dfan

You do lose all the strategic decisions involving the doubling cube, though.

I'm not sure what's going on with your computer (or with mine), but the lowercase t's (which I assume is what you are noticing) look fine on in my Firefox (75.0 on macos 10.14.5).

I took two software engineering courses at MIT around 1990 that used CLU as the implementation language (one was a general medium-scale software engineering course, one was on writing compilers). I found it to be a very pleasant language, although the main thing I had to compare it to was C (C++ was just getting off the ground). It's always nice to see it mentioned in lists like these.

I'm not aware of a single "official" name for it that everyone uses, like the Hypermodern style of chess in the early 20th century. In English, people say things "AI-inspired" or "AlphaGo style" (although a lot of the ideas come not from AlphaGo directly but from the public engines that followed in its wake).

I admit that I do not have a quantitative measure to support my claim (as you note, constructing one is difficult). But qualitatively:

1) People have learned a lot from new engines: joseki (corner patterns), general strategy (e.g., moves on the side are now considered less valuable, making large moyos (largely empty space loosely surrounded by your stones) is less attractive because AIs have demonstrated that they're more invadable than previously thought) and are able to actually explain the new principles in human terms.

2) All go professionals now play in the new style, to some degree; ones who tried to continue in the old pre-AI style performed badly.

So I am comfortable claiming that human play has improved by learning from the new engines.

I am not sure what you mean by "style", but top chess grandmasters certainly play in a different way today compared to, say, 20 years ago. Current play is much more concrete (less based on general strategic principles) and players are willing to take greater risks for rewards such as material gains, since AIs have shown that there are often many more defensive resources than was previously thought as long as the defender remains tenacious.

No, it isn't. The top humans are much better at Go than they were four years ago, largely due to learning from the new engines. If it were all just about sampling the phase space zillions of times, this would not be the case.

He is a smart and thoughtful person, and any given article by him is worth reading (as long as you balance it with other people's takes as well), but he has quite strong opinions on the subject, which is one reason he's always quoted. The main danger in my opinion is that it's not hard to read five separate articles about how AI is overhyped and think you are getting five independent data points from domain experts, when in fact you are really getting the same data point five times.

I assume from the title that this is either written by Gary Marcus or quotes him extensively. (I just checked and was correct.) It would be nice to hear this sort of critique from more than one person.

You are correct.

The extent of the Rock Band / Guitar Hero connection is that they were made by a company (Harmonix Music Systems) founded by two people who had been grad students at the Media Lab. They did indeed work on music tech while at the Media Lab, and the first Harmonix product, The Axe, was an outgrowth of that tech, although it only sold a few hundred copies. Guitar Hero came eight years later.

(I worked on all three projects.)

The Stockfish chess engine [1] is very nice, especially so since chess engines tend to be very kludgy.

[1] https://github.com/official-stockfish/Stockfish

Sensei's Library is older than Wikipedia.

It was a really vibrant place for a while, and is still the first place I go when I want to look something up that's Go-related, but it's pretty static at this point. I'm glad that it's still maintained; a lot of resources like that tend to wither and die once their usage starts to fade.

I realize this is classic old-man-yells-at-cloud, but I don't understand why every online article these days, even the technical ones, need to have a giant "amusing" gif every two paragraphs. Do people not pay attention otherwise?

https://towardsdatascience.com/its-only-natural-an-excessive... is another nice overview complementary to this one.

I am 95% sure we removed it long before ship, unfortunately.

I don't like calling it "inferior" (and we didn't, at the time). We solved a much more general problem in a slow and only approximately correct way. They found a much more limited problem that they were able to solve perfectly and blazingly fast. That's a feature! There was definitely room for both approaches.

The multi-height thing was a real annoyance, by the way. For example, my path-finding code really wanted to be 2D (and of course it was mostly expressed in 2D) but had to understand that you could do things like take a bridge over (or even jump over) a gully and then walk around and cross under your original path.

I'm the Dan Schmidt mentioned in the article. A few years back I wrote a few anecdotes about the technical side of working on Ultima Underworld:

https://dfan.org/blog/2011/02/20/the-dangers-of-self-modifyi...

https://dfan.org/blog/2011/02/21/ultima-underworld-bugs/

https://dfan.org/blog/2011/03/17/one-more-ultima-underworld-...

You left out most of the equation. The whole equation as posted is just an application of Bayes' rule: P(A|B) = P(B|A)P(A)/P(B). Your excerpt leaves out everything after P(B|A).

I was confused too; it has a weird format that I think hurts comprehension rather than aids it by making you think you're not looking straight at the data.

Ignore the fact that the "Training Days" axis is drawn diagonally. The system is creating about 40 agents per day; by the end of day 14 it's made 610 or so. The graph shows, for any given time of training (vertical axis, going down), what is the distribution of trained agents that it's chosen in order to be unexploitable (you wouldn't want to choose rock all the time in rock-paper-scissors, for example). So, for example, at the end of day 14, it's using a selection of agents with numbers 595 through 610 or so, which means they've all been created within the last day.

I noticed the same thing. I presume it is a typo. Later on f is defined as you would expect.

I have spent countless hours searching for the perfect programming font, and once I found Go Mono, I never went back.

There's Quantum Information Science (https://www.edx.org/course/quantum-information-science-i) on edX.

Pros:

- It's an MIT graduate class (this could be a con if you are a beginner in other areas and not just in quantum computation).

- It's taught by Isaac Chuang (of Nielsen & Chuang, the leading textbook on quantum computing) and Peter Shor (of the Shor factoring algorithm, the most famous quantum algorithm).

Cons:

- Shor does 90% of the lectures, and he is... not a good teacher.

- When I took it, there was almost zero support (e.g., when there were bugs in the problem sets).

I'm glad I took it, because it induced me to learn a bunch of material I had been wanting to learn for a while, but compared to what it could have been it was pretty disappointing.

When I was a kid, I learned Pascal from a book. We didn't even have a PC at home (it was the 80s) but my mother was a college professor and had access to the mainframe there. I wrote out a whole program to run Conway's Game of Life on a bunch of notebook paper, and one day when I was off from school I came in and typed the whole thing in at a terminal. I ran the command to compile it, and it spit out hundreds of syntax errors. Then it was time to go home.

It was not an auspicious start to my career in computer science.

If I understand your description of the rules correctly, it sounds like that removes any sort of strategy from the game (manipulating the board and the rack), making it a sequence of Scrabble puzzles instead with no between-player interaction. That's not necessarily a terrible thing, but I think it would remove a lot of the interest for many people (me included).

I have read about Nigel Richards, the New Zealander French Scrabble champion who doesn't speak French, and I think that the game he is the champion of is "regular" Scrabble (with French vocabulary). For example the Guardian article https://www.theguardian.com/lifeandstyle/2015/jul/21/new-fre... talks about his "pretty rotten draw of letters".

Edited: Ah, I see from his Wikipedia page (https://en.wikipedia.org/wiki/Nigel_Richards_(Scrabble_playe...) that he has won both Classique and Duplicate titles.

This has been tried (I think for some US Championship tiebreaks), and I like it a lot. If memory serves, the winning time bids for Black were quite low; people really value draw odds.

A just has a kind of a buzzy feel to it. It's not something I can really quantify or prove, just a sort of synaesthetic sense. I would say that I feel it the strongest in the two octaves below middle C. It is basically true for any piano. I don't have a mapping for all 88 notes on a piano, and it's not like you could hand me a timbre and I could identify the corresponding pitch. The whole thing is shifty enough that I assume something else is actually going on and this is just the way that my conscious brain makes sense of it. But it is part of the way that I perceive pitch.

Mostly I feel like the timbre of an organ note is very consistent from pitch to pitch, whereas, say, an A on a piano sounds buzzy to me while a Bb sounds more smooth and a B natural is kind of liquid. (Of course the timbre probably isn't really changing much but that's how my brain perceives it.)

Yeah, but this is sort of like saying "whenever you see red, remember that it's really green instead." It's not like you stop seeing and thinking red.

Oh, geez, I'm not going to go find some specific section of music that perfectly meets all these constraints. Often the pitches are at the same octave, by the way; I don't think that makes the connection any smaller. And often the texture is similar; many aspects of the music may be "recapping" and one of them happens to be pitch. I'm not "confusing anything with a pitch class connection" just because another connections happen to exist; I'm hearing more connections than I otherwise might.

I am definitely talking about non-tonal music in general, yes. Elliott Carter is a good example of a composer where I feel that absolute pitch helps me understand what is going on more than I would otherwise. But again I'm not going to dig out some particular measure of music that perfectly proves my point.

If this reply was insufficient for your needs (I fear it is) I apologize. I'm not trying to win some debate, just to elucidate a little how I feel that absolute pitch aids my music appreciation.