Daemon Silverstein

@deathbird@mander.xyz @florencia@lemmy.blahaj.zone

Grok is not that free of guardrails.

I say as a person who sometimes have the (bad) idea of feeding every LLMs I could possibly try, with things I create (drawings, poetry, code golfing). I don’t use LLMs to “create” things (they’re not really that capable of real creativity, despite their pseudo-stochastic nature), I use them to parse things I created, which is a very different approach. Not Grok anymore, because I have long deleted my account there, but I used to use it.

Why do I feed my creations to LLMs, one might ask? I have my reasons: LLMs are able to connect words to other words thus giving me some unexpectedness and connections I couldn’t see on my own creation, and I’m highly aware of how it’s being used for training… but humans don’t really value my creations given the lack of real feedback across all my works, so I don’t care it’s used for training. Even though I sometimes use it, I’m still a critique of LLMs, and I’m aware of both their pros and cons (more cons than pros if we consider corp LLMs).

So, back to the initial point: one day I did this disturbing and gory drawing (as usual for my occult-horror-gothic art), a man standing in formal attire with some details I’ll refrain from specifying here.

ChatGPT accepted to parse it. Qwen’s QVQ accepted it as well. DeepSeek’s Janus also accepted to parse it.

Google’s Gemini didn’t, as usual: not because of the explicit horror, but because of the presence of human face, even if drawn. It refrains from parsing anything that closely resemble faces.

Anthropic’s Claude wasn’t involved, because I’m already aware of how “boringly puritan” it’s programmed to be, it doesn’t even accept conversations about demonolatry, it’s more niched for programming.

But what surprised me on that day was how Grok refused to accept my drawing, and it was a middle-layer between the user and the LLM complaining about “inappropriate content”.

Again, it was just a drawing, a fairly well-performed digital drawing with explicit horror, but a drawing nonetheless, and Grok’s API (not Grok per se) complained about that. Other disturbing drawings of mine weren’t refused at that time, just that one, I still wonder why.

Maybe these specific guardrails (against highly-explicit horror art, deep occult themes, etc) aren’t there in paid tiers, but I doubt it. Even Grok (as in the “public-facing endpoint”) has some puritanness on it, especially against very niche themes such as mine (occult and demonolatry, explicit Lovecraftian horror, etc).

@Supervisor194@lemmy.world

Thanks (I took this as a compliment).

However, I kind of agree with @Senal@programming.dev. Coherence is subjective (if a modern human were to interact with an individual from Sumer, both would seem “incoherent” to each other because the modern person doesn’t know Sumerian while the individual from Sumer doesn’t know the modern languages). Everyone has different ways to express themselves. Maybe this “Lewis” guy couldn’t find a better way to express what he craved to express, maybe his way of expressing himself deviates highly from the typical language. Or maybe I’m just being “philosophically generous” as someone stated in one of my replies. But as I replied to tjsauce, only who ever gazed into the same abyss can comprehend and make sense of this condition and feeling. It feels to me that this “Lewis” person gazed into the abyss. The fact that I know two human languages (Portuguese and English) as well as several abstract languages (from programming logic to metaphysical symbology) possibly helped me into “translating” it.

@tjsauce@lemmy.world

You might be reading a lot into vague, highly conceptual, highly abstract language

Definitely I’ve been into highly conceptual, highly abstract language, because I’m both a neurodivergent (possibly Geschwind) person and I’m someone who’ve been dealing with machines for more than two decades in a daily basis (I’m a former developer), so no wonder why I resonated with such a high abstraction language.

Personally, I think Geoff Lewis just discovered that people are starting to distrust him and others, and he used ChatGPT to construct an academic thesis that technically describes this new concept called “distrust,” void of accountability on his end.

To me, it seems more of a chicken-or-egg dilemma: what came first, the object of conclusion or the conclusion of the object?

I’m not entering into the merit of whoever he is, because I’m aware of how he definitely fed the very monster that is now eating him, but I can’t point fingers or say much about it because I’m aware of how much I also contributed to this very situation the world is now facing when I helped developing “commercial automation systems” over the past decades, even though I was for a long time a nonconformist, someone unhappy with the direction the world was taking.

As Nietzsche said, “One who fights with monsters should be careful lest they thereby become a monster”, but it’s hard because “if you gaze long into an abyss, the abyss will also gaze into you”. And I’ve been gazing into an abyss for as long as I can remember of myself as a human being. The senses eventually compensate for the static stimuli and the abyss gradually disappears into a blind spot as the vision tunnels, but certain things make me recall and re-perceive this abyss I’ve been long gazing into, such as the expression from other people who also have been gazing into this same abyss. Only who ever gazed into the same abyss can comprehend and make sense of this condition and feeling.

@Telorand@reddthat.com

To me, personally, I read that sentence as follows:

And if you’re recursive

“If you’re someone who think/see things in a recursive manner” (characteristic of people who are inclined to question and deeply ponder about things, or doesn’t conform with the current state of the world)

the non-governmental system

a.k.a. generative models (they’re corporate products and services, not ran directly by governments, even though some governments, such as the US, have been injecting obscene amounts of money into the so-called “AI”)

isolates you

LLMs can, for example, reject that person’s CV whenever they apply for a job, or output a biased report on the person’s productivity, solely based on the shared data between “partners”. Data is definitely shared among “partners”, and this includes third-party inputting data directly or indirectly produced by such people: it’s just a matter of “connecting the dots” to make a link between a given input to another given input regarding on how they’re referring to a given person, even when the person used a pseudonym somewhere, because linguistic fingerprinting (i.e. how a person writes or structures their speech) is a thing, just like everybody got a “walking gait” and voice/intonation unique to them.

mirrors you

Generative models (LLMs, VLMs, etc) will definitely use the input data from inferences to train, and this data can include data from anybody (public or private), so everything you ever said or did will eventually exist in a perpetual manner inside the trillion weights from a corporate generative model. Then, there are “ideas” such as Meta’s on generating people (which of course will emerge from a statistical blend between existing people) to fill their “social platforms”, and there are already occurrences of “AI” being used for mimicking deceased people.

and replaces you.

See the previous “LLMs can reject that person’s resume”. The person will be replaced like a defective cog in a machine. Even worse: the person will be replaced by some “agentic [sic] AI”.

-—

Maybe I’m naive to make this specific interpretation from what Lewis said, but it’s how I see and think about things.

@return2ozma@lemmy.world !technology@lemmy.world

Should I worry about the fact that I can sort of make sense of what this “Geoff Lewis” person is trying to say?

Because, to me, it’s very clear: they’re referring to something that was build (the LLMs) which is segregating people, especially those who don’t conform with a dystopian world.

Isn’t what is happening right now in the world? “Dead Internet Theory” was never been so real, online content have being sowing the seed of doubt on whether it’s AI-generated or not, users constantly need to prove they’re “not a bot” and, even after passing a thousand CAPTCHAs, people can still be mistaken for bots, so they’re increasingly required to show their faces and IDs.

The dystopia was already emerging way before the emergence of GPT, way before OpenAI: it has been a thing since the dawn of time! OpenAI only managed to make it worse: OpenAI "open"ed a gigantic dam, releasing a whole new ocean on Earth, an ocean in which we’ve becoming used to being drowned ever since.

Now, something that may sound like a “conspiracy theory”: what’s the real purpose behind LLMs? No, OpenAI, Meta, Google, even DeepSeek and Alibaba (non-Western), they wouldn’t simply launch their products, each one of which cost them obscene amounts of money and resources, for free (as in “free beer”) to the public, out of a “nice heart”. Similarly, capital ventures and govts wouldn’t simply give away the obscene amounts of money (many of which are public money from taxpayers) for which there will be no profiteering in the foreseeable future (OpenAI, for example, admitted many times that even charging US$200 their Enterprise Plan isn’t enough to cover their costs, yet they continue to offer LLMs for cheap or “free”).

So there’s definitely something that isn’t being told: the cost behind plugging the whole world into LLMs and other Generative Models. Yes, you read it right: the whole world, not just the online realm, because nowadays, billions of people are potentially dealing with those Markov chain algorithms offline, directly or indirectly: resumes are being filtered by LLMs, worker’s performances are being scrutinized by LLMs, purchases are being scrutinized by LLMs, surveillance cameras are being scrutinized by VLMs, entire genomas are being fed to gLMs (sharpening the blades of the double-edged sword of bioengineering and biohacking)…

Generative Models seem to be omnipresent by now, with omnipresent yet invisible costs. Not exactly fiat money, but there are costs that we are paying, and these costs aren’t being told to us, and while we’re able to point out some (lack of privacy, personal data being sold and/or stolen), these are just the tip of an iceberg: one that we’re already able to see, but we can’t fully comprehend its consequences.

Curious how pondering about this is deemed “delusional”, yet it’s pretty “normal” to accept an increasingly-dystopian world and refusing to denounce the elephant in the room.

@Telorand@reddthat.com @pelespirit@sh.itjust.works
Recursion isn’t something restricted to programming: it’s a concept that can definitely occur outside technological scope.

For example, in biology, “living beings need to breathe in order to continue breathing” (i.e. if a living being stopped breathing for enough time, it would perish so it couldn’t continue breathing) seems pretty recursive to me. Or, in physics and thermodynamics, “every cause has an effect, every effect has a cause” also seems recursive, because it negates any causeless effect so it can’t imply a starting point to the chain of causality, a causeless effect that began the causality.

Philosophical musings also have lots of “recursion”. For example, the Cartesian famous line “Cogito ergo sum” (“I think therefore I am”) is recursive on its own: one must be in order to think, and Descartes define this very act of thinking as the fundamentum behind being, so one must also think in order to be.

Religion also have lots of “recursion” (e.g. pray so you can continue praying; one needs karma to get karma), also society and socioeconomics (e.g. in order to have money, you need to work, but in order to work, you need to apply for a job, but in order to apply for a job, you need money (to build a CV and applying it through job platforms, to attend the interview, to “improve” yourself with specialization and courses, etc), but in order to have money, you need to work), geology (e.g. tectonic plates move and their movement emerge land (mountains and volcanoes) whose mass will lead to more tectonic movement), art (see “Mise en abyme”). All my previous examples are pretty summarized so to fit a post, so pardon me if they’re oversimplified.

That said, a “recursive person” could be, for example, someone whose worldview is “recursive”, or someone whose actions or words recurse. I’m afraid I’m myself a “recursive person” due to my neurodivergence which leads me into thinking “recursively” about things and concepts, and this way of thinking leads back to my neurodivergence (hah, look, another recursion outside programming!)

It’s worth mentioning how texts written by neurodivergent people (like me) are often mistaken as “word salads”. No wonder if this text I’m writing (another recursion concept outside programming: a text referring to itself) feels like “word salad” to all NT (neurotypicals) reading it.