klik99

I mean great film, but how is any of that what we're living inside? The only thing I can see, is easily manipulatable president, smug tech bro, and AI exists, but literally everything past that premise hasn't happened - we haven't handed over defense to complete control of AI, and it hasn't seized control to make a totalitarian state (and at this stage it doesn't seem like we need AIs help for that).

I mean, again, great movie, but a lot of AI discourse these days seems to conflate what might happen for what is currently happening. Every advancement in the field is seen (by some) to be irrefutable evidence for predictions that date far back into the last century. So my read on this article is it's self-referential. This is media part of a discourse has always been part of AI research (at least as far as I've been paying attention - I've been interested in AI/singularity stuff since hearing about it on the MIT campus in early 2000s) so it's really just showing that at least some people have been aware of the potential dangers since then. We weren't living inside it in 1970, and we currently not living inside it today. If the president were to hand any control over infrastructure to an AI, with no human in the loop, then at least we'd be in the first 20 minutes of the film.

I honestly think it's dangerous to keep presenting what _could_ happen as what is currently happening. Once it becomes clear the world doesn't end because of LLMs, people will become more lax about the potential dangers, kind of like how after the millennium bug failed to stop the world economy everyone treated it like a joke, not realizing that it failed to materialize due to a lot of people working overtime.

I asked my wife to not share any info with Muse and she said “Well they know everything already”. Well I was able to convince her not to with some stories about hallucinations of agents getting things horribly wrong as agents in the past, because the privacy concerns just didn’t phase her. Shes even more aware than most people since I’m very privacy aware. 100% it’s very echo chambery to claim that it’s not 1990 anymore. It’s like a historian saying that history won’t repeat itself because we all know what happened last time

Yes, but I don't feel that counts as a danger of the technology - that's a danger of these escalating, unicorn based investments that have to make bigger and bigger bets to survive until the bets as big as the US whole economy. The issues with our financial markets are a whole other set of concerns, but it's not inherently "about" AI. AI is just a vehicle for these smash and grab shenanigans.

I’ve noticed a weird thing about the discourse around this, that it can only be a marketing thing or a true belief, as if everyone working in AI has a monolithic opinion. The tweet kicking off this article has that assumption “it’s not a marketing thing, many people truly believe it”.

It’s not an either/or. Many sincerely held beliefs can be used by cynical actors in cynical ways. It can be both a cynical marketing ploy by the C-suite and a genuine fear.

Personally, based on my experience with AI, the only dangers with the current technology is a) massive loss of employment, b) deployment by humans to control critical infrastructure that LLMs shouldn’t be in control of. Both of those are real, society changing risks that framing the issue as “once we reach the singularity everyone could die” minimizes. Obviously their concerns are possible, but >10% is a random guess. The loss of jobs and how that affects an already K shaped economy is already happening, and nothing is being done for that

This is true but every recent president has expanded powers, and had their party agree. Due to my age, I first noticed it with Bush then when Obama continued it, it felt weird to be fine with allowing more power in the executive branch when it’s someone you support, because that’ll add to the precedent for the next person who you don’t support. The presidency has come a long way since Washington and Trump 2.0 is a clear result / continuation of what’s been happening for as long as I’ve been paying attention.

Typically I agree, NYT on China is usually fear mongering, but in this case the article is very neutral and just the facts. I mean, it's the same kind of thing Palantir is selling to US and that various three letter orgs have started building, but more comprehensive (for the time being) because at least in US there is still a small amount of accountability. This is a glimpse into the future in every country that can afford it, just some countries will take a little bit longer to get there.

I ask that you please read the comments before you make your own comments, just as you asked people to read the paper before commenting (I did!)

> Your assertion that it is all culturally relative ("nurture") and none from the structure of the most efficient algorithm for recognizing the harmonic series ("nature") has no basis in fact

I never said that. I said "Though it's built on harmonics and overtones which can be explained objectively, you cannot make a fully scientific theory of harmony since cultural expectations factor in just as much (and maybe even more) than the expectations around harmonics." To put it in your terms, "you cannot make a fully scientific theory since nurture factors in just as much (and maybe even more) than nature". The world is not black and white, just because I believe sociology to be critical to understanding the emotional effect of music on people, does not mean I'm claiming there is no objective, measurable component.

> If so, then an octave is too simple to fire the recognizer"

Ok so the happiest music is perfect fifths? Or is 3/2 ratio also too simple? So when does this recognizer get complicated enough to fire? 5/4? For someone claiming to have a scientific theory, you are making a lot of assumptions to make this theory work. Scientific theory is about testing hypothesis to reveal truth and not making hypothesis up to match what you already believe. You make an assumption that the brain wraps/aliases frequencies, create a theory of the recognizer, and then make that recognizer have a very convenient threshold that it makes a physical process suddenly the reasoning for "happy". To the "fully scientfic theory" above, I absolutely agree that ratios play a part, simple ratios are boring, and get more interesting as they get complex, then get too complex and it sounds dissonant. That part I think we, and most people, can agree is objectively true, and fits in the theory of expectation vs surprise at a microscale. I just don't agree this translates to "happy" or "sad". Please show me an experiment that indicates the kind of recognizer that you're talking about, or some kind of cross cultural link between ratios and emotional quality. I'd actually be very curious about that, as it's been a while since I've been reading papers about this.

> The music was played into the brain with the root removed; when coming out of the brain of the owl, brain had restored the root

Yes, I'm sure you know this, I'm just showing I know this too - this is called psychoacoustic bass, and happens as part of the feedback loop between the brain processing and the actual ear - the ear actually generates the root tone! This could also be a product of pattern matching though, as we're used to hearing related harmonics (say 200, 300hz) with a root (100hz), so if we just get the harmonics without a root (200, 300hz) the psychoacoustic part of your brain helps out by filling in the 100hz - this is an alternate theory to the aliasing/halving/doubling theory you put out, not saying either one is absolutely true, just that there are other explanations for the owl thing.

It also has nothing to do with the emotional effect of the music.

> So at Berklee they teach recognizer algorithms and artifacts thereof?

Again, I did not say that. I'm not talking about the speculative, non-experimentally backed parts of your theory, just the parts of your essay that are part of the common scientific consensus. I said "ratios and how chords derive from that was absolutely part of my education at Berklee". Nothing more.

Harry Partch has a book Genesis of a Music, which is half autobiography and half a theory that all melody derives from the micro pitch variations of speech - as in melody is a stretched out version of how we speak. I don't know if this idea has ever been pursued more academically but I find it very interesting. If it's true, I also wonder if there's feedback, like our vocal inflections are influenced by music melodies just as much as vice versa.

You might like the book Emotion and Meaning in Music - it starts from some outdated psychological ideas (that emotions are only intensity and that we assign positive/negative valence post-hoc) but ties that to early information theory and makes a really compelling theory of music that's all about the interplay between expectation and surprise, across both the physical harmonic ratio stuff this paper talks about and the cultural associations too.

Yeah, that's really getting to the heart of why the idea of a totally objective theory of music seems legit. Major/minor doesn't exist in a vacuum, there's a lot of other signifiers in the music that are highly correlated. It's not just the major triad, it's the energy of the beat, playing, singing, the bright instruments. If you were a 2 year old and just heard a major chord played on a piano, then a minor chord, then you probably wouldn't get "happy" or "sad" from it. Maybe by 3 or 4 you've already picked up the correlation between "happy music" and "major triad". You pick this stuff up really quick and it's part of your core memories/mental model, so it can feel inherent. It's like the fish that doesn't know what water is.

I've been working with some people building a music model from scratch and it really surprises me how much correlation there is across different features. It's been difficult applying that to general sound effects, because music is highly correlated in ways more general sound isn't.

I meant "fully scientific theory of harmony" in the same way your second paragraph does, not as predictor of popularity but predictor of what kind of effect it raises in the listener. A big part of how music effects the listener is sociology, which sure is scientific, but I tried to use scientific in the way the article uses the term, IE from objective first principles and not culturally specific/learned response.

I read through a lot of this article and it does a good job explaining the ratios and math behind harmonics and how that makes chords, but ideas like lower ratios are more pleasant and minor chords feel like something is missing (from the article "I do not hear the Harmonic Series itself -- something is missing") is just wrong - if it was right, then the happiest music would be entirely octaves. Though it's built on harmonics and overtones which can be explained objectively, you cannot make a fully scientific theory of harmony since cultural expectations factor in just as much (and maybe even more) than the expectations around harmonics. Music is pleasurable when it mostly meets expectations but surprises you as well, IE it sits just outside of what we expect. It then becomes part of what you expect in the future and music evolves.

Ultimately minor sounds sad because it's been used that way for a long time and it's what we expect. I was listening to Trout Mask Replica and my wife said it sounded like her home town funeral music (I later heard some myself and yeah she was right) - Captain Beefheart doesn't sound sad to me, but it does to her. The funeral music uses major 5 note scale.

And on "But I've Been a Musician All My Life / Studied Music In College and I've Never Heard Any of This Before!" - ratios and how chords derive from that was absolutely part of my education at Berklee. This article is good for the objective parts, but really self-congratulating for the subjective/experiential part of it.