Han

There's a blogger you might enjoy reading whose name is Ramin Shokrizade: http://www.gamasutra.com/blogs/author/RaminShokrizade/914048/ . He's some kind of consultant for video game monetization schemes. I think he's a little bit hyperbolic and overwrought sometimes, but he has a lot of direct experience and textual evidence collected from other designers at companies like Zynga.

I think there are a lot of psych topics that are relevant for freemium games but not normal gambling, which means they're a great zone for research. Normal gambling games like poke... (read more)

I think you're right. I'm badly overlooking a subtlety because I'm narrowing "describe" down to "is a suffix of." But you're right that "describe" can be extended to include a lot of other relationships between parts of the big sentence and little sentences, and you're also right that this argument doesn't necessarily apply if you unconstrain "describe" that way. (I haven't formalized exactly what you can constrain "describe" to mean -- only that there are definitions that obviously make our sledgehammer ar... (read more)

I thought Gurkenglas' solution was a lovely discrete math sledgehammer approach. There's a lot of subtly different problems that Thomas could have meant and I think Gurkenglas' approach would probably be enough to tackle most of them.

(Attempting to summarize his proof: Some English sentences, like the one this problem is asking you to dig around in, are countably infinite in length. If some English sentences are countably infinite in length, and any two of them have different infinite suffixes, then there's no way the text of this sentence contains both of them.)

Not a long note or a detailed dissection, but just a reminder: whenever you take single-dimensional data and make it multidimensional, it becomes harder and more subjective to analyze it. (EDIT: To clarify, you can represent multidimensional data multidimensionally. But mapping multidimensional data to a lower-dimensional space usually involves finding a fit, which can introduce error. Mapping it to a lower-dimensional space is usually an important step in explaining it.) I suspect you'll find that if you have this many dimensions for people to respond by,... (read more)

I think there's a rule-of-thumby reading of this that makes a little bit more sense. It's still prejudiced, though.

A lot of religions have a narrative that ends in true believers being saved from death and pain and after that people aren't going to struggle over petty issues like scarcity of goods and things. I run into transhumanists every so often who have bolted these ideas onto their narratives. According to some of these people, the robots are going to try hard to end suffering and poverty, and they're going to make sure most of the humans will live f... (read more)

I think it's a little bit worse than this.

A lot of people who gamble compulsively don't do it because the odds are beyond them. (It's really easy to play slots a bunch of times, lose a lot of money, and realize you lost a lot of money.) There's something neurologically strange about people who gamble frequently even though they lose, and it's hard to pinpoint it, but it seems like variable reinforcement is winning out over logic.

If you buy a large number of lottery tickets, you're pretty likely to win some sort of prize. Related example: slot machines are ... (read more)

Thank you for the information! My brain does something weird when I see the word "actually," so I don't think I was charitable when I read your post.

Oh, absolutely! It's misleading for me to talk about it like this because there's a couple of different workflows:

• train for a while to understand existing data. then optimize for a long time to try to impress the activation layer that konws the most about what the data means. (AlphaGo's evaluation network, Deep Dream) Under this process you spend a long time optimizing for one thing (network's ability to recognize) and then a long time optimizing for another thing (how much the network likes your current input)
• train a neural network to minimize a loss f

I'm confused. Isn't it evident from the rest of my comment that I agree with you?

(On an unrelated note: I think my upvote button has vanished. Otherwise I would have clicked it for your post!)

You're probably right! (At least some of the time.)

In music, I know a lot of people who think about things the same way you do, and they sensibly learn to use versatile tools like FM synthesis because FM synthesis covers a wide range of sounds really broadly. A lot of them even know how to make human voice-like sounds using these tools.

On average if you stick to those tools you'll do pretty well. They still fall back on using physical instruments for a lot of techniques, because you can do elaborate expressive things with physical instruments a lot more ea... (read more)

Thanks for the trouble of posting this!

I don't see the paint of exploring many different kinds of 2D painting. I would expect that a digital pen beats most other tools. Especially in the future as technology advances.

There are a lot of people who say that piano is the most versatile instrument, and they're right about that on a superficial level. You can do polyphonic things with a piano that you can't do with a clarinet or a trumpet. And like a digital pen, a digital piano can simulate a lot of other instruments, especially if you hook it up to flashy synthesis software that knows all the d... (read more)

The risk with an AI is that it would be capable of changing humans in ways similar to the more dubious methods, while only using the "safe" methods.

I think what you're saying makes sense, but I'm still on Dagon's side. I'm not convinced this is uniquely an AI thing. It's not like being a computer gives you charisma powers or makes you psychic -- I think that basically comes down to breeding and exposure to toxic waste.

I'm not totally sure it's an AI thing at all. When a lot of people talk about an AI, they seem to act as if they're talking abo... (read more)

Oh, thanks for the link!

I think you misunderstood me, or maybe I wasn't clear. I meant "of the strategies which we used to search for musical ideas, none of them involved solving NP-complete problems, and some of them have dried up." I think what neural nets do to learn about music are pretty close to what humans do -- once a learning tool finds a local minimum, it keeps attacking that local minimum until it refines it into something neat. I think a lot of strategies to produce music work like that.

I definitely don't think most humans intentionally sit down and try to solve NP-complete problems when they write music, and I don't think humans should do that either.

I really like your thread: thank you for writing me back!

I think you have good intuitions about how sound works. I don't think I can determine whether there's a consensus on what is good: I'd venture to guess that any audio humans can perceive sounds good to someone. A friend of mine sent me an album that was entirely industrial shrieking.

But I agree with you that there's a limit to the distinctness -- humans can only divide the frequency spectrum a certain number of times before they can't hear gradation any more, they can only slice the time domain to a ... (read more)

I think two of your premises aren't necessarily true:

So if I hit random piano keys with my hands a few times and call it a song, the consensus of music listeners would be that Beethoven's Fur Elise is a better song.

Probably, but I think your example is a little bit too extreme to demonstrate your point. There are a lot of genres, like taarab, that won't sound like good music to you because of your cultural background. Acid house probably wouldn't sound good to people who were raised in the 1800s, either. There are commonalities between how people appre... (read more)