Live data from Hacker News

Attempting to interpret sperm whale clicks with AI, then talk back

hakaimagazine.com

91–100 of 138 posts

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#91
post #87
post #62

Earlier quoted context omitted.

It has to do with comment quality, not bad words.

There are so many non-substantive comments in this thread and you chose this one. Stop tone policing.

Uyuxo, tone policing is part of dang's job, and it makes this site better.

FYI, your comment was dead and I vouched for it, not because it's a particularly good comment, but because it's not spam. Please try to add value, though.

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#92

Earlier quoted context omitted.

Your point stands, but I want something wholly different from that. I want the equivalent of the LHC for cross species communication. You're telling me to go do some communing with physics, because we've already got some good textbooks on quantum mechanics.

> I want the equivalent of the LHC for cross species communication. This is not a valid concept. What would the "equivalent of the LHC for communication between ordinary Americans and the Twitter API" look like? What uses would it have? There's nothing out there that can receive arbitrary messages. Communication -- of any kind, with anything -- is limited to messages that the other party is capable of receiving.

This is Wittgenstein's argument about why we'll never communicate with lions (iirc) right?

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#93
post #85

Earlier quoted context omitted.

Yup. I knew it because I fell in a Wikipedia rabbit hole about animal intelligence not long ago. It was very intriguing to me, as someone with no previous knowledge, how this was assumed to be an only human trait yet the fact was pretty much glossed over. I find fascinating the idea of a step between being stuck with the information that others emit and being able to and request arbitrary information at will, being p…

I'm working on a sort-of language as a side project and have fallen down many of the same rabbit holes that you have. I don't think questions are one of the key features that make language special because in the language I'm working on questions are an emergent property. There are two structural words in my language, "propose" and "tell". From these words you can build complicated ideas such as lying ("I propose to y…

>The feature of language that seems the most surprisingly powerful is placeholder words. The words like "someone", "somewhere", "somehow". >I call these the entropy words because they are the words for the information you don't have. They represent sets of possible things rather than a literal specific thing.

As an aside, you might know this, but that's how questions are formed in Chinese. You basically use an enunciative phrase with a placeholder word in the information you're missing ("this pencil is whose", "the dog is where", etc).

Going back to the point, I'm not sure whether the limitations of the language are actually a limitation of the brain that uses it. Sure, if you get into your mind that you want to ask a question, and the opposing party is able to understand that, even gestures will do with a little bit of ingenuity. But that has a lot of prerequisites: you need understanding that some information can be useful to you, understanding that other living beings own that information, some abstract thinking to express an idea that is not literal...

It's not even about questions I think, there's a jump between "tiger here now" and "tiger potentially here at some point, watch".

But still, questions particularly seem to be a clear cut "non existent outside of humans", even in weird emergent forms. There must be a reason they're not there, right?

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#94
post #2

Whale: “Hey, I found some fish.” Whale Eliza: “Interesting. How does that make you feel? Whale: “What the fuck?”

Not sure if it's the case with whales, bit as far as I know there is no recorded use of questions in the animal kingdom - all communication seems to be enunciative, or orders. Questions are exclusive to humans...

There seems to be a lot of wishful thinking in the research. The article mentions the Great Tits having 2 calls for alarm, which if made out of order have no effect, and calling that grammar. You could also call it a single call. Grammar to me means being able to construct new meaning from parts. While interesting and maybe useful, if this sort of research bears fruit I believe lots of people waxing lyrical over how amazing it will be to talk with the animals are going to be rather disappointed. Basic conversation requires a vocabulary in the hundreds, and so far it seems we are looking at word-equivalents maybe in the dozens, in the more capable species. And then you have all the concepts we take for granted except in infants, such as 'I' and 'you' and self awareness.

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#95
post #6

Earlier quoted context omitted.

Only the "finally". We've been talking to other humans and then killing them for resources for a very long time.

We've always been able to talk with those humans though and most of the times we fight them for their resources, not their bodies.

Humans were often raped and enslaved when they were pillaged. Sometimes used for food or entertainment.

It is a sad truth that humans had little trouble viewing other humans as resources.

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#96

Earlier quoted context omitted.

Brain size may not necessarily correlate with intelligence because it isn't the number of neurons, it's the organization and optimization of them. Just like we make much more powerful CPU chips in the same volume of silicon as before. And it could be that it just isn't necessary for whales to optimize brain density, like it is for humans and crows.

Agreed! But afaik the neuron process "node" across species is not so different. The relevant metaphor I would think is humans have pcie and a large register count while many (all?) others are still on pata and register starved. Maybe a real biologist can come in with more facts and less bad metaphors

So far it is looking like the number of neurons in the cerebral cortex is the best indicator of intelligence.

http://mitp.nautil.us/feature/227/the-paradox-of-the-elephan...

We started as basic primates, started cooking our food, getting more nutrition, allowing our guts to shrink, allowing us to spend more energy on our brains, growing more and more neurons in the important bits, at which point things spiraled away using our new language and technology to make better technology.

(Not a real biologist)

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#97
post #74

God I cannot wait to talk to whales. Their oral histories must be incredible. People have fantasized about communicating with extraterrestrials for ages. I don't understand why we haven't invested significant resources in trying to communicate with the other animals on our own planet. What an incredibly weird and non translatable experience it will be to (finally?!) start this adventure with whales. Tangent time. If…

I mean.. hate to be a party pooper, but just because we created a the equivalent of gpt3 for whales, does that mean we can do anything useful? Like talking to whales.. we haven't even established how their language works. What language even is, is a good question. I read it once demonstrated as this; some species of monkey has a specific call they do when they see a panther, and it results in all the monkeys who hear…

I think your entire post is hugely optimistic despite the negative overtone! The fact that other people see there are these important and fundamental questions that can be addressed by this nascent field gives me hope that this won't be another one off attempt like so many before. I'm very hopeful that it can snowball due to the hype around machine learning even if I'm personally skeptical that machine learning is a particularly fruitful avenue.

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#98

Earlier quoted context omitted.

Your point stands, but I want something wholly different from that. I want the equivalent of the LHC for cross species communication. You're telling me to go do some communing with physics, because we've already got some good textbooks on quantum mechanics.

> I want the equivalent of the LHC for cross species communication. This is not a valid concept. What would the "equivalent of the LHC for communication between ordinary Americans and the Twitter API" look like? What uses would it have? There's nothing out there that can receive arbitrary messages. Communication -- of any kind, with anything -- is limited to messages that the other party is capable of receiving.

I'm not sure I understand your complaint or why you've caught the Twitter API into it?

I imagine a center for cross species communication research would coordinate scientists across a couple important fields also interested in the question and allocate their resources to the research which seems most promising to make progress. Just like any other institute.

I imagine the fields would be biology zoology linguistics computation and probably since experts for the specific target species.

Re: Attempting to interpret sperm whale clicks with AI, then talk back

#99
post #74

God I cannot wait to talk to whales. Their oral histories must be incredible. People have fantasized about communicating with extraterrestrials for ages. I don't understand why we haven't invested significant resources in trying to communicate with the other animals on our own planet. What an incredibly weird and non translatable experience it will be to (finally?!) start this adventure with whales. Tangent time. If…

I mean.. hate to be a party pooper, but just because we created a the equivalent of gpt3 for whales, does that mean we can do anything useful? Like talking to whales.. we haven't even established how their language works. What language even is, is a good question. I read it once demonstrated as this; some species of monkey has a specific call they do when they see a panther, and it results in all the monkeys who hear…

I would expect the vast majority of their comms to be :

Nice water over here

Loads of food here

I'm down to F, anyone else?

Do we really have any evidence that suggests it'll be anything else?

Post reply on HN