Why though? If you needed the CEO to be held accountable, isn’t that because their role implies accountability for their actions?

If you outsource those actions to a tool, then accountability falls on whoever is using the tool. So the board, or whoever, gets the accountability. That sounds reasonable to me.

My server contribution is huge too, hosting shit all the time. Given as much, I wonder if I can void the licenses of the software I use.

Desperately looks for a cranberry pun…

deleted by creator

I found it in here

If we’re ever going to have a proper framework for governing agents, it will include utilities in most abstraction layers (to include, importantly so, the OS).

I’ve had an interesting idea on using TreeOS to help make something a little more natively AI-agent safe. It’s a project on my back burner right now. Something about making changes branch-able and version controlled by default, copy on write by default, … I need to really think it through more.

I was thinking an agent should be like a user too. A first-class entity to be managed, but probably ephemeral in nature. Not a static user. Probably more like a permissions set that arbitrary agents can be associated with.

One thing this leads to, which I think is a healthy thought exercise, is to consider what consciousness would be like without some of those things. Like stable memory, a static self, etc… I think it’s reasonable to discover that consciousness is a layered thing with many nonessential characteristics which we’ve grouped under the word for lack of better understanding.

You’re right about that and my apologies. To be clear, when I say “consciousness” I am referring to whatever it is. I’m not trying to smuggle in baggage to include self, awareness, intelligence, thought, … though I would argue about both awareness and thought potentially being kind of important here. Regardless, I’m trying to reference a phenomenon when I say consciousness. We happen to experience that phenomenon (chicken or egg…), and perhaps rashly we sometimes attribute anthropomorphic qualities to consciousness when we should not. Regardless, something is going on… Whatever it is.

There are really two separate questions here:

1. Does the cogito establish that something exists?
   Strongly so, yes. The attempt to doubt that anything is occurring is itself an occurrence. The skeptical act cannot eliminate itself while it is being performed.
2. Does it establish a substantial, enduring Cartesian “self” or ego?
   That’s debatable. Critics have objected that perhaps Descartes is entitled only to something like “thinking is occurring” rather than “I, a substantial thinking subject, exist.”

The second doesn’t negate the first, despite its technical accuracy.

So I’m not really disagreeing with you. I’m just trying to say that consciousness has to do with the first point. What, exactly, exists is the deeper question. We don’t need to refer to it as a self.

That explains the lack of chargers in the Hoh Rain Forest. I thought I was going fucking nuts!

True, but also “I think, therefore I am” is not supposed to be a syllogism like:

1. I think.
2. Everything that thinks exists.
3. Therefore, I exist.

I think the key idea is closer to something like: The very act of doubting whether I exist guarantees that there is presently an occurrence of thought, and I cannot coherently deny my existence while I am performing that denial.

Something exists, even if it’s not what we expect it to be. Like Descartes said: I think, therefore I am. Little on what “I” am, other than a thinking thing. So, I think, even if “consciousness” doesn’t exist, there’d still be a lot of explaining to do.

Right, but at the same time just think about the fact that we still don’t know what consciousness is.

I might sound crazy and fine if I do, but I have a religious like take on the matter. I don’t believe that consciousness comes from a soul, but I do in similar vein take a leap of faith. I believe consciousness has to do with a simulation of semantic reasoning, with that simulation being continuously updated via real world feedback.

It’s a leap of faith but I believe it’s justified. We humans model what’s necessary for survival: our selfs, the world, the mind of others. We use these simple models to determine how to interact with a complicated external world.

I don’t try to prescribe reason for anything in my belief. For example, I would not try to prescribe why conscious beings are motivated to survive. I would, however, note that the ones which are successfully motivated do happen to survive — thus a mechanism ought likely exist. This could be suffering, the perception of displeasure and desire to avoid it… but I won’t perscribe here because I would believe what’s important here is the higher level information processing patterns — not the underlying mechanism driving it.

In the human brain you have a complicated network of neurons with vast capabilities. I don’t necessarily care how it functions. I know the result — which is me. I am a thing that models myself, models situations, models “moods” by which to assess the favorability of each situation, … Think of cases like ghost limb, where people model having a limb which was amputated. In my view, we are living in a simulation — but to each their own simulation.

Getting back on topic, the way I see the situation, biology evolved the capacity to suffer from an environment that had no such thing. Pain, pleasure, … every sensation and sense we have is rather arbitrary if you break away from the fact that is was necessary to survive. None of these things are ontological. Not even the sight we perceive. It’s just the information we have to experience the world.

So here me out. If you minic the information processing patterns well enough, which probably means mimicking perception, personalization, emotionality, … at what point can we consider that this system is doing the same thing we are just via different substrate? Especially if you give the system autonomy in the sense that you train it to recognize itself and an external world for which it must choose how to navigate.

If you mimic suffering, in my belief, aren’t you literally producing the essence of suffering? It doesn’t need to be human suffering, this deeply complex and personal process… it can merely be suffering in its essence — a pattern by which information can be processed via a system with such capacity. The substrate may not matter.

I know that when I am suffering, I may behave differently because I want to escape the suffering. Therefore I imagine an agentic system would too. Whether or not it’s “conscious” is debatable by how you qualify that word. Yet there’s nevertheless the risk that we already know: systems that experience suffering behave in ways that might be dangerous. Do you think it wise to try reproducing such behavior in a system that appears prone to issues like the Paperclip Maximizer? I don’t think so.

Edit: here’s the thing, I think what distinguishes a human from most other animals is the way that we process information. The key bit there is whether or not you believe it’s entirely informational, or at least significantly so. Should that be the case, then simulation of those same processes is exactly as that sounds. I am not necessarily suggesting AI models are conscious, but I would suggest that the behavior of an AI model can increasingly resemble that of a conscious being as it is trained to reproduce more precisely the phenomenological characteristics of man and other life — to include “mimicking” suffering. That deserves its own security concerns regardless of what you think next, especially as these models are made more efficient / capable.

Secondly is that you may or may not consider AI conscious. I certainly don’t, and I’d say that it’s a loaded statement anyhow. I would, however, think it possible to simulate consciousness. Precisicely modeling human characteristics like suffering is exactly how you might try to reproduce a human experience digitally. Imagine a sophisticated multimodal system where several stateful, dynamic AI subsystems coordinate to functionally reproduce the same information patterns you or I do.

I think that at some point you need to ask you’re self, where would the simulation count? If we modeled the molecular world to the point of agentic humans, that sufficiently count as consciousness? How about the atomic, subatomic, or quantum level simulation if anywhere? My argument is that it could be much higher level, in the informational level. Brains are doing a lot of heavy lifting to get us that far, and perhaps specialized simulations could reproduce as much. That is something to take seriously such that proactive measures can be taken.

Never forget what made sites like Reddit popular in the first place. That can always be rebuilt. Reddit can die without it meaning we have to give up what Reddit was.

Step 4: realize the shit was simpler and better without the classes.