Hacker News Reader

Comment by baerrie | Hacker News Reader

It is less like an alphabet, we already know that there is a discreet number of click clusters (basically an alphabet). Here they have found that the sperm whales modulate tempo of these clusters, rhythm within the clusters, add additional ornamental clicks, and change the rhythm of individual clicks within clusters over time (rubato). This adds an insane order of magnitude to their language much like our use of tone, context, and all the other ways we take our alphabet and enrich it with exponentially more meaning. This is the first time this has been proven in an evolutionary lineage separate from ours!

View on HN

What are the chances of us being able to decode some of that in our life time? There must be some meaning to it, even if it's not as complex as human speech.

We should do it just to make sure they're not saying "so long and thanks for all the plankton."

Extremely likely, especially with the increasing abilities of LLM to decode unknown languages. Then the test would be for us to produce these sounds and see if the whales respond as expected.

How would we train the LLM to actually decode it though? Don't we need some way to weigh the results?

My guess: train a generative model to predict whale sounds, based on recordings of real ones, and hope that the resulting latent space will map to the one of a human-trained LLM. We'd need a stupidly large amount of recordings of whale songs, a tokenization scheme, and few already translated sounds/phrases to serve as starting points for mapping the latent spaces.

Exactly. Also, I think an alternative to LLM that is more generally trained towards identifying large linguistic patterns across a language could be cross referenced with the aforementioned more standard llm to at least point to some possible meanings, patterns, etc

We'd need contextual tracking of what the whales are actually doing/communicating to match to the songs. An LLM would be excellent at finding any correlated patterns between the language and actions, and then mapping those to similar English concepts, but that all requires the behavioral data too. Cameras strapped to whales maybe?

I unironically want this more than another particle collider.

We'd do it without its "help" and give it the results which it would then recombine and hallucinate.

Would just need a way to tokenize, then use predictions to map back to some positive interaction symbol. Something like we think a certain phrasing means "food-fish-100m-down" and whales respond consistently to that.

I dunno if sometimes the language would be contextual, and utterances could not be understood without taking into account the context of what is occurring, or the speaker. Yes I know human language can be subject to these variables too. Anyhow it's all speculation and the dream of talking to animals is surely exciting.

Also, a Youtube doc about researchers attempting to teach dolphins english: https://www.youtube.com/watch?v=UziFw-jQSks

Imagine if they are communicating using a lot of pronouns.

I can’t even understand some other people when they keep switching the target of the pronoun without being explicit.

“He is tired. He dropped the ball on his foot. He yelled at him for being tired.”

(How many people are here?)

I've heard that "da kine" in Hawai'i Creole English historically was, and still may be, used exactly in situations where the speakers share plenty of context, allowing them to figure out what it denotes, but leaving listeners largely unenlightened.

compare "dude" in Fig. 1 of https://acephalous.typepad.com/79.3kiesling.pdf

In a language such as Thai, pronouns are left out in most cases, and only added when you need to disambiguate. No plurals either, requiring you to add this information with extra words when it matters. But nobody forces you to communicate effectively, or use Oxford commas.

> Imagine if they are communicating using a lot of pronouns.

That's fine. The idea is to record them with lot of metadata in situ. Recording what is going on with the whales. (are they feeding? are they traveling? are they in a new location or somewhere they have been for a while? How many wales there are?) And also about their surrounding (sea state, surface weather, position and activity of boats, prey animals etc etc.)

What increasing abilities of LLMs to decode unknown languages are you referencing?

(I possibly missed a paper)

See e.g. here: https://ai.meta.com/research/no-language-left-behind/

If you scroll down, the very first step they describe is for collecting datasets of existing translations. They aren't translating even unknown human languages, let alone completely alien ones.

No idea of this is even vaguely in the right direction, but this comes to mind: Unsupervised speech-to-speech translation from monolingual data

https://research.google/blog/unsupervised-speech-to-speech-t...

We'll just fine tune our existing models with data scraped from the whale internet. Surely that will work.

You could train an llm on all existing whale sounds, get it to “listen” to live whales and respond with what it “thinks” it should, then do human analysis on the results, maybe find one shred of meaning, rinse and repeat.

That's literally impossible. Imagine trying to learn Japanese by talking to a Japanese man on the phone, with neither of you being able to understand or see each other or what you're each doing. Without shared context communication is impossible. Best case, you and the Japanese man would create a new context and a new shared language that would be neither English nor Japanese that would allow you to communicate about whatever ideas fit through the phone line. Maybe words like "sound", "word", "stop", etc.

Well even one of those words could be enough. If I knew he was in danger by the terror in his voice well then probably one of those words is “help”

Impossible in a single step, but perhaps impossible is too strong a word to reject the possibilities that arise when we consider how the statistics of words, or sounds are connected. If you can work out the statistical correlation between groups of sounds you can start to gain an idea of how they are interrelated. This is a stepping stone on the path to understanding.

I could see that being possible with a human language, but a non-human language? No way near enough context, I'd think.

>the statistical correlation between groups of sounds

That assumes that the speaker is similar to the person correlating the sounds. For example, if you had statistical data for utterances of English sounds in the context of Magic the Gathering tournaments, and you tried to decipher the speech of a Swahili electrical engineer talking about transistors, you could very well decipher something that's seemingly coherent but entirely incorrect.

It would be an overgeneralization to assume that whales speak about things in the same statistical patterns that humans do.

You just applied the concept of conversation and topic to yourself when the proposal did not suggest that.

Then what's needed is a sea drone that tags along with whales and collects context for their language.

Nice. Probably just start with whale twitter and go from there

Where do you get this idea that LLMs can be useful “to decode unknown languages” at all?

You would need some way to convert the whale LLM to human language though. Otherwise you would just be making pre trained GPT4 for whales. One option would be to label data according to induced reactions in whales to whale language completions (i.e., let the LLM complete whale language and use the reactions to try to induce some understanding. But it feels unlikely we would get further than providing a chatgpt for whales that only they can understand.

You wouldn't necessarily need that. You don't actually need translated text for every single language pair a LLM will learn to translate.

ie train a LLM on English, French, Spanish data. This data only contains parallel text in English-French. Can this LLM still translate to and from Spanish ? Yeah.

You still have a bridge and each of those languages are not just from the same species but the same language family. If there’s English to French and French to Spanish there’s a semantic relationship between English and Spanish.

There exists no bridge to whale any more than there is aliens from Alpha Centauri.

Common concepts are common, what species the language is in is not as relevant as you think. Text and Image space, two entirely different modalities are so related in high dimensional space, you can translate between them with just a simple linear projection layer.

There's an enormous gap between being able to "decode" it and actually understanding what it means.

Wittgenstein's quote on lions is still relevant.

Wittgenstein didn't present any evidence.

Reminds me of the Chinese room [1] argument: Does a computer really understand Chinese language if it can respond to Chinese inputs with Chinese outputs?

[1] https://en.wikipedia.org/wiki/Chinese_room

> Searle's thought experiment begins with this hypothetical premise: suppose that artificial intelligence research has succeeded in constructing a computer that behaves as if it understands Chinese. It takes Chinese characters as input and, by following the instructions of a computer program, produces other Chinese characters, which it presents as output. Suppose, says Searle, that this computer performs its task so convincingly that it comfortably passes the Turing test: it convinces a human Chinese speaker that the program is itself a live Chinese speaker. To all of the questions that the person asks, it makes appropriate responses, such that any Chinese speaker would be convinced that they are talking to another Chinese-speaking human being.

> The question Searle wants to answer is this: does the machine literally "understand" Chinese? Or is it merely simulating the ability to understand Chinese? Searle calls the first position "strong AI" and the latter "weak AI".

(Emphasis added)

If we were to make an analogy to contemporary machine learning, we're talking about the difference between an LLM (with context) and a Markov Chain. 'Understanding' requires novel reuse of recollections. Recollections require memory (i.e. context), and the novel reuse of those recollections require a world model with which to inference.

I am personally in favor of the use of the word 'understanding' with LLMs, but it should be noted that many people strongly disagree with that use of the term.

You're saying that LLMs meet the stronger definition of "understanding?" I disagree: You're confusing necessary with sufficient. [0]

Take the original analogy of a person with an if-then opaque phrase table, and make it a slightly fancier two-step process where they must also compute a score based on recent phrases or words to determine the proper output.

So now the system has "memory" and gives better output... But was that sufficient to create understanding? Nope.

[0] And perhaps not even necessary, if we consider an operator who does understand Chinese but their short-term memory has been damaged by an injury.

I think the confusion comes here in believing that humans have some exceptional mechanic for understanding. Ultimately human intelligence is the product of neural networks. The only thing that's separates us from ML is both scale (in terms of compute and memory) and agency, the latter giving us the ability to train ourselves. We take inputs, we output behaviors (including language). The idea that we are, in all actuality, deterministic computers is really at the heart of the existential panic around AI.

> Ultimately human intelligence is the product of neural networks

Citation needed?

Sounds like "human intelligence is the product of <current tech hype>"

You do realize that artificial neural networks are based on the organic ones? Unless you have some dualist interpretation of intelligence that involves some ethereal intangible soul that somehow eludes all detection, nor holds any explanatory power for intelligence?

If you were obliterate your own neural network, by, I dunno, shooting yourself in the head, might that have some affect on your intelligence? Head trauma? Brain cancer? Alzheimer's? Senescence? Do I really need a citation that your brain is where intelligence comes from? *scratches head*

Artificial neural networks have departed any biologic principles.

So comparing ANN to NN no longer makes sense to me.

> Artificial neural networks have departed any biologic principles.

ANNs are perceptrons. Perceptrons mimic biologic principles.

> Citation needed?

"Speech and Language Dysfunctions in Patients with Cerebrocortical Disorders Admitted in a Neurosurgical Unit"

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6417298/

> Sounds like "human intelligence is the product of <current tech hype>"

What? Neural networks is a biological/medical term. Are you confusing it with _artificial_ neural networks?

In a weird way "hype" undersells how history repeats itself: Rene Descartes promoted the idea of the brain as a pump for the spirit, then in Freud's era it was the steam engine, etc.

You can only be trolling if you say brains aren’t made of neural networks. Poe’s law in effect

> if you say brains aren’t made of neural networks.

Except biological brains are not made like hyped-up artificial neural networks. The latter are from a simplified hypothesis then massively twisted and cut-down to fit practical electrical engineering constraints.

> The only thing that's separates us from [machine learning] is both scale [...] and agency

... And the teensy tiny fact you keep refusing to face which is that they don't work the same.

Declaring that the problem is basically solved is just an updated version of a perennial conceit, one which in the past would have involved the electrical current flow or gears.

> simplified hypothesis

There's no hypothesizing at this point. Neurons have been studied in the lab since – checks notes – 1873. Modern neural nets have largely taken Occam's Razor rather than precise biomimicry, mostly due to 'The Bitter Lesson' that basic neural networks show more generalizable emergent behaviors when scaled vs being clever about it. e.g., dendrites themselves have been shown to behave something like a multilayer perceptron on their own. So it's really perceptrons all the way down when it comes to brain circuitry.

https://en.wikipedia.org/wiki/Golgi%27s_method

> electrical current flow or gears

Perceptrons were built to be mathematical models of neurons. When gears or 'electricity' were first created/harnessed, there was no intention to build a model of the mind or to mimic neurons whatsoever. There really is no weight to this argument.

> Declaring that the problem is basically solved

I'm not making that declaration for whatever you might be terming 'the problem' here. I'm just stating that 'understanding' is still (incorrectly) rooted in our belief that our means of computation is exceptional. As far as anyone can tell, 'understanding' isn't substrate dependent, and as far as we know, our 'understanding' comes from neuronal computation.

> The question Searle wants to answer is this: does the machine literally "understand" Chinese? Or is it merely simulating the ability to understand Chinese? Searle calls the first position "strong AI" and the latter "weak AI".

I don't know why anybody thinks this is a profound question worth spending time thinking about. All it boils down to is whether or not we reserve the word "understand" for beings with souls or whether we allow it to describe machines too.. and BTW there's no empirical test for souls (not even for people!) so it's really just asking if "understand" is some sort of holy word reserved for people who have faith and believe in souls.

The question isn't about machines and their empirical capabilities at all, it's just about how humans feel about that particular word.

Isn't a large company more handy model than a "Chinese room"?

Hordes of these in practice own our species. They come up with the decisions on carbon footprint, on whether to colonize Mars, on where to shift collective attention.

Although these entities are only one of achievements of our culture, for now they dominate on global level.

Big achievement. "Consciousness" not strictly needed. Whatever the definition of "consciousness", a big company is driven by a glorified stack of excels and definitely couldn't write Chinese (assuming none of employees does, after Searle).

So, running a planet and overriding any individual human is not enough, let's focus on entertaining an individual human with an interesting conversation in Chinese? What do you even seek with that model? /s

it would be interesting if data mining and embeddings could reveal meaning hidden in the patterns and context.

Good point on the difference between decoding and understanding. The paper "What is it like to be a bat?" [1] comes to mind. We can certainly try to decode the behaviors and language of whales, but understanding their mindset is an entirely different matter.

[1]: https://en.m.wikipedia.org/wiki/What_Is_It_Like_to_Be_a_Bat%...

Probably that I miss the deepness of this quote, but lions or whales are mammals like us and they must be chatting about their basic needs which we should be able to relate to.

For Wittgenstein words don't so much express things in themselves, but are deeply integrated in nonspoken contexts and impressions. It's hard enough to really /get/ somebody from a different cultural context, whereas a lion lives in a massively different lifeworld, with a very different body.

What does a word like "hungry" mean to you, when your diet is gorging yourself on freshly killed meat every couple of weeks? What does feeling "tired" mean, when you're the size of a bus and sleeping is done in increments of 10 minutes because you must constantly wake and resurface to breathe through a hole in your back? You can use language to gesture at these experiences, but that's different from really knowing what we are talking about.

Not that I understand Wittgenstein very well at all, even though it seems like it should be easier than understanding whales and lions ;)

> It's hard enough to really /get/ somebody from a different cultural context

This is such an eye roll statement. The world “really” is the one which does all the work in it. One would say “nah, communication is actually amazing, and through listening to people / reading their words and being attentive one can understand many different cultural contexts”. And whoever wrote the original sentence can always come back and ask “but did you really really understand them”?

> What does a word like "hungry" mean to you, when your diet is gorging yourself on freshly killed meat every couple of weeks?

I don’t understand the problem here. “Hunger” is the feeling which make you seek out food. If your normal is to gorge yourself full and don’t eat anything for weeks then you are not hungry for most of those weeks and then when you start feeling hungry that is when you start preparing for a new hunt, chasing a herd and so on.

We already and in our everyday life understand such differe ces. We know for example that newborns need feeding a lot more frequently than adults, so when they become fussy we ask “is he hungry maybe?” Despite the fact that our own hunger works on a different schedule. We also undestand that when a holocaust survivor describes the deep hunger they suffered from in the camp that is a different feeling than the one you feel 2 hours after lunch when you start thinking about opening a pack of chips. Somehow we “get it”. But do we “realy” get it? Yeah i mean we know that the baby will become more and more fussy the longer it is not fed, and will die eventually. We know that the lion on the hunt will eat us if we look easy enough prey. We understand that people in the camps turned to eat things they would have considered inedible before, and the experience left lasting psychological scars on many of them. And we understand that the person thinking about opening a pack of chips will stop thinking about it if something distracts him/her for a minute. “But do we really really really understand them?” Idk. You tell me what you consider “really really really” understanding anything. Because one can play this game forever.

> What does feeling "tired" mean, when you're the size of a bus and sleeping is done in increments of 10 minutes because you must constantly wake and resurface to breathe through a hole in your back?

If that is normal for that being then they wouldn’t describe that state of being as tired. It is just their normal. But if something (illness, activity, noises) disturbs their normal which make them seek out more sleep they would call that being tired.

>And whoever wrote the original sentence can always come back and ask “but did you really really understand them”?

Asking questions that are very difficult to answer precisely and well, and then trying to do so, is kind of the whole deal with philosophy. How /do/ you know the things that you think you know? You can roll your eyes at it and choose not to investigate further, or you can try to get super granular about sussing out an answer. Philosophers generally try to do the latter.

If your answer to "what does it mean to know things" or "how do we actually use words to navigate meaning" or "how well can we understand other living things" is "I don't care, that's nerd shit" then I highly recommend that you don't read Wittgenstein, because you're gonna get real bored real fast.

> highly recommend that you don't read Wittgenstein,

I did read Wittgenstein. I still hold contempt in my heart to those aping his thoughts mindlessly.

> Asking questions that are very difficult to answer precisely and well, and then trying to do so, is kind of the whole deal with philosophy.

Yes. And declaring that things are impossible is not that. "If a lion could talk, we wouldn’t be able to understand it." is a feel. It's not the conclusion of a reasoned argument. (I'm talking about the sentence in the context it is found in the Philosophical Investigations of Wittgenstein)

> If your answer to "what does it mean to know things" or "how do we actually use words to navigate meaning" or "how well can we understand other living things" is "I don't care, that's nerd shit"

No. Those are all very interesting questions. But if you start with "understanding is impossible" then I hope you get a hug from a loved one. That stuff is hurt speaking. It is not a usefull starting point to answer any of those questions.

I don't think anybody here is upholding Wittgenstein as an exemplar of an emotionally well-adjusted person. I'm not proselytizing for him, just explaining his take to the best of my ability.

I don't disagree with this criticism of his position either, for what it's worth - Personally I think Wittgenstein ignores empathy as a transfer of embodied nonspoken knowledge, for one.

It'd be one thing if his example were of something very radically different (IE space aliens), but social mammals like lions and whales and humans all learn to navigate the world through the presence of others first and foremost. Long before you have any awareness of language, even long before birth, you will have an awareness of closeness. You will primarily learn to navigate the world by way of your relationships to others. That goes whether you're a lion cub or a human baby.

> It'd be one thing if his example were of something very radically different

Yeah absolutely! I just think Lions would be easy. :D But imagine trying to explain to a tree what a "having a bad commute" is. :)

a big part of the project is to not just collect whalesong data but also measure what activities the whales physically engage in as they “speak”.

Temba, his arms open.

Darmok and Jalad at Tanagra.

The topic paper is part of a big project called Project Ceti that's aiming to do just that.

On the theoretical side for why it may be possible, see https://arxiv.org/pdf/2211.11081 https://arxiv.org/abs/2303.10931

I am now anticipating some form of an existential crisis around whether animals that can talk via a specific language are any less entitled to rights than other humans.

We barely give certain humans rights, we'll have no issue ignoring the rights of animals

Don't worry, we've had a pretty pisspoor track record with fellow humans too. There's always room to turn it around

It's always surprising to me when people seem to be unaware that things like human rights or international laws only exist as long as every authority (i.e. everyone with the power to enforce it with violence if necessary) agrees with them.

The US literally passed a law allowing itself to invade The Hague if necessary to free any US service member brought before the International Court of Justice. Even the signatories of the UN human rights conventions could simply refuse to cooperate and experience no meaningful consequences. International laws are just treaties and treaties can be cancelled or revoked. Human rights are no more reliable than company mission statements.

Not to mention that even on a national level laws only exist as limitations the state imposes on itself (or to chisel away at limitations it imposed on itself before). This is why it took the civil rights protests and literal terrorism to arrive at universal suffrage and civil rights for women and Black people. Of course this only works if the government would rather change its laws than escalate protests to a civil war and which one your government chooses might surprise you.

Don't forget us gays, too.

And yep, everybody's just a monkey with or without a stick.

So what you're saying is to some extent everything's made up and the points don't matter, so maybe we can rope animals into this same messy web of loose promises too!

No, I'm saying unless those animals pick up guns and fight for their rights, they'll have to rely on how icky society at large feels about mistreating them and how willing we are to cause a ruckus about that to outweigh the influence of the moneyed interests that benefit from their lack of rights.

Of course if for some reason the domestic meat industry collapsed and people relied on imported meat there would be an incentive for praising lab-grown meat or veganism because it would benefit the domestic companies but that's more or less the inverse of what's happening in the US.

I expect there will plenty that think animals speaking is nothing more than a politically motivated hoax, no matter what research shows.

Clever Hans

https://en.wikipedia.org/wiki/Clever_Hans

I expect most of our communications with them will be political in nature or will involve apologizing to them.

woah a double rainbow!

Animals already can be considered to have certain rights. The interesting question is whether a truly sentient animal could be considered a person - then they would have ALL the rights humans have.

"A truly sentient animal"?

Aren't all animals sentient? And what is the difference between "truly sentient" and "sentient"?

They probably mean sapient. We consider most animals sentient but only humans sapient.

That said, rights don't happen through scientific discovery but through reform or revolution. There are strong economic incentives to keep animal rights minimal and there is only very weak political will to change that. It also doesn't help that in the US there are strong religious doctrines that oppose the notion of animals deserving rights at all.

It's worth mentioning that there are indeed jurisdictions (not sure about the US) where animal cruelty laws are actually framed as animal rights (usually with some limitations, e.g. only considering vertebrates) rather than as protection of private property but even they usually don't include a right to life or bodily integrity and are flexible about what constitutes cruelty and what can be done to reduce it.

> Aren't all animals sentient?

How far are you willing to go with this? Is a roundworm sentient? An amoeba? A bacterium?

No, an abstract entity such as a company is already given the status of a person, but it doesn’t mean it has the same extent of rights a human being is given in the same jurisdiction.

Corporations aren't natural people, my friend.

Wales, maybe.

Thank you for giving me some friendly response. :)

I understand the feeling here, definitely wales and corporations are not of the same ontological class.

What I was trying to point out however, was that just because a jurisdiction give an entity the status of person, it doesn't mean it give to it the same extend of rights.

Could be just like that classic Gary Larson canine decoder cartoon.[0]

Could be more musical. I hope / fear it's something more sophisticated.

[0] https://twitter.com/Andr6wMale/status/1501487704399298560/ph...

> even if it's not as complex as human speech

I wouldn't make that assumption. Just because whales don't have opposable thumbs and have a harder time making and using tools doesn't mean that their speech isn't advanced.

The people hand-rolling their own hydrophones think there might be a lot going on we simply haven’t been able to hear yet.

> This adds an insane order of magnitude to their language

Do we assume they only have one language? We humans have thousands of languages.

So more sophisticated than a mere alphabet, if anything

Exactly

Maybe it's more like phoneme set.

Reply on news.ycombinator.com