Stealing this, thanks.

Do you know what it’s from? I can’t place it at all.

Right, i mean if you made the context window enormous, such that you can include the entire set of embeddings and a set of memories (or maybe, an index of memories that can be “recalled” with keywords) you’ve got a self-observing loop that can learn and remember facts about itself. I’m not saying that’s AGI, but I find it somewhat unsettling that we don’t have an agreed-upon definition. If a for-profit corporation made an AI that could be considered a person with rights, I imagine they’d be reluctant to be convincing about it.

Yeah I think for it to be a proper strange loop (if that is indeed a useful proxy for consciousness-- I think there’s room for debate on that) it would need to be able to take it’s entire “self” i.e. the whole model, weights, and all memories, as input in order to iterate on itself. I agree that it probably wouldn’t work for the current commercial applications of LLMs, but it not what being what commercial LLMs do, doesn’t mean it couldn’t be done for research purposes.

There’s no reason an LLM couldn’t be hooked up to a database, where it can save outputs and then retrieve them again to “think” further about them. In fact, any LLM that can answer questions about previous prompts/responses has to be able to do this. If you prompted an LLM to review all of it’s database entries, generate a new response based on that data, then save that output to the database and repeat at regular intervals, I could see calling that a kind of thinking. If you do the same process but with the whole model and all the DB entries, that’s in the region of what I’d call a strange loop. Is that AGI? I don’t think so, but I also don’t know how I would define AGI, or if I’d recognize it if someone built it.

I’ve always wanted to deploy Chaos Monkey for its actual purpose, but I’ve never been in charge of a big enough infra to make it worth the time. I have turned off databases just to see who files a ticket, which seems in the same spirit.

symlink: !vexillology@lemmy.world

The horror of being a senior admin is realizing that the whole thing could live or die based solely on your actions and decisions. And that you will be blamed.

Validate your backups regularly.

Also, make backups.

Dr. Horrible’s Sing-Along Blog. I know Whedon’s a dirtbag but they are catchy tunes performed by a stellar cast. I keep them all in the shuffle.

EDIT: I decided to watch it again, and now that I’m reminded: “Everyone’s a Hero” didn’t make the cut. I forget about that one.

That’s not really new, or unique to AI. The whole “field” of eugenics was created to give racism the mantle of scientific legitimacy. People will pick through a haystack of data to find a needle that supports (however tenuously) whatever they want to be true. LLMs are just a more convenient way to find or invent those needles.

I still think you’re mistaking the murder weapon for the murderer. AI is just a program, it can be used for whatever purpose the mind can devise. If someone uses an airplane to traffic children, I don’t think a reasonable response is to say that airplanes are child traffickers.

Also I don’t mean “live with” to imply a surrender to how other people (including fascists) use AI. We should do everything in our power to build the world we want to live in, and that means dismantling the power structures of those who abuse them. I mean accepting that AI tools exist and then planning from there. Wishing that they had never been invented is a perfectly fine thing to do, they are something of a headache at the moment, but they’re here and can’t be un-invented. We can either find a comfortable existence in this reality and strive for that (perhaps by limiting their use), or resign ourselves to the doom we find ourselves in.

Now that you mention it, I see that it’s pretty blatantly slop parody. You got me – with the emojis , the em dashes, and most especially the lists of comma phrases that don’t really add to the text, I was sure that this was LLM spam 😅

Well played 🎯

AI isn’t fascism, it’s a tool. Fascists use art in their propaganda, does that make art fascist? No. Fascism doesn’t create, it just corrupts. The problem we have, that we have always had, is with people, not the tools they use. The tools may be terrifying in their hands, but we can’t just wish them gone any more than we can wish nuclear weapons away. We have to figure out how to live with them.

Well, now I’m curious as well. If I only kind of have to pee, like I just noticed it, it feels entirely voluntary to hold it, but if I really have to pee, it does feel like one one part of my brain is sending “pee now” signals that another part of my brain, the conscious decision-making part, has to fight against, which makes me think they have to get involved in the decision somehow. Maybe that physical motor control fight just is how those two parts of the brain mediate each other. Neat.

This just in: New evidence suggests some people think before making a decision? We’ll tell you what this means for your weekend at 11.

For all y’all on here saying you have no survival skills: computers and electronics aren’t going to just go away, and the fact that you’re having this conversation on a federated internet forum means you know at least a little bit more about computers than your average bear. For any scenario where extinction is a possibility, being able to operate a computer and use it to communicate with other humans would be a huge asset, and the more of those people we have around the easier it’ll be to keep it going when things go to hell. Don’t sell yourself short.

They are literally part of the medium.

In my experience, people who live with people who use information for abuse learn to protect information as a first course of action, because it’s hard to predict what information might be dangerous to share. In extreme cases, the only safe opinion to express is that of whoever’s in charge. It can be hard to tell what information can be safely expressed, which I think can make people quick to flatter or agree if they don’t feel safe. It may be that you feel safe to express thoughts about the boss to their face, but they didn’t. It’s a cultural divide I’ve seen both sides of. I’ve worked with people who clearly did not feel comfortable criticizing me even after I encouraged honesty, because they had had bosses before who had said the same thing and abused the privilege of trust. I have also worked with people I did not trust with certain information and I withheld it, even after discussing the matter with peers. I think the things said in confidence can sometimes be harmful and deserve to be rebutted the same as when they’re said in public, but the existence of those things doesn’t make confidential conversation per se bad.

Beyond a certain level of apocalypse we will probably need traveling network engineers, people who are willing to make the difficult journey between human habitats in order to repair and maintain network equipment.