, acoustic signaling to form sexual pairs may be conductive to speciation because once populations diverge in calling song they can remain reproductively isolated even if, in other respects, they are remarkably similar. However, the fact that some sibling species of crickets lack distinctive songs and have been recognized through studies of life cycles or allozymes strengthens the notion that sibling cricket species would be frequent even if crickets were mute (e.g., Alexander and Bigelow 1960, Howard 1983, Masaki 1983).

(From TJ Walker and Sinzo Masaki, "Natural History", in Cricket Behavior and Neurobiology (Huber, Moore, Loher eds.). Many of the papers cited are available on Walker's website.)

I was watching a movie, or in a movie. Simulations of things are indistinguishable from things, you see.

Fanshen

getting trashed is fun but how about something else

Fanshen: A Documentary of Revolution in a Chinese Village, by William Hinton, is good propaganda. "good" both in the sense that it is effective as propaganda, being entertaining and educational about false things, and "good" that it's a good book, entertaining and educational about true things.

It's about the Communist revolution of the 1940s in China as concerns a single village, Long Bow (张庄村, Zhāngzhuāngcūn) in the norethern part of the country. It's not so much a regular history text as it is a "documentary", as explained by the book itself; individual people are very important, they're sometimes interviewed separately, that sort of thing. It really does read like a documentary film watches, which is not something I think I've seen before.

I've tried to write out a rundown of the book four times. I don't think it's happening. So: If you want a book on the practical consequences of a Communist policy in the small scale, with numbers and charts - in China at this time specifically, or in feudal societies, or whatever - this is a good pick. Or if you like seeing the concrete policy statements of Communists, which are mysteriously hard to find in my American upbringing.