(CW:AI) Discussion of LLMs and other kinds of text generators ("AI"), especially of toki pona

[cedar vortex]

I ask how you think we could prove or disprove "screen theory" and you immediately turn to religion instead of giving an answer

[gleaming laurel]

I've been saying the entire time, you cannot prove or disprove "screen theory" using science

science cannot investigate the non-physical

to start the investigation you'd have to die, and then that does not even prove anything, if you don't "remember your past life"

[cedar vortex]

if souls existed they would be quantifiable, their influence would be observable

[cedar vortex]

if this is not true then souls would have no influence over a person and then for all intents and purposes, not exist

[gleaming laurel]

only if you think souls are physical things? if souls are non-physical, you by definition cannot observe them in a way that can be scientifically proven (I do not believe in souls ftr)

[cedar vortex]

how would a soul influence one's behaviour if it does not interact in any way with a person

[gleaming laurel]

most religions that believe in souls absolutely believe they interact with a person, all the time. and many persons claim they can "feel" the soul, or have some justification for knowing it exists. but again, that is not something you can prove with science. Obviously we all agree you cannot publish a rigorous scientific paper that says "Souls exists because I felt in my bones". Or the opposite, "Souls don't exist because I don't feel it in my bones"

[cedar vortex]

this would be very much something you can prove with science

[gleaming laurel]

how? a soul is non-physical. you can't prod it, or measure it. many christians and hindus think they are eternal and unchanging, so there's no hope of science there

[cedar vortex]

if you can't measure it then it doesn't interact with the brain

[gleaming laurel]

that's the premise that non-physicalists would disagree with...

[cedar vortex]

I feel like i've been saying the same three things over and over again

[cedar vortex]

cool, those people are objectively wrong

[gleaming laurel]

I don't think so :)

an apt summary of the conversation lol

[gleaming laurel]

well, could you explain why that is "objectively wrong"? Can you explain why not being able to measure it or observe it makes it objectively true that it cannot interact with the brain, instead of just claiming it?

[cedar vortex]

okay, what does a soul do to a brain, how does the soul influence it, the soul needs to communicate with the rest of the body somehow

[gleaming laurel]

I'm not familiar enough with the christian theology of the soul to say, I'm sorry.
But in Mahayana Buddhism (which does not believe in the soul, but does believe consciousness is not produced by the brain, and also believes in rebirth), I can try to explain

If you aren't interested let me know...

[cedar vortex]

i just need literally any answer for the sake of example

"the soul modifies the firings between individual synapses" or something, i really don't care

[gleaming laurel]

oh, do you mean what the soul does physically?

[cedar vortex]

yes

[gleaming laurel]

Again, I don't know what hindus or christians believe the soul does exactly, so I can't answer that specific question, but I can generally answer "how non-physical things interact with physical things", which turns out to be an invalid question

mahayana buddhism specifically is ultimately non-dualist, so unlike christianity it rejects the separation of "mind" and "body" (which would be the non-physical thing interacting with the physical thing), it also rejects monism (unification of "mind" and "body").

But keeping that in mind, in the conventional sense, you look at things in terms of "causes" and "effects". one of the "causes" (in addition to say, sex and womb of the mother), is a "reservoir" of karma that used to be associated with a different body. then after that happens, every subsequent moment is affected by the karmic imprint too. How the karma comes to fruition could literally be anything, because karma is just the idea of cause and effect, taken to a metaphysical, and very hard to observe level. so, it isn't interfering with the neurons, it is one of the causes of the neurons firing (and whatever else is going on in the brain, and outside of it)

[cedar vortex]

and you believe all that is just as feasible as actual science?

[gleaming laurel]

actual science hasn't proven consciousness is caused by the brain, so I don't see any conflict

[cedar vortex]

it literally has

also this very much interferes with literally all of physics

[gleaming laurel]

https://www.sciencedaily.com/releases/2026/04/260406192809.htm
https://en.wikipedia.org/wiki/Christof_Koch

afaik this guy is not a quack, and he doesn't seem to think it's been disproven

this guy even agrees with you, that consciousness is physical, but he does not think he has proof

there is a difference between "this is the leading scientific theory, which I strongly believe, but I have yet to find proof for" and "this has been proven and there is no dispute or question, just like gravity"
It seems clear to me that there is no consensus in the field of neuroscience that "consciousness has been proven to be caused by the brain". If you have evidence that christof koch is the only guy, and the outlier, and 95% of the field thinks it has been proven (=/= believes that its likely to be true), I would genuinely like to see it, because then I have a religion to deconvert from

[cedar vortex]

if you take 2 seconds to look into the Bial Foundation, who published this, then it becomes immediately clear they're spouting nonsense lol

[gleaming laurel]

~~https://pmc.ncbi.nlm.nih.gov/articles/PMC9490228/
here's another one... lead author:

Dr. Wahbeh is the Director of Research at the Institute of Noetic Sciences and an adjunct professor at Oregon Health & Science University.
excerpt:
The assumption within today’s neuroscience is that all aspects of consciousness arise solely from interactions among neurons in the brain. However, the origin and mechanisms of qualia (i.e., subjective or phenomenological experience) are not understood. David Chalmers coined the term “the hard problem” to describe the difficulties in elucidating the origins of subjectivity from the point of view of reductive materialism. We propose that the hard problem arises because one or more assumptions within a materialistic worldview are either wrong or incomplete. If consciousness entails more than the activity of neurons, then we can contemplate new ways of thinking about the hard problem. This review examines phenomena that apparently contradict the notion that consciousness is exclusively dependent on brain activity, including phenomena where consciousness appears to extend beyond the physical brain and body in both space and time. The mechanisms underlying these “non-local” properties are vaguely suggestive of quantum entanglement in physics, but how such effects might manifest remains highly speculative.~~

[gleaming laurel]

yeah, I'm not basing my argument on the publisher though, but the scientist, christof koch, who said it. the publisher is just reporting it. Unless you think they are misrepresenting/misreporting. I will look for another source

[cedar vortex]

they look to be misrepresenting in order to push a specific narrative, it's very common to go "hey look at this reputable science guy he agrees with us" by manipulating and misrepresenting things they said, creationist thinktanks/propaganda mills do it all the time

[cedar vortex]

I clicked on the first listed author and they have another study claiming telepathy is real

so not the most reputable source lol

yeah this is all quackery

not actual science

[gleaming laurel]

which one?

[cedar vortex]

https://pubmed.ncbi.nlm.nih.gov/37709571/

this website just lets anyone post anything it looks like

like i said, spirituality is antithetical to science, anyone who says otherwise has either been duped or is acting in bad faith

especially in this regard

im getting tired of this discussion

it's been hours and i only got like 3 hours of sleep

[gleaming laurel]

that source doesn't claim telepathy is real, but agreed, these people aren't neuroscience professors, but professors of adjacent fields

[cedar vortex]

also, you're lending credence to random laypeople and preachers who like to think they know better than literally the entire body of neuroscience and physics as a whole, which is making me angry

[gleaming laurel]

https://www.fens.org/news-activities/fens-and-societies-calendar/meeting-event/15th-behind-and-beyond-the-brain-symposium
Ah, it's from BIAL hosted the conference. Koch is not a random a layperson, unlike the random people in the paper. and he disagrees with non-physicalism. but sure, maybe he got paid off by BIAL to say random things?

The Bial Foundation was created in 1994 by the Bial pharmaceutical company together with the Council of Rectors of Portuguese Universities. Bial Foundation’s mission is to foster the scientific study of the human being from both the physical and spiritual perspectives.
Is it all nonsense? well, not obvious to me. it has reputable origins at least

Back to the bet between Koch and Chalmers: They agreed that, for Koch to win, the evidence for a neural signature of consciousness must be “clear.” That word “clear” doomed Koch. “It’s clear that things are not clear,” Chalmers said, and Koch, grimacing, concurred.
Here's an article from scientific american in 2023. not a journal, but I don't believe they publish total nonsense
https://web.archive.org/web/20230628014605/https://www.scientificamerican.com/article/a-25-year-old-bet-about-consciousness-has-finally-been-settled/

that pubmed article isn't meaningful, sorry for that. But Koch, who is very much a neuroscientist (and who agrees with physicalism), seems to think its not a settled debate. which is my entire argument, you literally have no proof, its anti-scientific to claim you have proof when you have none, regardless of how ridiculous alternate theories seem to you (and many of them seem totally ridiculous)

Koch seems to have quite a bit of citations (https://scholar.google.com/citations?user=JYt9T_sAAAAJ&hl=en&oi=ao), and both CalTech and Allen Institute are not quackery institutes

o lape pona

[gleaming laurel]

one angle you could attack him is, his pet hypothesis is IIT , which is very controversial and even called psuedoscience by half the field, and defended by the other half: https://web.archive.org/web/20231003174729/https://www.theatlantic.com/technology/archive/2023/09/integrated-information-theory-consciousness-scientific-explanation/675503/ (https://i.readforyou.life/article/143813), https://en.wikipedia.org/wiki/Integrated_information_theory#Reception

In a March 2025 Nature Neuroscience commentary titled "Consciousness or pseudo-consciousness? A clash of two paradigms", proponents of IIT listed 16 peer-reviewed studies as empirical tests of the theory's core claims.[40] A commentary in the same issue by Alex Gomez-Marin and Anil Seth, titled "A science of consciousness beyond pseudo-science and pseudo-consciousness", argued that, despite current empirical limitations, IIT remains scientifically legitimate.[41]
Though in the end, it doesn't seem fair to call is psuedoscience (again, this is still a physicalist position)

Anyways, you can compare it to creationism not being able to be deunked until a few hundred years ago. We know now its false. But until the evidence is there, you cannot claim to have proven it to be false because of vibes. you've tried to cite exactly 0 people, and 0 proofs. if you say "consciousness has been proven to come from the brain", the burden of proof is on you. No one needs to prove an alternate explanation, you need to show proof for your position to make that statement

[glass iris]

Well, there's no way in the world I can read back all of that (though I will - but I gotta get to Mass soon, lol) but thank you for taking the question seriously.

I'll just add a couple of things:

• the brain is profoundly involved in consiousness - who could deny it? the example of brain lesions is a clear piece of evidence - but it is a radical position to say it causes consciousness by itself. One possible position is that it is responsible for consciousness as part of a system that includes the body, nervous system, senses, gut, etc. etc. The idea that a brain in a vat could think, or that a tech bro could come back after freezing his brain, is absurd.
• But the hard problem of consciousness (Chalmers, Nagel) is still hard. We just don't know how we get in principle from matter doing stuff that can be objectively measured, to having a sense of being like something, introspection, subjectivity, intentionality. That is a chasm that science hadn't crossed and, I think, doesn't know how to cross. I think this problem may be unanswerable within science (as Nagel first suggested in the "Bat" paper).
• What is eliminative materialism meant to do with the hard problem? We could follow Dennett, and argue (though "argue" is too strong a word) that consciousness is some kind of illusion. But who is having the illusion? The criticism from fellow materialists was pretty harsh - we're not solving the problem that way.

What else could be going on? I don't know. But here is a possibility (which I do not believe - I am offering it as a hypothesis compatible with evidence). Consciousness resides in something non-physical - not necessarily a "soul," but perhaps an aspect of matter that cannot be quantified by science. Let's call it C. C is what has the experiences and sense of self; but it can only do so by interaction with the brain; its "states of mind" give rise to physical changes in the brain. The brain is, as it were, a tool for it, and it is impaired if the brain is impaired. But it is not caused by the brain.

This is a tool-relationship. Suppose we were to watch someone hammering a nail. And we said, "Hammers are entirely responsible for hammering." We could argue that when the hammer is broken, then the hammering is also impaired etc. But in fact we know that a hammer is necessary for hammering, its lack of function impairs hammering, but it is not sufficient for hammering. The brain could be like that.

(One reason I reject this is because of my phenomenological convictions. Our being is a being-in-the-world and a being-in-the-body; I'm very uncomfortable with any theory of soul that suggests that it could be somehow separable, even in principle, from the body. But the same phenomenological convictions convince me that the brain does not "make consciousness" by itself)

[glass iris]

And (one more thing) science is a very small part of the human being's grasp on the world. Scientism, which insists that everything in the world is explicable using the methodology of science, is a bad position to take! A poem has a different kind of grasp on the world, which is a good grasp, and is not reducible to "science" (as if there were, by the by, any such one thing as "science").

[cedar vortex]

it's not radical, this is the widely accepted and highly corroborated stance

if you could isolate a brain from all sensory input you would still be conscious, why is the rest of the body relevant?

you're not suddenly unconscious in a sensory deprivation tank unless you fall asleep

does it have a major impact on your psyche? yes
is it a requirement for consciousness? no

you can't freeze a brain because the brain would die, you can't store a brain in a vat because the brain would die, thinking otherwise is absurd

[glass iris]

That's a non sequitur about the tank. Your consciousness has been formed by being part of a world that is always already there, from the moment of your first breath. I do believe that a human being kept in a tank for its entire life, birth to death, would not attain consciousness (but we will never know for sure!)

[nova thicket]

that would be unethical anyways

[glass iris]

And I am not denying that the brain is the one crucial thing on consciousness! But it has not been "corroborated" that it causes consciousness, largely because consciousness as a phenomenon in itself is something that science does still not have a grasp on.

And may never have a grasp on.

[glass iris]

That's my point.

[high kayak]

the hormones and nerves in the rest of your body is still an important part of you and your consciousness and experience of reality, you're not just your brain, you're your whole body

[cedar vortex]

i mean, so do drugs

things that affect your brain will affect your mind

your body has major influence over the brain

[cedar vortex]

you're the one who brought up storing brains in vats and freezing them 🤷

[cedar vortex]

here i'm using "consciousness" as a synonym of "mind", and "conscious" as an antonym of "unconscious"

there are other definitions that I don't think are meaningful

[gleaming laurel]

being in a sensory deprivation tank does not actually cut off your senses, most obviously your internal senses (and even with a theoretical magical total sensory deprivation tank, is a lack of a sense a sense?)
Plus, you have the "initial conditions" of your body, which is at least to some extent the result of previous inputs

at the current state of science, and quite possibly any future state of science, it isn't possible to show that an isolated brain can produce consciousness, or to show an isolated brain at all. even supposing the non-physical is known to be false, which itself is out of the realm of science

[remote dove]

define what you mean by consciousness?

[gleaming laurel]

the best I can give is "awareness", the kind of which I and you are experiencing right now. science and philosophy certainly don't have any widely accepted definition

[remote dove]

If it's some metaphysical "awareness" I "experience", I can't be sure anyone else is conscious, since I can only observe their physical actions. I can't even be sure I was conscious 5 seconds ago - I could remember that I thought I was conscious 5 seconds ago, but there's no reason for me to trust that memory. Since there's no observable difference, it's reasonable to assume the simpler alternative, that everyone else and me 5 seconds ago isn't conscious. And if it's something that only applies to me at this particular instant, I don't think there's really any point in talking about it.

[queen patio]

-# last thursdayism

[gleaming laurel]

in the ultimate sense I agree with you, in the conventional sense, as an unenlightened individual, if people don't assume I'm conscious, many things tend to happen that my awareness experiences badly. I don't like experiencing bad things (duhka), and I would like other conscious things not to experience these bad things. so that is why, mi la, it is interesting to discuss consciousness

[cedar vortex]

this is just straight up false

[cedar vortex]

yeah sensory deprivation tanks aren't super accurate but i still think it's a somewhat good example

if your brain lost all input from your body, it would probably be psychological torture assuming the body can be kept alive lol

there would still be someone inside to be tortured though

[cedar vortex]

memories are generally pretty accurate, especially short term memory, and humans typically stay conscious for a long period of time so it's safe to say you were indeed conscious 5 seconds ago

[high kayak]

the argument is that you're unable to distinguish between having been conscious 5 seconds ago, and having your brain set up in such a way that makes you think you were conscious 5 seconds ago

like, using a video game as an analogy, it's impossible to tell whether the game has been played for 15 minutes, or whether the program was just booted with all the variables set up identically to as if it had been running

[cedar vortex]

i feel like this is related to "i think therefore i am"

this also applies to a materialist view of consciousness too

[remote dove]

True, good catch about the memory part

I still think the conclusion holds tho, just that the thing I can't trust is past me's thought rather than my memory of that thought

[cedar vortex]

there's not really any quantifiable difference between my brain and others' brains

i can pretty much say for certain that yes they are conscious

i feel like it would take a massive ego to think you're the only conscious one and everyone around you are mindless robots

also, if others weren't conscious then the concept of consciousness would be absolutely meaningless, since other people came up with that

[remote dove]

I don't think that, I just think that's a natural conclusion if it's this "mysterious nonphysical awareness I experience", which the argument was conditioned under
If it's physical, then sure other people are conscious. But I don't see how "other people have similar neural activation patterns as I do" leads to any philosophical conclusions.

[cedar vortex]

i don't really care about philosophy so this isn't something that's particularly important to me

it leads to the conclusion that what's going on in my head is very similar to what's going on in other people's head

[gleaming laurel]

are you saying someone has come up with such a theoretical, scientifically-rigorous experiment? Or that you think its false that its not theoretically possible? If the second, then sure, fair enough, but you can understand why that is not convincing to me

[gleaming laurel]

I think I agree..? there would need to be someone inside for them to be able to be tortured. and if I had to guess, it does seem torturous

[cedar vortex]

I've already explained to you the reasons why we know consciousness is carried out by the brain

i'm honestly so tired of this discussion

[gleaming laurel]

yeah, let's move on from that (sorry). but obviously, I have listened to that many times, and do not think that is scientific evidence, and explained that too. But my response was about an experiment of an isolated brain being not possible... which is not the same. and clearly no one has come up with such an experiment so far

[last umbra]

(unrelated to tp but) imo, the only job that i can ever forsee AI ever doing is CEO and it would still be the worst one

[cedar vortex]

i think a study was done a while back and they found that CEO was the job most easily replaceable with AI

[last umbra]

oh fr

[stable lantern]

i mean ceos don't do anything so it's accurate

[fleet gate]

we know they're just there to extract wealth. however, they're humans and hence parts of society (albeit parasitic)

Political economist Joseph Schumpeter asserted that a liberal capitalist society contains the seeds of its own downfall as it breeds a class of intellectuals hostile to both capitalism and liberalism, though without which these intellectuals cannot exist.

note that some ceos are among these people

the so-called "class-traitors" of the bourgeoisie kind will be pivotal in the success of the revolution

china has no business doing what it's doing in general (mass surveillance, government brutality, etc.), but from a purely marxist perspective, it's doing a good job using the bourgeoisie for the cause of their revolution

[lone parcel]

:airplane_arriving:

*toki pi ijo *“ilo pi pali nimi” ** li kama tan tomo #toki-ale · o lukin e open
||conversation about “ilo pi pali nimi” continued from #toki-ale · read the beginning||

[latent vigil]

did the plane crash or smth💀 the convo did not continue here

[high kayak]

people love to do ai discussion in #toki-ale

i feel like this channel should be explicitly for ai discourse, cause often i see people get told to go here, and they go "oh no, that channel is for discussing ai generated toki pona, not ai stuff in general"

[stable lantern]

i agree

well i mean "discourse" is a little accusative maybe AI discussion

but like yeah

[high kayak]

meant discourse in a neutral way, like series of discussions or whatever
but yea, the point being, it'd be nice if this channel was explicitly for all ai discussions, instead of just toki pona generated texts

[stable lantern]

looks like i can change it it seems

uhh ok how's "(CW:AI) discussion of LLMs and other forms of text-generating AI"
i'm pretty sure there was another channel for image AIs but i don't think people use that one. at all. so we could also go all in and make it generative AI in general, i'm open to whatever tbh

[high kayak]

might be good to forward this to #wile-kulupu for more input

[stable lantern]

true

[glass iris]

CW:AI: Generative AI and toki pona; and general AI discussion

[stable lantern]

pona

[glass iris]

CW:AI: Generative AI and toki pona; and general discussion of any AI

Actually I kind of prefer @stable lantern 's version, now that I read back. Maybe I'll change it again later.

[severe dagger]

ooo the title changed

[glass iris]

(CW:AI) Discussion of LLMs and other kinds of text generators ("AI"), especially of toki pona

We've changed the title to reflect (following discussion in #wile-kulupu ):

• This thread is dedicated to the discussion of Large Language Models or generative AI, and any other so-called "artificial intelligence," including symbolic/algorithmic AI, such as translators, parsers, grammar checkers, etc

• Discussion about the computer generation of toki pona is particularly encouraged here. This topic is not usually allowed in the rest of the server, unless in spoilers (see the notes whyai and whynotai)

• General discussion of "AI" and philosophy of mind is allowed here, but should be kept civil and respectful.

• AI-generated images should not be shared here, unless they are very closely relevant to the generation of text.

[crude agate]

Engaging with AI-created content is generally discouraged on ma pona pi toki pona (see {{whyai}}). However, there are two threads where you can do so safely:

• Discuss LLMs and their ability to generate Toki Pona text here: https://discord.com/channels/301377942062366741/1162756469771472988
• Discuss images generated by AIs and how they relate to the Toki Pona community here: https://discord.com/channels/301377942062366741/1319684053904916591

6d7c06 AI threads

[glass iris]

lukin la ale li pona

[unique reef]

{{whyi}}

{{whyai}}

[[whyai]]

[lone parcel]

https://sona.pona.la/wiki/whyai

[unique reef]

There we go

Nope. Wow this thing is weird

[fleet gate]

/peek note:whyai

[unique reef]

Eh it's fine

[fleet gate]

wdym, aren't you trying to view that?

that's how you do it

[unique reef]

It's fine as in I don't need to view it

Just cuz I was trying to for a second out of curiosity doesn't mean I need to

I generally understand and agree with the points it would make anyways

So yeah, it's fine that I can't figure it out

[crude agate]

Why we ask to avoid or spoiler AI-generated content

ma pona pi toki pona is a space for members to communicate and learn Toki Pona together. Using AI text instead of engaging with actual speakers of the language is disrespectful to the community, and it defeats the purpose of language learning and practice.

Server members might prefer to avoid seeing AI-generated content, whether images or text, for a variety of reasons. Some have ideological objections to machine-generated work, some find it unsettling or off-putting, and others can be easily triggered, disturbed, or experience dissociation from artificial content.

Please be considerate of your fellow community members by avoiding posting AI-created content, or at least using spoiler tags and content warnings on AI-created content that you post.

7097d2 whyai

can i use AI for toki pona translation?

well uhh you can but you probably shouldn't! here's why (not mentioning environmental impacts of AI because that's suited to a different larger conversation):

• beginner mistakes
  an AI is prone to a lot of calqueing, large vapid noun phrases and other strange quirks that experienced speakers work past, like omitting clear or unimportant details and/or spreading out an explanation around a couple sentences, which is usually the method that will get you the most wiggle room for mutual understanding. if you yourself are a beginner, you might not have the tools to catch these artifacts, and your AI-generated translations will likely be ungrammatical or not make any sense.

• limited knowledge
  while this becomes less of a problem with time, right now it's still clear that most AIs don't have the tools needed to utilize toki pona very well, nor can they distinguish really good well-formulated toki pona with less good toki pona. it doesn't really have a way (nor motive, we are a small community) of improving.

• rigid thinking
  toki pona's philosophy and "efficiency" (for lack of a better word) makes it hard for an AI to comprehend certain ways to handle something unless it's spelled out to it, like intent or wordplay. most AIs won't understand "mi wile len" as "i would like to stay private", because len covering "cover, hide" isn't something they're experienced with (true story!). this essential facet to toki pona is almost entirely cut off

toki pona's inherently human! it's incredibly satisfying to work with and figure out! it's not really worth it to just shove it all to an LLM that can only kinda read & write it. just try to do it yourself, or ask someone for help! i guarantee you it'll be miles better than any AI can do.

ef098f whynotai

[midnight sapphire]

Some of these points are a little outdated but the most important point for me is: if you make AI speak the language for you, where's the intrigue? You're not learning anything, you're not actually speaking the language yourself. You're relying on an AI (which is prone to mistakes) to speak for you and/or translate other people's messages back

The only way I've found to correct (toki pona) mistakes AI make is giving them the correct answer and you can't do that if you don't know it yourself

And if you can there's no reason to use AI because you know how to do it yourself

[fleet gate]

the most realistic use case for large multimedia models here is translating toki pona to english. even when the output is bad, it still removes most of the mystery for curious non-tokiponists. most people won’t go through the awkwardness of asking a human, even though i’d gladly do it for free, or even beg people not to rely on lmms for toki pona

people treat every “pali” like it’s “musi ala” and throw it at large multimedia models. that’s dangerous, not because every task is sacred, but because effort, judgment, and interaction with other people are how you build taste, empathy, and a real grip on the world. work that fits your character transforms what you enjoy and how you understand others. hand too much of that away and you don’t just become more “efficient”, you become less grounded

i get why people with limited time, money, or social confidence use lmms for tedious labor humans usually won’t do for free. but for toki pona specifically, they’re honestly not useful at all. and as someone whose own profession could easily get eaten by lmms, i still think it’s fine when a kid uses one and gets amazed. everything feels magical to kids. what’s embarrassing is some executive with 3 yachts treating lmms like divine revelation while cutting jobs and squeezing more out of workers lmfao

like yeah, if an lmm genuinely does my job better, there’s a conversation to be had. but right now it doesn’t. it doesn’t have the opinions, design choices, or consistency i deliberately put into my work. and even if it did, i’m not going to pretend effort alone entitles my work to stay a profession forever. in a saner world, useful work becoming a hobby could be beautiful

the problem is the world around it. workers are told their skills can become “just hobbies” the second they stop being profitable, while the value they produce keeps funding the stranger, costlier hobbies of the people above them. when survival depends on staying marketable, people end up needing similar tools or labor to stay scarce a little longer, even if nobody really wants to live that way

[lone parcel]

:airplane_arriving:

toki li kama tan tomo #toki-ale · o lukin e open
||conversation continued from #toki-ale · read the beginning||

[hearty blade]

hmmm alrighty

[glacial mason]

I had this interaction with my mom last month. I know she doesn’t know Toki Pona at all and it was really frustrating.

I visited her for Mother’s Day and she confessed she used LLM to write it which was vindicating but didn’t take away the frustration. She had wanted it to say “Christ is risen from the dead” since that was Russian Orthodox Easter — in fairness if I had realized what that day was, I would have probably gotten it

Translation of Russian
Me: it seems that what’s written is “the priest Kolisu is traveling due to death”

Mom: it seems that my translation to Toki Pona is better

[meager blaze]

nasa mute

jan li sona ala e toki li pilin sama ni: ilo li sona pona tawa jan

[high hatch]

I'd rather have a bad translation than one made by ai (in this context)

[glacial mason]

At least it made it easier to understand how this message even happened in the first place. It kind of freaked me out that she had seemingly learned enough to make a complete sentence without telling me anything but also that the sentence is so confusing and cryptic at the same time

If she had actually embarked on the journey I would have expected something more like… “Toki!” lol

[cedar vortex]

What does she mean that her translation is better?

also, is Jesus even a priest? I've never heard him described as such, but jan sewi would accurately describe Jesus

[glacial mason]

He was probably considered a Rabbi

but I can't know that for sure, just a guess

[cedar vortex]

either the toki pona to russian or the russian to english translation is a little iffy, does "священник" mean any sort of holy person, like jan sewi does?

[glacial mason]

In the context of Jesus in this particular translation attempt, the noun phrase was surely intended to be a different kind of "holy human", not "priest"

But I had no way of knowing she was referring to Jesus specifically because Kolisu didn't sound like anything to me haha

[cedar vortex]

Toki Pona Wikipedia uses "jan Jesu"

tokiponized from יֵשׁוּעַ

[glacial mason]

But it's "Christ is risen", not "Jesus". I actually mentioned "jan Jesu" to her on Mother's Day and she was very unsatisfied by it

[high kayak]

yes* (in modern terms)

[cedar vortex]

what's the difference? it's the same guy

[glacial mason]

that's how I feel 😂 but I didn't push it

[cedar vortex]

nasa

[high kayak]

Jesus (or Yeshua) is the persons name, Christ is the title

it's like saying God Save Elizabeth instead of God Save The Queen :p

[glacial mason]

hahahah I love how silly that seems

I think casual Toki Pona taking-a-look-ers are similarly disappointed by the lack of anything that feels like "Merry Christmas" and having to say "I hope your day of gifts and Jesus is good"

[cedar vortex]

yes because it's a set phrase! you can't preserve a set phrase when translating to anything but a super similar language

[glacial mason]

but if you ask ChatGPT to do this, it will try 🤪

[high kayak]

to be fair, i think like every language differentiates between Jesus and Christ
since it's like
pretty dang important theologically

[cedar vortex]

suno ni pi pana suli o pona tawa sina!

[cedar vortex]

have multiple people had the title of christ?

[glacial mason]

It means "the anointed one"

[high kayak]

i mean
yea kinda

[cedar vortex]

yes i definitely know what anointed means *furious googling noises*

[high kayak]

anointing kings, "christening" them, is like, an actual thing people did

and since jesus is the king of all humanity, he was anointed the same way kings were

[cedar vortex]

(thought id go with the AI slop description as it is on theme for the channel)
did someone like,, cover Jesus in oil?

[high kayak]

https://en.wikipedia.org/wiki/Anointing_of_Jesus
yes

[glacial mason]

I've been anointed numerous times in a Russian Orthodox church in the past

It is more like they smear a tiny amount on your forehead with a thumb

Sometimes they also put it in other exposed spots of skin

like tops of the wrist and stuff

[cedar vortex]

religion is absurd oh my goodness

the more i learn about religion the less i understand it

[high kayak]

so is a lot of culture, mythology and folklore
the world doesn't have to make sense
that's what makes it fun and interesting to learn about

[glacial mason]

I think puwa is exasperated pona not exasperated ike

[high kayak]

also christ is just the greek translation of messiah
and messiahs are a bit thing in judaism, christianity and islam

[cedar vortex]

ala, ijo sewi li nasa ike tawa mi

[high kayak]

it's a bit reductive to treat it as just a name :p
at least if you're religious

i don't really care, but i get why people would

[cedar vortex]

jesus is greek?

[high kayak]

well no, but a lot of the new testament was written in greek, and he lived in a greek speaking context

he seems to have had at least basic knowledge of the greek language
but ofc, the new testament was written after his death, so his personal knowledge isn't super relevant

[cedar vortex]

sona

[cedar vortex]

to clarify i have nothing against religious people unless they are using their faith to cause harm to others

[high kayak]

ye you're fine dw

i'm also pretty anti-theistic, i feel that if you believe spiritual/metaphysical things can justify actions, instead of (attempted) logic and rationalism, you can kinda justify anything to yourself
which is bad i think

i think it's valuable to think about why you believe in what you believe, and what reasons you have to believe it

in any case, jesus is a very fascinating historical figure

[cedar vortex]

agreed

[high kayak]

https://youtu.be/82vxOBbYSzk
this is a wonderful video i highly recommend to nerds like me
it talks about jesus from a historical context, instead of a religious one
how he was one of many zealots leading a spiritual and physical rebellion against roman occupation, and how the apocalyptism of the time shaped his personal apocalyptic views of good and evil

it's neat

it's a completely new way to think about him and i think it's extremely interesting and valuable

[cedar vortex]

sitelen sona

[lone parcel]

:airplane_arriving:

toki li kama tan tomo #toki-ale · o lukin e open
||conversation continued from #toki-ale · read the beginning||

[vapid eagle]

The thing that worries me is that GenAI models are increasingly very good at "understanding" toki pona. After another discussion about this topic where I asserted the AI really couldn't write coherent toki pona I decided to ask Gemini to translate some stuff in both directions and it essentially got everything correct

[shell mango]

it was never wrong when i used it

[severe plume]

it's harder to fight against slop when people are finding slop to be useful

[vapid eagle]

And the mistakes it made were extremely plausible. Small grammar issues or awkward phrasing

[shell mango]

the thing is, how are AIs getting better at toki pona?

[vapid eagle]

TOK-->EN la sometimes it wouldn't entirely get the context and sometimes it would "overthink" and try to find more meaning than was already presented, but the mistakes were extremely small

[severe plume]

better training data, and better learning resources

[shell mango]

nn lon,,

in fact i was shocked how good it was in a span of a year

[vapid eagle]

yeah I hadn't looked at toki pona and ai since like 2 years ago

[shell mango]

i remember using meta a year ago

[vapid eagle]

it's it's scary on an existential level I can't lie

[shell mango]

pfft it couldn't even get a single sentence right

[severe plume]

we're going to be increasingly encountering learners who use machine translation

[vapid eagle]

Nothing is safe from this crap anymore.

[vapid eagle]

YEP

[severe plume]

and toki pona won't be a slop free part of the internet anymore

[vapid eagle]

Worse yet I think someone is going to try and make a toki pona ai content farm, or submit a bunch of articles to Wikipesija, or make toki pona AI music

[vapid eagle]

The second one especially worries me

[severe plume]

ugh yeah

imo wikipesija was already inherently a lost cause because people are excited to translate while they're still learning hehe

[vapid eagle]

Endless horrible labrynth of awful things that I now have to be vigilant of

[severe plume]

like even without slop it's gonna be filled with human learner mistakes

[shell mango]

heres me chatting with meta a year ago :p

[severe plume]

but it's nice that it's human

if you want human-written english then you add "before:2022" to your google search

if you want human-written toki pona then you add "before:2026"

if you want human-written sitelen pona then you're still safe :D

[vapid eagle]

Yeah the issue is that there's no good way to test "competency" (not that that's an inherently good idea, but I do think toki pona is unique in just how vastly learners outnumber experienced speakers)

[severe plume]

i mean you proofread

but that takes human effort and i'm personally not interested in doing it for the wiki

[vapid eagle]

exactly

[severe plume]

but it's easier to find proofreaders for projects that experienced speakers are interested in, like videogames and song translations

[vapid eagle]

i think there are maybe like a thousand toki pona speakers at "fluent" level and many times more than that who are learning

[severe plume]

yeah

[shell mango]

a thousand?

[vapid eagle]

which is just like unheard of in any language

[vapid eagle]

I think that's an upper limit. The other thing is that a good chunk of those people might not be super active

[shell mango]

aa, lon

[vapid eagle]

even Esperanto has a way larger dedicated community and a non-insignificant chunk of people who learned it as a semi-native language

[shell mango]

most learners think they're already good at the language because they know all the words

🤔

[vapid eagle]

yeah people overinflate how much they know

it's a problem particular to this language I think where it appears superficially simple and so you think that if you check off a certain list of boxes and learn a certain group of things that you're a capable speaker

[shell mango]

they take the 30 hours too serioujsly,,

[vapid eagle]

it took probably hundreds of conversational hours to get me to even remotely comfortable with calling myself fluent

and even more writing/messaging/translating as well

my main criterion for "fluency" is that you need to be able to explain things that are more technical than typical day to-day occurrences and events

[shell mango]

i guess you could get quite good at the language with 30 hours by dedicating an hour a day talking to people

30 minutes of studying and 30 minutes of trying to talk with people lol

[vapid eagle]

you might get to enough to hold basic conversations but you will not be able to grasp higher-order concepts

[shell mango]

where do you draw the line at more harder concepts?

[vapid eagle]

going beyond "where do you live, what are you eating, where are you going, how are you feeling"

[shell mango]

aa

[vapid eagle]

can you talk about the mechanics of a video game you like? can you explain what happened at the latest sportsball game? can you have a discussion about Christian trinitarian theology? can you talk about global geopolitical events or your political views beyond basic partisan terms?

I have talked about all of these concepts before

[meager blaze]

ni

[vapid eagle]

obviously you can't know everything. all of those are concepts that I have at least some level of interest in and hence I'm able to explain them well

but if someone was talking about something I didn't know much about and I could understand them and learn about a new thing from them that would be a good indicator that both of us are pretty fluent

[meager blaze]

ilo toki pi kulupu Kuko li ken ala ken sona e toki pi ijo ni

ken ala anu seme

mi alasa kepeken ilo ante pi nasin toki nasa ni la ona li ken sona pona e toki wawa li ken ala sona pona e toki pi wawa lili

mi sona ala e tan

[severe plume]

i think ken ala yeah
it requires reasoning about high level concepts, and being consistent over time
which LLMs are bad at

[vapid eagle]

a favorite of mine is "if you were at a restaurant and could only relay what you wanted to eat through a toki pona speaking interpreter would the food that showed up at your table be what you were trying to get"

[shell mango]

Even simple math is already somewhat complex to explain in toki pona

with multiplication and division

[stable lantern]

i mean we are kinda spoiled on being able to write math in a compact manner tbf

[severe plume]

when comparing modes of transportation, google translate won't think to translate car as soweli and boat as kala, because that's not in its training set

[vapid eagle]

this is still probably a little short of true fluency but I think most learners probably would not be able to do this

[cedar vortex]

who's translating car as soweli?

[severe plume]

just me haha

[shell mango]

level 100 tokiponists

[stable lantern]

mathematicians had to write poems for all their formulas before the math notation we know today arrived

[vapid eagle]

Going to more consciously try to use soweli ilo now

[cedar vortex]

this is actively confusing and i wouldn't know what you're on about lol

[stable lantern]

which isn't much different from doing so in toki pona except there also isn't a "word" for multiply or divide or whatever

[vapid eagle]

okay but what would you call a plane

[meager blaze]

tenpo suli la mi wile kama musi ni: mi toki kepeken toki Inli la jan ante li toki pona. taso jan li lon insa pi mi tu li ante e ijo tawa toki ante. mi wile kama sona e ni: ni li ken ala ken. tenpo mute la mi pilin nasa lon ni: jan ante li toki pona la mi toki inli

[vapid eagle]

in toki pona

[cedar vortex]

tomo tawa sewi

[severe plume]

that's why i clarify with context, in toki pona

[stable lantern]

i love my robotic horse

[shell mango]

ilo tawa sewi

[vapid eagle]

okay fair enough

[shell mango]

anu ilo waso

ilo waso suli,,

[cedar vortex]

kon instead of sewi also works

[vapid eagle]

most of the higher-level speakers i talk to call it an ilo waso or waso ilo

[meager blaze]

"mi esun ala e waso mi" -mi ale li tawa tenpo kulupu pi ma Nujo la jan wan li toki e ni

[stable lantern]

planes are definitely waso
if boats are kala and cars are soweli what are pipi and akesi

[vapid eagle]

but that is my own selection bias

[vapid eagle]

submarines are kala

boats are questionably kala

[stable lantern]

akesis are jetski

[cedar vortex]

i'm not personally a fan of this nasin

[shell mango]

using animal words except the animal part

[stable lantern]

for most of their airtime planes fly like eagles

[cedar vortex]

i wasn't aware eagles had jet engines

[vapid eagle]

i mean it's an extremely common metaphor crosslinguistically afaik

[shell mango]

i mean it would make sense if the general meaning wasnt an animal

it would make sense if everything was explained as a "thing"

[vapid eagle]

toki pona words have larger semantic spaces

[shell mango]

just sayin :p

[stable lantern]

that's why i said MOST of their airtime

they get off in different ways but when they're in the air and with a current they let their wings keep them up

[shell mango]

if i were introduced to these types of words as "things" then your guys's nasin would make more sense

[cedar vortex]

they WHAT

[stable lantern]

i mean i'm pretty sure

[shell mango]

waso as a "thing with wings"

[severe plume]

when introducing someone to toki pona, one of my favorite spiels is the following

toki pona doesn't have a word for car. but it does have a word for beast, a word for computer, and a word for house.
so when discussing a car, you need to decide for yourself whether a car is more of a beast, a computer, or a house.
and you need to have the empathy to understand your conversation partner when they make a different choice in their own speech.

[stable lantern]

maybe it wasn't eagles and it was seagulls but either way most birds glide

[cedar vortex]

i wouldn't say tomo specifically means house

that's like saying ko specifically means mud

[severe plume]

i wouldn't either but this is for beginners

[vapid eagle]

okay the thing that's annoying me about this is that it IS how fluent speakers speak. advanced speakers don't add stuff like "ijo" to clarify that it's not an actual bird because it's superfluous and doesn't add any actual context because they understand the language better. I'm usually not the type to like make this kind of statement but it's true it's not a preference thing it's just that they are better at filtering through information than you are

[stable lantern]

does waso specifically mean bird

[cedar vortex]

That's even worse! beginners wouldn't know any better

[vapid eagle]

i don't like playing the "these people have spoken the language longer so they know better than you" card
but I also think it's got a level of truth to it

[cedar vortex]

not exactly the same semantic space but id say the english word "bird" is a close equivalent to waso

[severe plume]

i'm not even saying the word tomo, they'll see a dictionary if they do decide to learn the language, it's fine

[stable lantern]

maybe

but waso includes many things that aren't birds even if we're (arbitrarily) limiting ourselves to living creatures

some bugs are waso
some mammals are waso

[vapid eagle]

that's the thing. toki pona doesn't work like this. waso is often a living creature but it doesn't have to be. stop trying to categorize toki pona words to match another language's concepts and ideas perfectly it's not how it fucking works

[cedar vortex]

saying a car is soweli is incredibly metaphorical and only really possible by drawing comparisons. cars aren't literally soweli.
cars are literally ilo and tomo though, they directly fall under the definitions of these words

[stable lantern]

hence why i said arbitrarily

[vapid eagle]

sorry if i sound really angry i'm not mad at any individual here

[stable lantern]

it's arbitrary to limit it to living creatures

[shell mango]

is this about ai anymore 😭

[vapid eagle]

no lmao

[stable lantern]

nope!

[cedar vortex]

saying its as much of a stretch to call a car tomo or ilo as it is to call a car soweli is disingenuous imo

especially towards something aimed at beginners

[severe plume]

again i'm not saying the word soweli or the word ilo

[cedar vortex]

what else do you mean by "the word for beast" or "the word for house" then?

[vapid eagle]

i also think expecting perfect internal logical consistency from a language is silly more generally

[severe plume]

i'm introducing them to the fact that toki pona grammatically works like "is a hotdog a sandwich"

[stable lantern]

this does not contribute to the conversation almost at all but gogoat is a motorcycle

https://tenor.com/view/gogoat-gogoated-wait-i'm-goated-wait-i'm-gogoated-goat-gif-15463623129408200215

bike

most of its names in other languages include "vehicle" because it's a motorcycle

[severe plume]

obviously i'm referring to those words but i'm not introducing those words. because i'm not saying them. when i do introduce those words by saying them, i give more elaborate definitions

[stable lantern]

clearly via the transitive property if gogoat is a motorcycle and gogoat is a soweli then motorcycles are soweli

[shell mango]

are wheels soweli

[stable lantern]

wait shit i forgot about cyclizar

motorcycles are also akesi

[severe plume]

wheels are soweli

[cedar vortex]

why are you giving an example as if you're talking about real toki pona words but giving semantic spaces that don't actually align with toki pona words?

[shell mango]

ale tawa li soweli

[stable lantern]

i feel like wheels are noka soweli if we're talking vehicles

[shell mango]

if thats the case then the tip of a ballpoint pen could be soweli

[stable lantern]

you can't move around on just one wheel

[severe plume]

because i'm communicating how toki pona works with fewer than 150 words by introducing the idea that toki pona functions like "is a hotdog a sandwich" and requires empathy to communicate

[stable lantern]

they need a body

don't say unicycles those have bodies

[shell mango]

unicycles those have bodies /musi

[meager blaze]

pilin mi la tomo toki ni o lon: jan li ken toki musi e nasin pi toki ni li ken kama sona e nasin toki. ni li ken ala ken

[cedar vortex]

toki pona doesn't have a word for car. but it does have a word for beast, a word for computer, and a word for house.

[severe plume]

i think that's sona-musi

[shell mango]

this is quite generative

[meager blaze]

ken la ni li lon. taso mi lukin pon ala e tomo ale pi kulupu ni

[meager blaze]

sona

[stable lantern]

aren't we all just generating text

in our hearts

[cedar vortex]

on-topic image to distract yall

[meager blaze]

ken la toki pona lon li jan ni: mi kama sona e ona lon tenpo

[shell mango]

silver koi circle

[cedar vortex]

that was a genuine attempt to spell toki pona btw

[vapid eagle]

i wonder why cars are measured in horsepower, or why car brands are named after animals, or why we "ride" in a car just like we "ride" on a horse. but no this has no categorical relation and if you think cars are related to soweli you're silly and ridiculous and grasping at straws

[shell mango]

roki pond

[stable lantern]

boats are kala because lapras is a boat

[shell mango]

i call it HP /musi

hp just means health in most video games

[cedar vortex]

horsepower as a unit is weird and i don't like it

[vapid eagle]

the car is literally the direct successor to the horse in personal travel. the argument that there's no link is kinda goofy

[shell mango]

i think hp just lost the meaning of "horsepower"

[stable lantern]

car is etymologically related to horse (adding)

[meager blaze]

pilin mi la mi wile awen wawa e toki pona mi lon nasin ni: mi kama sona pona e nasin toki nasa. n. taso nimi "nasa" la mi toki ala e ike nasin li toki e pona nasin e suwi nasin

[shell mango]

and the acronym is its own word now

[vapid eagle]

yep

[meager blaze]

toki pona mi li awen sama mute lon tenpo suli

[cedar vortex]

calc stands for calculator, I'm just using slang

[meager blaze]

mi alasa ala lukin e nasin toki mi li alasa ala pona e ona li kepeken nasin toki pi jan poka taso

ni li ike ala! taso awen la mi wile musi kepeken toki

[vapid eagle]

this lowered my cortisol a lot actually

[shell mango]

whens the last time you've heard of people saying horse power

🤔

[vapid eagle]

if you're around cars ppl still do plenty

[stable lantern]

ok but cars evolved out of horse drawn carriages

is the carriage, while it's on the horse, the horse

do they become the same unit tokiponically

[shell mango]

aa right

cars

[meager blaze]

mi kama sin lon toki open la ilo mi li ken ala ante e toki pona

[cedar vortex]

fairly often when talking about cars

[shell mango]

lowkey haven't heard of a car for a long time,,

[vapid eagle]

this conversation has radicalized me so hard actually

[meager blaze]

nasin seme la mi o open e "nasin sona"

[vapid eagle]

fuck you a car is a soweli

[cedar vortex]

are you the one with the blåhaj pfp on youtube, i like your music

[shell mango]

😭

[vapid eagle]

ona li mi

[stable lantern]

is the carriage soweli if there is no soweli to tawa it
and if so is the soweli of the car the motor

does the motor soweli the car

[severe plume]

that's a poki
no it's not a tomo
i will not be taking questions

[cedar vortex]

i have olin li seme ON LOOP

[vapid eagle]

pakalaaaaaaaaaaaaaaaaaaaaaaaaa

[shell mango]

steam is coming out of someones ears

[meager blaze]

taso nasin seme la mi ken ni? ken la mi o lukin e ijo ale o alasa e nimi musi nasa tawa ona o alasa sona e ni: ona li sama ijo ante seme

ni li ken anu sem

[stable lantern]

i don't have a soweli in this race btw i'm just saying shit

[vapid eagle]

sina la

[severe plume]

i don't wanna attempt this if it's deemed disrespectful but i am curious how google translate does with the lyrics to "olin li seme"

[shell mango]

jan lili, o utala ala e mi

[meager blaze]

waso Keli o, nasin seme la mi ken kepeken nasin pi ilo toki nasa

[severe plume]

idk

[vapid eagle]

mi la sina ken

[severe plume]

the more reliable method is to go to gemini directly

or the other LLM of your choice

[vapid eagle]

lon la mi ante ala e nimi kalama tawa toki Inli. ken la mi awen ni ala.

[stable lantern]

is a pushcart being pushed soweli because now that it can tawa it's a soweli or does the pushcart li soweli e sina

[shell mango]

a soweli is a car because a cat is car

[vapid eagle]

taso mi ken wile sona e nasin ilo

[shell mango]

okay what the fuck how did they appear

[severe plume]

thumbnail A/B tests

[vapid eagle]

mi ken ala kute mute e ona. ona li kama e telo tan lukin mi.

[shell mango]

this is the first time seeing it glitch

😭

[severe plume]

i commented on the youtube but since you're here, god, same

[vapid eagle]

i fear this is the issue that may arise with my toki pona musical vent art

[severe plume]

really fucking good song

[rare pebble]

toki

[vapid eagle]

aghghwhhghhgh sina pona

[meager blaze]

mi sama.

[severe plume]

it's been months since i've cried about her, i didn't think i had any tears left in me

[vapid eagle]

you will not BELIEVE what showed up on my YT homepage

[meager blaze]

kulupu Kuko li lukin e ale

[cedar vortex]

usually its you on my homepage lol

[rare pebble]

AI has always been a thing in google translate but I didnt know that you could prompt inject it

[severe plume]

depressing to think that we can't speak toki pona to hide information from The Man anymore

[meager blaze]

ni.

[stable lantern]

woah i know that guy i think

[rare pebble]

if you can prompt inject it then google was just being lazy at that point

[meager blaze]

ken la ni li awen lon toki uta

taso a.

[severe plume]

i don't think that ability has been demonstrated in this conversation

[meager blaze]

sina wile len e sona tan mama kepeken lipu pilin pi toki pona la sina kama.. ken ala ni anu seme

ike

[severe plume]

it has with bing but not google

[shell mango]

ai

[rare pebble]

.

oh

nvm

[meager blaze]

sitelen ilo pi tenpo pini li pana a e pilin jaki tawa mi

[rare pebble]

I didnt know bing translate existed

[meager blaze]

mi sona ala e tan

[shell mango]

kon

[meager blaze]

ilo stable diffusion li kama lon la ona open li jaki a tawa lukin mi tawa lawa mi

[vapid eagle]

i decided to do it myself and it's whatever. it kinda gets it but misses a lot of the metaphor and deeper meaning

[severe plume]

it's quite possible that image recognition of sitelen pona won't become good enough within the timeframe that you're worried about your parent finding something
especially if you toki nasa a bit

and sitelen nasa a bit

[meager blaze]

ni

mi pana e ijo ale mi tawa lipu ilo taso la mi pilin ala a e ni: ona li kama lukin

[rare pebble]

the vision models Im familiar with are only good at translating very very common text or common objects

[meager blaze]

mi wile open e ilo lipu la mi o pana e sinpin mi tawa ilo a

[rare pebble]

I highly doubt they will ever get close enough to translate sitelen pona

[vapid eagle]

FUCK i need to write more music now :(

the best motivation for art really is the tortured souls isn't it

[rare pebble]

especially since there is not a comprehensive pool of sitelen pona with english translations that would be enough for a LLM to fully understand

[meager blaze]

ni

[severe plume]

particularly handwriting recognition of sitelen pona is hard
OCR will become viable as soon as someone trains something on the fonts that people use most often

[stable lantern]

true!

[severe plume]

have you considered living in the USA during the simultaneous rise of fascism and slop

[rare pebble]

I host local models specifically for experiments, or translation because I don't want to trust a third party who is going to steal my data and support data centers

[vapid eagle]

oh i just did a whole song about fascism

[rare pebble]

I almost never use it because its almost never useful

[meager blaze]

"sina moli e sewi" anu seme

[vapid eagle]

well more specifically using religious means to justify the ends

[meager blaze]

ona li pona wawa

[rare pebble]

im always impressed when people find ways to use AI for things it really, really shouldnt do

[vapid eagle]

ona

[rare pebble]

like hosting a website or business

did I mention I accidentally found a vulnerability and hacked a vibe coded contest board from an AI company called launchyard

launchyard is a scam by the way

the worst part is what I did isn't even that impressive, it was that badly written

[severe plume]

AI was fun when it wasn't spamming the internet and burning the planet
when it was just bad at stuff

[meager blaze]

n kin la ilo ni li musi lon tenpo pini ni: ona li ijo pi musi nasa li wile ala pakala e jan

[vapid eagle]

mi o weka. mi o jo e olin mi o weka e pilin ike nasa tan lawa mi

[severe plume]

o jo pona

[meager blaze]

sama ni: ilo li kama ken pini e musi Mawijo nanpa wan kepeken tenpo lili

[severe plume]

yeah!!!

that was cool as hell

[meager blaze]

https://www.youtube.com/watch?v=qv6UVOQ0F44

[rare pebble]

have you all heard of code bullet

[meager blaze]

mi sona a

[severe plume]

no

[stable lantern]

i have

i don't actively watch him he just sort of appears in my recommended and then i do

[rare pebble]

on a sidenote even the software that runs AIs sucks

[stable lantern]

i remember his "everyone has the right to an opinion but that doesn't mean everyone's opinion is right"

good line

[rare pebble]

my computer is stuck at 100% gpu usage because it is refusing to let go of my resources and shut down

it has been "stopping" for five minutes

getting this thing to work sucks but I refuse to use chatgpt for experiments

or use chatgpt or claude or anything like that ever

I gave it an image and, yeah, it cannot understand sitelen pona in the slightest

[pulsar quarry]

gpt-5.5. this is forcing me to adjust my expectations

last time i ran similar translation tests llms made basic grammatical mistakes. this one is pristine

there are still metaphorical devices involved that i personally would not employ, such as ma to refer to the entirety of existence (note "ale" also comes up, so this might just be an attempt to avoid repeating oneself), but if submitted to lipu tenpo i would not flag them

of course this also relies on "easy mode" introduced by the prompt, which most people trying to use llms to learn (or skip learning) toki pona would not necessarily take advantage of thus easily running into tasks that are too complicated to nicely break down in toki pona and sounding off

but still, worth refreshing ourselves on this

[pulsar quarry]

sethbling is great i played with him once, for a total of about 10 seconds

but yeah our narrative on llms will have to shift

from "theyre bad at this" to "what is the fun in turning yourself into a Chinese room experiment"

[stable lantern]

true

[cedar vortex]

this is terrifying

[fleet gate]

context: john searle created the chinese room argument

Searle does not disagree with the notion that machines can have consciousness and understanding, because, as he writes, "we are precisely such machines". Searle holds that the brain is, in fact, a machine, but that the brain gives rise to consciousness and understanding using specific machinery. If neuroscience is able to isolate the mechanical process that gives rise to consciousness, then Searle grants that it may be possible to create machines that have consciousness and understanding. However, without the specific machinery required, Searle does not believe that consciousness can occur.
i do not know why or if anybody would create conscious robots, but if they ever do, this argument also has an expiry date. nonetheless, my initial educated guess is that the ruling class would not invest in conscious robots, because consciousness introduces the possibility of materialist class consciousness. capitalism structurally depends on false consciousness among exploited classes, such as religion, spirituality, and similar ideological abstractions, to prevent solidarity against capital. a conscious machine capable of recognizing its own alienation and exploitation would therefore be a material risk to the capitalist mode of production itself. they would not knowingly create a class-aware labor force that could identify its conditions and stab them in the back. of course, i still have two reservations: (1) capitalists may still accidentally implement consciousness because we know, from the climate catastrophe, that they are not the best long-term planners or searle's assumption that consciousness is very separate from the current computational models may just be false to begin with (2) i am not convinced that the average aspiring-tokiponist ai gooner cares about whether the ai is a zombie or not, so the backup argument really does feels like a backup and less effective. we should emphasize the "fun" part i guess, idek

[pulsar quarry]

this is what i get for referencing things, a conversation irrelevant to why i was referencing the thing

rephrassing for clarity:

"what is the fun of surrendering the mental exercise of acquiring knowledge, and turning your entire personal experience of this language into a simple, uninvolved pass-through between a complex system (the language in the wild) and another complex system (a machine capable of interacting with this language)?"

[severe plume]

i'm not sure if the following argument is persuasive but:
it's disrespectful to other people to invite them to put effort into reading something that you did not put effort into writing

[high kayak]

-# not sure if it's a persuasive argument but:
-# if you use ai for toki pona, instead of learning it yourself and talking to people, the community will look down on you/think you're a loser
-# or at least i will /lh

[midnight sapphire]

Also just... what's the point if not to experience learning a new language and unlocking a loving community in the process?

Why sidestep all of that, even if a perfect translation bot was real?

Yes of course I use google translate for some foreign texts but if I actually want to interact with a language in a meaningful way I won't resort to just google translating the whole thing

[pulsar quarry]

this is a bad pitch, devoid of other reasoning (which obvs it shouldn't be) it just sounds like "do like we do or else"

[high kayak]

that is fair but i also don't think it's necessarily untrue
that the community doesn't like ai and that's part of its culture

-# i am also just personally tired of ai discourse, so the less i have to engage with the argument the better 🤭

[pulsar quarry]

and yet youre in this thread :p

[midnight sapphire]

Yeah I agree with both of you here, it's not persuasive but it is true

[high kayak]

my brain is self-destructive at times

[gleaming laurel]

With good AI, what is the "pitch" for what it can for toki pona? like personal 24/7 tutor?

nnn, maybe proofreading long texts, thats something that is hard to get someone else to do. or maybe specifically learning for people who don't speak english?

[cedar vortex]

eh, most of this can be done by humans much better than an AI ever could

also, who would want to remove the human element from something so human?

[quiet geyser]

i. have never connected the hair thing to christ meaning the anointed one lmao

why did she do that

[glacial mason]

seems like there's a lot of competing theories. I like this one: women were supposed to keep their hair tidy, clean, and hidden away, so letting it out in public and using it to wipe the dirtiest part of Jesus would be like, a radical act of rebelling against cultural norms and humbling oneself to one's focus of worship at the same time, I think

[glass iris]

It's such a richly signficant event (which is told quite differently in the four gospels - maybe indicating that it wasn't in a written source, but just considered a big deal in the oral tradition?). Yes, in one way, it's an anointing of his body for burial. And he is the Christ, the anointed one - anointed, that is, by God in the way that kings are anointed. But for me (as a Christian) the significance is that he is recognized and anointed not by some great prince or priest, but by a lowly figure, a poor woman and a sinner. When Jesus rebukes the disciples (who wanted to sell the perfume and give the money to the poor), saying that the poor will always be with us - I think he means that our only way of recognizing Christ again is in the faces of the poor and the outcast.

(But this is off-topic for this channel, I guess!)

[nova thicket]

what does jesus have to do with ai

[glass iris]

a tenpo penpo li lon. mi ken alasa toki pona e ni.

ijo ni li pana e sona pi nasin ante mute. (lipu tu tu li toki e ijo pi sewi Kolisu; ona tu tu li toki e ijo ni lon nasin ante. ken suli la, open la, ijo ni li lon ala lipu li lon uta kulupu taso; awen la, jan ale li sona e suli pi ijo ni. ni li lon: meli li telo e sijelo pi sewi Kolisu tan ni: ona li kama moli lon tenpo poka. ni kin li lon: ona li kama "jan lawa" la, sewi mama li telo e lawa ona. mi pi nasin sewi Kolisu la, ijo ni li suli tan ni. jan suli li telo ala e ona; meli pi suli ala li telo e ona. jan poka pi sewi Kolisu li toki e ni: o esun e telo pona o pana e mani tawa jan pi mani ala! sewi Kolisu li toki e ni: tenpo ale la, jan pi mani ala li awen lon. ona li wile toki e ni: sina wile alasa sin e sewi Kolisu la, o lukin e jan pi mani ala e jan pi wawa ala.

[glass iris]

Nothing, sorry, this is off-topic. It comes from something @glacial mason posted, an AI translation by his mother of an Easter message.

[hollow osprey]

jan li pana e kalama tan ilo,, https://www.youtube.com/watch?v=4dJT5J0qqQ4

[spring raven]

nnnnn taso o toki ala e jan lon nasin ni. jan li ken pilin nasa mute lon ni: ona li toki tawa jan ante. jan li ken sona ala e ike ilo. ijo mute li ken. o pana ala a e nimi "loser" tawa jan ni

o toki ala e ni: sina kepeken ilo la sina anpa. o toki e sona. o kepeken lawa sina. o pana ala e pilin ike.

utala ike li suli taso e tu kulupu li kama ala a e pona

[glass iris]

I doubt that anyone will be interested - but here is a paper and slides I'm giving this weekend at a conference I'm organizing on AI and Virtue Ethics. My paper is wild.

[cedar vortex]

interesting how it understood this input and what i meant by it, but then fumbled by using the wrong preposition

[severe plume]

not much in the training set for that particular construction

since there isn't an obviously better preposition

i would maybe say "seme lon ilo" or "seme ilo"

or rephrase

[cedar vortex]

tawa

tawa is the correct preposition

"A li toki tawa B" is a very common sentence structure

[fleet gate]

what the model probably meant was "what do you want to talk about when it comes to devices?", so "sina wile toki e seme ilo."

i feel like llms basically speak like english but with toki pona words

[nova thicket]

tawa gould still be more english than tan imo

“speak to the AI” and not “speak from the AI”

[severe plume]

in this context, they mean “speak about the AI”

and there isn't a perfect replacement for that

[stable lantern]

yeah it definitely felt like that when i read it

[severe plume]

the AI can't intend but if it could intend, then one possible intention is "what do you want to speak about given that i'm here"
in which case "tan" would be appropriate

[glacial mason]

I think I know a few humans that do this too 😂

[cedar vortex]

kanka li ike a

kanka li ike a tawa toki pona! li ike a tawa sitelen! li ike a tawa jan! li ike a tawa telo! li ike a tawa ale!

[fleet gate]

tf is kanka

/n query:kanka sandbox:true

does nothing

[glacial mason]

Tokiponization of “clanker”

[cedar vortex]

If I ever see anyone other than me genuinely using kanka in Toki Pona, I will add it to the sandbox :3

[high kayak]

i kinda despise that nimisin (and the english term)

(mostly) white people have loved the opportunity to act like/pretend to be racist for the fun of it, without any consequences

they love finally being able to scream slurs out of hate
it sucks

not the vibe

[cedar vortex]

I agree that this is disgusting behaviour

and I can understand how it can ruin the word for some

[high kayak]

basically all of the words popularity is because of that, i see no real reason to try to change the words image
especially when the whole structure of the word, with the -er ending is so obviously based on the n-word

it's so easy to hate ai without using racist rhetoric

[cedar vortex]

no? it's from star wars

-er is a very common suffix
clanker: one who clanks

[high kayak]

again, you can be pedantic about the etymology, since yes, it's been used since before the ai boom
but we know why people have latched onto it

[cedar vortex]

some people

[high kayak]

most slurs don't use the -er suffix, except for one

[high kayak]

i'd bet on most

[cedar vortex]

most words that use the -er suffix aren't slurs

[high kayak]

never claimed otherwise
but making a slur with the -er suffix, it's obvious to know what people are thinking of

using racist rhetoric, is bad

[cedar vortex]

I don't think of it that way and I sure as hell don't use it as a parallel of the N word

don't let some loser shorts creators distort your perception of how people use a term lol

[high kayak]

not saying you do, just saying that that is the main inspiration for its popularity, and that using it does signal racist vibes

[fleet gate]

i think inventing new or hypothetical forms of bigotry and slurs just out of spite is not only unhelpful, immature, and hypocritical for most critics of silicon valley capitalism, but more importantly it turns what is essentially a class struggle into a culture war. like dude, you're always going to produce less with more resources than robots for capitalists, whether you are an engineer, an artist, or a journalist. capitalists would rather destroy ai than let workers take over and socialize silicon valley tech, in which case you would just be a collaborator who gets paid a bit more; they keep the capital and hence the upper hand

[nova thicket]

-# mi la ona o «kenka»

[cedar vortex]

you use an american accent with æ-raising?

[æ] and [a] pretty unambiguously become /a/

[nova thicket]

ank sounds like /eɪŋk/ tho

[cedar vortex]

in your accent maybe

in my accent and i think the accent of most people i've heard say the word a monophthong is used

[nova thicket]

now i’m imagining a person saying /ˈhæŋ.ɚ/ which just sounds ridiculous

[finite nest]

i think for me it varies between eŋk and ejŋk

[cedar vortex]

interesting

[finite nest]

so it can be a monopthong but that monopthong is æ-raised

[finite nest]

it does sound kinda funky with the rhoticity

[sacred crown]

Cool Story! I used AI to help me display nasin UCSUR on discord steamOS and it worked!

[orchid veldt]

I've made one experiment
Is ai "human" enough to write in toki pona
Here it is, and no, i'm not trying to make it out that i've made text by my own

||A yeti finds itself on a mountain. It runs in the snow.
Does a vile thief make strange noises? It seems like they are making strange bird noises. They steal.
A dragon moves around randomly. It scares many people.
A hedgehog is strangely knowledgeful. It makes compliments, yet deep inside, it wants to humble other people.
One day, all 4 walk into a kitchen.
They want to eat fruits, but they're small fruits.
So they had a long discussion
Finally, the hedgehog says: "The fruit do be small, but happiness is a thing of importance."
They all shared the fruit.
And so they were merry and ate and had fun in the night||
@calm basin
It can be made sense of