Eugen RochkoMachine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
From what I've observed, people who claim that LLMs can replace artists don't understand art, people who claim that they can replace musicians don't understand music, people who claim that they can replace writers don't understand literature, and people who claim they can replace translators don't rely on translations. If I had a button that would erase LLMs from the world but it would take machine translations away (which is a false dichotomy anyway), I would absolutely still press it.
Seyfeddin BassaracTechnology is not inevitable. We've decided not to have asbestos in our walls, lead in our pipes, or carginogenic chemicals in our food. (If you're going to argue that it's not everywhere, where would you rather live?) We could just not do LLMs. It's allowed.
Domingos Faria@Gargron It is a technology that humanity has been seeking for a long time. At least since the 1950s, with Turing and his colleagues.
@df No, this is marketing. OpenAI, Google, Anthropic &co want you to believe that what they're doing is artificial intelligence. My professional opinion is that LLMs are a dead end technology to creating actual intelligence. And if any of those companies did create actual intelligence for the purposes they pursue, it would be slavery, for which I cannot advocate.
@Gargron LLMs are not exclusively a product of large corporations or just marketing. Much of the research and development also takes place in open source and academic communities. The codes for these LLMs are public and can be audited or run locally. Furthermore, I argue that serious ethical reflection is necessary, but prohibition is not the way forward.
Papa Exmatrikulatus 🏴🏳️🌈@df
Consciously not using something ≠ prohibition
Edit: Also, who cares who worked/ envisioned or works on this now? If you think about LLMs enough, you will likely see enough good arguments about the resource waste, centralization of power and multiplication of slop which describe LLMs. We lived without it before and we can live without it in future times.
Consciously not using something ≠ prohibition
Edit: Also, who cares who worked/ envisioned or works on this now? If you think about LLMs enough, you will likely see enough good arguments about the resource waste, centralization of power and multiplication of slop which describe LLMs. We lived without it before and we can live without it in future times.
Joshua Grochow@df @Gargron Academics may study LLMs out in the open, but I don't think academia has been able to produce LLMs whose outputs are sufficiently marketable compared to the current commercially available ones. Because the first "L" ("large") is - in our current, limited understanding - crucial for the verisimilitude of the synthetic text, and only corporations (and governments, but they mostly haven't gotten to this yet) have the scale to get large enough for that so far.
❣️u.s. pause federal tax remittances@joshuagrochow and the lack of moral compass or publicly stated ethicals standards that would allow university employees to steal large enough sets. small sets of text are read and understood by humans who can, far more efficiently, apply appropriate prior written and other formats of source material to a specific use case.
programming a calculator only makes reasonable sense if the computation requires enough repetition to warrant the resources used in building it, or it's a closed set without novelty... like, for example, a numerical calculator. ; )
edited for typos and clarity: it was killing me, apologies for the notification disruption.
ResunaThe code for the LLM interpreter is relatively simple, and bears the same relationship to the actual LLM as the C compiler does to an operating system. The models are the real software and the ones big and complex enough to be useful are the product of large corporations and mass copyright violation.
millennial fulcrum@Gargron they'll never create intelligence because intelligence requires will and they do not understand will. they dont even posses one of their own: their own behaviour is driven by feelings and shaped by a commercial playbook. there is zero chance they will ever create intelligence.
🇺🇦 haxadecimal 🚫👑@Gargron @df
Yes. The flowchart has three boxes:
1. Create LLM
2. Then a miracle occurs
3. Profit from AGI !!!
The companies pushing so-called "AI" have completed step 1. Some of them try to tell us that they've nearly got a handle on step 2, but that's just an attempt to swindle more investors. There is literally NOTHING that fits in the hole of step 2.
PatrysTransformers are neural networks.
LLMs are transformers wrapped in some Python scripting.
Every neural network can be accurately represented as an Excel sheet, even if it ends up having billions of cells.
Since it's just addition and multiplication, the model is fully deterministic. Same input, same output. Not intelligent.
It's Python code that does probabilistic sampling of the output. It's just a few lines of well-understood math plus a dice roll. Again, not intelligent.
Fish Face@patrys @df @Gargron does determinism imply non-intelligence?
If you hooked up the computer to a Geiger counter for true random noise and used that to modulate the output, would that have any bearing on its intelligence?
Or from the other side, what makes you think our brains are non deterministic, and why does that make us more intelligent than if the exact same history and sense-data always produced the same response?
@FishFace @df @Gargron If it’s deterministic, it can be unrolled into a giant lookup table. Did we kill phone books because they were on the verge of achieving AGI?
To me, intelligence implies a lot of things, like being able to form higher-order abstractions, learn, and thus remember things (no, being passed your “memories” as part of every prompt does not count). It also implies being curious.
@patrys @df @Gargron given that the lookup table would generally be infinite, I don't even see what that would have to do with anything. What about the Geiger counter?
I don't think those things are really needed for human-like intelligence, and something like curiosity can easily be simulated by a rules-based system.
@patrys a computable function can generally produce infinitely many different outputs. You're still not saying why a non-deterministic part affects intelligence.
@patrys LLMs are intelligent only in the sense of pattern recognition; that is, they possess logical intelligence. However, some psychologists argue that there are multiple intelligences that cannot be reduced to logic, nor are LLMs capable of possessing them. See psychologist Howard Gardner.
@df @FishFace @Gargron This pattern recognition is an artifact of the training process, not something that occurs at inference time. It’s like having termites dig some tunnels in an earth mound, then removing the termites, pouring aluminum into the mound, and attributing the resulting intricate shapes to the intelligence of the mound. The patterns it carries are from human artifacts used as input for the model before its weights settled.
fr0g@FishFace @patrys @df @Gargron
"Or from the other side, what makes you think our brains are non deterministic"
Us having free will/being non-deterministic is pretty much the base assumption we all operate on to even be able to function as humans. That of course doesn't mean that it's automatically true, but it makes the question of why do you think your brain is non-deterministic a no-brainer to answer: because we can't help but perceive ourselves as such.
@frog_reborn @FishFace @df @Gargron The very fact that you can read and mid-sentence learn something that changes your perception of the world means that you have brain plasticity that no neural network possesses. It’s deterministic AND rigid because training and inference happen separately.
GLC
Aetherial
William B Peckham
DJGummikuh@Gargron while all your examples are 100% valid, I seriously question whether we would be able to manage to do that today. With the utter shambles most democracies are in currently, multi-national Corporations can run roughshod on environmental protection, worker safety, child protection and just about everything that past generations fought hard for.
imagine for a moment, the billionaires have been beheaded and the yachts sunk into the sea. the value in the output of workers 100% reinvested into local communities. all of it. none for colonial masters far away. the 20 hour work weeks and all human workers hands full of the satisfaction their efforts are meaningful... no more busy work for shareholders to skim value out of. only meaningful work. custom artisanal everything. housewares repaired by local handicrafters. clothes sewn and tailored to each body. homes and townhomes and communal living spaces built and maintained by cooperative owners. neighboring towns and regions and nations translating with loving care between the communities of meaning... interconnected with care. 💜
Mason Loring Bliss
The ServitorAnd that lasts 1-2 generations before new people who don't understand the problems that lead their parents to create the paradise chafe under their constraints and begin changing the system to something its originators wouldn't like, this creating conflict, diversity of thought, and continuing the cycle of history.
See: reality.
@TheServitor hm. so you do not believe in evolution then?
you ignore the myriad plants and beasties with behavior change documented in the geological record? or the changes in human written language documented over millennia? or dr. martin luther king's statement on the arch of the moral universe? or even the historical shifts in graphic novels and comic book tales over 80 or so years?
weird.
nihilism has never had much sway with me. i accept humans believe such a framework applies beyond the self and assert omniscient determinism universally. however, i do not abide such a concept as the universal human. we are varied beasties and i have faith in evolution of arrangements of living matter as well as of patterned practices.
see: the data.
@TheServitor of course, the u.s. and israeli "leadership" could destroy the species in the next week and whatever survives the nuclear fallout would have a very different future. humans are weird.
LisPi@DJGummikuh @Gargron I'm sure the oldschool labor activism methods would still work.
arceuthobium@DJGummikuh @Gargron I would argue that LLMs shouldn’t require special regulation, just no exemptions from ordinary regulation. A lawyer who files a specious brief full of fake citations should be sanctioned the same regardless of whether an LLM wrote it.
Tekchip@Gargron would you know if you've seen a good outcome of an LLM? You'd somehow be able to identify when the LLM got it right?
I assure you you've experienced good LLM output and don't even know it. Because that's what good LLM output looks like. Indistinguishable from human output.
Your examples are perhaps false equivalencies. Take asbestos. We didn't abolish insulation. We developed better, safer insulation. We didn't stop dying food colors, we just developed safer dyes etc.
@Gargron ultimately LLMs like any other software is a tool. It's all about how a human uses them.
Lets take photoshop as an example. Humans generate vast amounts of garbage photoshopped images. Ever been to deviant art?
And yet the same tool is used by professionals all day every day to create stuff we like and enjoy.
The same applies to LLM use, and back to my first reply. What you lament is low quality output a human shared. Meanwhile the tool gets used masterfully to great effect elsewhere
Marta Threadbare@cygnathreadbare @Gargron yeah, that's a garbage way this technology has been developed. Unfortunately if we threw away every technology built on the back of people doing bad things we wouldn't have much technology, unfortunately.
I don't fault lamenting how it's come to be and even how it's used broadly. But claiming it's useless because some folks use it poorly isn't really accurate indicator of the technologies usefulness.
Tim Mills@Tekchip @Gargron The technology doesn't yet do much of what is claimed for it; and it is already expensive in terms of externalized costs: memory, energy, water. It really looks like the future of LLMs depends on mass acceptance of the "what-if" scenario - those hoped-for advances where it works better, uses less energy, and somehow doesn't wipe out thousands of middle- to low-level jobs.
ClipHeadLet me ask you this: It's your birthday.
5 of your friends met some days before and wrote a song for you. It's not really good, the text doesn't even rhyme...but they did this for you and they had fun.
They enjoyed the act of creating.
5 other friends wrote a prompt and pressed a button to generate a song.
Which song will you remember?
Kiloku - Secretário do Caos@Tekchip @Gargron the tiny potential for very rare good outcomes are not worth the constant poisoning of humanity's collective information corpus.
For every "good" generated content there are dozens of thousands of terrible slop that are difficult to separate from genuine useful information or material when doing research or code reviews, etc.
Not to mention that these "good" outcomes are much costlier to humanity than creating by hand, with no benefit.
@Kiloku @Gargron the problem is you want to assume they are rare outcomes. I don't believe they are. Unfortunately that's where we're at an impasse. It's literally impossible to measure the good outcomes.
I agree the environmental outcome is terrible. I don't like that part. What we can look forward to is the technology improving. General computers used to use WAY more power than they do now. The same is going to happen with LLM technology. Hopefully sooner than later. Folks are working on it.
@Tekchip @Gargron (also, most of what "AI" boosters *think* is good generated content is actually laughably bad to anyone who knows the subject matter of the content it generates. I'm certain you've shared something that you thought was indistinguishable from human created content that other people knew and saw a bunch of problems with as soon as they examined it further than a cursory glance)
Ben AvelingOh LLM output always looks good, as long you don't understand what it's talking, then it looks great.
Very beautiful, very plausible.
But if you actually understand whatever it is that the LLM is talking about, then it rapidly becomes obvious that it's just spewing all the right words in a random order.
@Tekchip @Gargron
Very beautiful, very plausible.
But if you actually understand whatever it is that the LLM is talking about, then it rapidly becomes obvious that it's just spewing all the right words in a random order.
@Tekchip @Gargron
카를로스@iscarlosmolero @Gargron unfortunately that's nearly every technology and it's not a problem that's country specific. Rail lines built by slaves, battery materials mined by slaves, coal power plants destroying the environment, plastics littering the planet, I could go on. Do we stop using trains, generating power, using batteries, using plastic? Maybe in some cases we limit their use, but generally no, we work to make the technology safer and more equitable.
@iscarlosmolero @Gargron many countries, yes america included, are separated from their externalities. Just because the slavery doesn't happen here anymore we're more civilized or something. Never mind the fact the supply chain that gives you X thing starts someplace else that you, we, I, should be responsible for.
Sandor Spruit 🇪🇺🇳🇱🇺🇦🇨🇦@sandorspruit @Gargron in the case of generative art it's purely subjective, as all art is. So that's not a matter of knowing something in advance. In terms of things like programming code generation you don't have to know in advance how the code works as long as the output doesn't break or is what you expect. In terms of Q&A output most LLMs now include reference so you can assess, post output. Long way to say I don't think you need to know anything in advance if you have the wherewithal to...
@sandorspruit @Gargron ...perform some due diligence. How do I know a wikipedia article is true without prior knowledge? I check the references and/or the information fits the framework of whatever other information/data has shown to be true.
Jacob Ibáñez 🔒
@Gargron we were not doing them 5 years ago, shouldn't be that difficult, right? Not even the cell phone was so quickly introduced in our lives, it's a complete madness!
Mastodon MigrationMachine vs. Human translation of fiction is an excellent analogy. Good translation involves an understanding of complicated material in an intuitive and nuanced way, and conveying those subtleties cleverly using equally complex forms in the target language while retaining the beauty of the writing. It involves much higher level thought than what LLMs do.
Likewise software engineering is much more complex and involves higher level thinking than prompted LLM code generation.
Sarcasm is all I have Left@Gargron You sound like me arguing against the inevitability of mass use of the cell phone.
I never understood why we gave up crystal clear audio, a two way simultaneous connection (yes, both parties could talk at the same time and hear wha5 the other had to say), and phone books for unintelligible garbled speak, dropped calls, delays, and no way to look up the damn phone number.