No I think using it against brand new troll accounts has essentially no reasonable possibility of false positives. We're talking about the people who make rape threats, spew nothing but abusive insults and slurs, and spam porn, gore, or offensive memes. Usually in the first ~40 minutes of account creation.
So forgive me for not wanting to give them the benefit of the doubt.

@jatone@lemmy.dbzer0.com I feel like using AI to ban the most offensive of troll accounts is not only justified but would benefit the community. You can have manual review for older accounts, but accounts under some low age and post history threshold should get automatic bans if they troll or abuse others. They're not worth putting huge amounts of human effort into. Plus it can allow them to survive and do damage way longer if people need to manually ban them.

What have the piefed devs done?

What is the LLM equivalent of neuroplasticity?

Believe it or not this does actually exist in a sense. There is such a thing as model surgery where parts of models are removed, bits from multiple models recombined, or model layers duplicated. Sometimes this is used to make an LLM with more performance or less resource usage. Models can then be "healed" by continued training so they behave correctly after surgery.

If you want a hardware and systems level example look no further than data center level redundancy and network routing systems that adapt from failures.

Epigenetics I am not sure would have an equivalent given we are not talking about biological creatures.

All the actual mechanisms that things like psychiatric drugs operate by?

Have you never heard of AI drugs using adversarial examples or activation engineering?

Activation engineering has been used in studies like this to manipulate emotion concepts in LLMs: https://www.anthropic.com/research/emotion-concepts-function
Here it's used for steering LLMs: https://arxiv.org/abs/2308.10248

It's also used to uncensor LLMs.

I am not going to sit here and get into an argument around if LLMs are or aren't sentient because we don't know enough about consciousness to make that determination. We are still a long way away from solving the philosophical hard problem of consciousness. We don't really know what parts of animals or humans are necessary components of sentience or are just implementation specific details made by evolution. I also think that you are judging an LLM based on poor understanding of how they work and the ecosystem built up around them.

LLMs are not going to be able to progress into being AGI, because there are many things Brains can do that LLMs cannot... for one, forgetting things, unlearning false things.

Catastrophic forgetting is literally one of the main problems with current models that ML research is currently trying to overcome. But whatever.

Nah if you are making a claim you should provide sources.

Which I've done, and not a single person has looked them up. The reason for that is that no one here is actually interested in the subject - they just cannot accept their feels about humans being special snowflakes not having any support in the science.

LLMs are not really cold calculating machines. They are more grown than made and can display all kinds of novel unintended behaviours. The meme about unknowable shoggoths are closer to what they really are than much else in the public discourse.

I'm not sure lovecraftian horrors are a great metaphor to make people feel better about them lol, but I get your point, the hallucinations are part of why I don't trust the info they give me either tho

that kolektiva thread is extremely frustrating if you're over here on the threadiverse side and know what's actually going on

I mean if someone called me a Zionist of Nazi would be very upset, but that doesn't seem to be the part they're taking issue with...?

Well, ok, sure, if you can frankenstein together different 'brain lobes' from different LLMs, graft them together, and then run through some process that makes them function as a more cohesive whole... yeah!

I can see that as a kind of neuroplasticity... sort of...

Its sort of like a very extreme version of setting up a group of adversarial agents that basically try to bullshit check each other, but you're doing it on another level, internally, actually ending up with a kind of hybrid model at the end.

But its also not like neuroplasticity, in that... nothing like neuroplasticity is present within a single LLM. Whereas in brains... every neuron, every constituent part has a 'mind of its own' in a sense, at a cellular level, from its genetic code and the rules of chemsitry that govern how a single one of them works, and allows for or gives rise to emergent structures and complexity, where the ability of brain to significantly restructure itself is one of those emergent complex properties/phenomena.

But but, I can also see how the ability to have networked and failover redundant hardware does at least somewhat approach a more biological neuroplasticity.

Maybe an interesting experiment would be to try and set up a network of LLMs that are actually designed to mimic the various different sections of the human brain, with their differing areas of relative 'expertise', their tendencies to handle different kinds of 'computational loads' ... and then link them together to form a whole total human brain analog... perhaps that could yield interesting results.

Now, when you talk about activation engineering... and then link the idea of basically talking to them in a certain way... that to me comes across more like gaslighting, social conditioning, brainwashing, etc.

What I was trying to get at was the functional neurochemical mechanisms at play, that moderate how a biological experiences reality, makes decisions and observations... that the complexity and subtlty of the electrochemical soup and all the interaction rules of physics and chemistry at play there... that is significantly different from the binary fit flipping that is the base level of how an LLM works.

But yes I am aware that LLMs have emotional vectors that can be tracked, measured, activated, manipulated via repeated patterns of external stimuli.

I've done a crude version of tinkering with or exploring this myself, with my local LLM setup.

But that page itself says that these are emulations, approximations of human emotions.

I was trying get at the actual distinct foundation level mechanisms that give rise to things like emotions, that are different between brains and LLMs.

Yeah you could. There just wasnt that much of an audience back then. Did you know the heavens gate site is still up amd running

I know, right?

It's the Hypocrisy. One of the core tenets of right-wingers and right-aligned people in general.

Not to mention editorializing other people's posts through federation, at the viewer/render level. Basically Piefed has native tooling to gaslight its users.

In your own words, what was the central point of my message?

1. technology progresses and doesnt sleep
2. you are not everyone

https://arxiv.org/pdf/2602.16800

People really need to stop with the emotional reactions. Being a zionist or nazi is not something that you just "are". It's what you do or say. So if someone speaks in support of Israel they're a zionist. Another popular one is "transphobe". It's not something you are and then do penance for, it's not something you disprove by being a very very good ally, you can do and say transphobic shit accidentally, with the purest of intention, simply out of ignorance and the correct response to being called a transphobe in response is not "how dare you I am very diligent about my pronoun usage I'll have you know!" but a simple "what did I do?". Even if the "accusation" is incorrect, or was done in a bad faith, maybe even slanderous intent, "what did I do?" communicates so much more effectively that you dont consider allyship to be a performance, a list of chores you make sure to tick off and if the accusation was done in good faith will make the other party much more receptive to tell you what happened from their POV and if it was in bad faith makes the other side look silly.

People would rather assume than actually verify the things they dislike and openly hate.

It's disappointing to see what's become of PieFed. I was hanging around in the Matrix channel for a while and it seemed like an alright place. I even helped improve the documentation at one point and made a theme too. This was all done under a different username so I'm well distanced from all this now.

Some drama happened a while ago with the Lemmy devs and people started taking an interest in PieFed. There were all these new moderators and instance owners in the matrix channel and things definitely started to feel different.

One of the bigger instance owners gave off some vibes I wasn't comfortable being around and everyone else was just quietly okay with it. I guess. I don't really know how to describe the feeling but I know that it was my cue to quietly leave.

I saw the signs in PieFed itself and the community developing around it so I'm not surprised. Just disappointed.

That the average person couldnt just get on the internet and blast random stupid shit. Which was simply not true. Most people just didnt care enough yet, but there were still plenty around that did

Very well said. Yet another reason it's valid to be intolerant of Nazis and Zionists, is because they can set down their ideology and stop being hateful any time they want

I wasn't trying to make people feel better. If you look at AI safety and alignment it's all quite scary. Lovecraftian horror isn't that bad of a description really.

Yeah you aren't quite getting what I am trying to explain here.

You used the example of Pheneas Gage as an example of neuroplasticity and neural remapping in the event of extreme damage. By talking about model surgery I was trying to make the point that LLM training algorithms do the same sort of thing. They manipulate each individual parameter of the model in a similar way to how neuroplasticity manipulates synapses in humans. It goes below the level of neurons onto the level of individual connections and biases affecting those neurons.

As for different models handling different things while this is actually done in practice we also have the concept of mixture of expert models. It's worth looking into.

pay the dues

I find this viewpoint absolutely baffling. No one owes you for your data when you give it away.

Every single thing we post is graffiti on the sides of a public view board. There are no magic words you say that protect you from other people, or in this case, anyone on the internet from taking what you post.

The type of thing you submit doesn't change that. And one of the reasons this baffles me is... Any idiot knows you don't post your password online. You don't post your credit card information online. You don't post your social security number online.

Why?

Because it will be stolen.

Why people pretend that scraped comments or someone's art should be immune to this just... baffles me.

If you want it to be different, you have to actually do something about it.

It's the problem with liberals today. People will spend five billion hours online whingeing that LLMs shouldn't steal art, and not a single minute actually organizing protests or lobbying their government representatives or building a site or setting up a server. Literally complaining nonstop without lifting a finger to make change.

Also I am so fucking over how people call every group they don't agree with blank-bros. This is another kind of female erasure and it pisses me off.

Our instance is known for having a higher than average number of autistic and ADHD members. Telling us to go read a word wall when a tantalizing pedantic linguistic argument is waiting to be had is just not gonna fly.

Oh but on a more friendly note, please feel free to look up tips for writing for ADHD readability. Many of us see big blocks of text like that and our eyes straight up skip it.