Live data from Hacker News

Capitalism is a paperclip maximizer

thoughtinfection.com

91–99 of 99 posts

Re: Capitalism is a paperclip maximizer

#91

Capitalism depends on price theory, which predicts that making too many paperclips will cause them to be in glut and the price will drop. So a paperclip-maximizing AI might oughta be built to take that into account.

Point missed. It's not that any one product is being produced. "Capital" (or profit or wealth) here is the combined market aggregate of all production. The problem is that this too, as with any single optima process, leads to madness. In the case of capitalism, ecternalities, equity/equality, sustainability, and systemic global risk are all undervalued, and simple short-term profit maximisation tends strongly to "und…

I get the point - it's necessary to reanimate the mouldering bones of Malthus now and again. I just disagree with that because the emiprics of the case are pretty lopsided against him. He could not have known about engines. Which now appear to cause one whale of an externality.

The only thing Capitalism did was replace Mercantilism at scale. You had production before.

What's weird is that economists talk about ways to actually price things like externalities, equality, sustainability and/or risk. That might be madness or it might be reason.

If we can't price them, then are we not dropping the checkbook and reaching for the sword?

Re: Capitalism is a paperclip maximizer

#92
post #84
post #72

Earlier quoted context omitted.

Again, I don't think you are aware of what "indistinguishable from reality" actually entails. The capabilities are so extreme there that you'll have essentially created your own universe. And what do you care if you live in a different universe? You don't even understand this one well enough to know the difference. So you're basically making this decision on the grounds of your own ignorance. What if this world is a…

Fair point. What I had originally thought of as an experience machine was drugs that would produce chemically equivalent reactions as those produced from our experiences. Spiritual usually has an anthropomorphic meaning which is what I object to: Defined as "of, relating to, or affecting the human spirit or soul as opposed to material or physical things" Maybe a better way of stating it is that I am not an empiricist…

That's a better way of putting it, though what I was really getting at was that, whatever reason you have for believing in the importance non-empirical information, you could employ that same reasoning to believe in anything spiritual as well. If what you believe isn't supported by evidence, then a rational person has no reason to agree to it.

Don't underestimate how easy it is for a human brain to confuse wishful thinking with rational belief. I know what you are talking about -- a sense of objectivity -- but I see no reason to believe it is a valid point of view, rather than a pretense of human desire.

Re: Capitalism is a paperclip maximizer

#93
post #66

Earlier quoted context omitted.

Every time I start writing a response to this, I can't help but feel it will simply be misinterpreted. I thought my original response was pretty succinct and clear. If you'd like to continue, please select one discrepancy and I'll be happy to discuss it further with you.

Our disagreement may be definitional - your original response was succinct and clear, but did not make an argument, merely an assertion. A) I see no reason that 'being made up of units which are intelligent' ought to classify a system as intelligent. B) I see no way in which you can usefully talk about anything without intentional behavior as a 'maximizing intelligence'. You can't call 'capitalism' the 'driver of our…

I hope you don't mind, but I'm only going to address one thing right now. If the conversation continues, I will continue to discuss more with you.

Lets start with (A). You are correct, the way I described it does not explicitly explain how it is intelligent. Lets adopt the strictest definition of intelligence which is the ability to predict.

When I refer to capitalism here, I'll be referring to capitalism implemented, not the theory. Capitalism is intelligent. It is made up of people which are intelligent. These components are at times working together, and at times working against each other. Much like your conscious mind and subconscious mind. Conflicts are resolved in the market or through the state via legislation and trials. The results of components working together and resolving their conflicts is how capitalism has decided to behave. Having behavior is not sufficient for intelligence. A virus has behavior but is not intelligent.

So does a capitalist system make predictions? One example should be sufficient to make the answer to this question yes. If I can show that a capitalist system makes predictions, then I will have proven that it is intelligent.

Lets use America for this example. One thing American Capitalism seeks is to mute the highs / lows of the inherit booms and busts in-order to increase the level of stability of the economy. The Fed watches the rates of lending / borrowing to predict bubbles and raises interest rates to slow / stop the growth of bubbles. Once it has popped, and lending/borrowing slows, the Fed then lowers interest rates to promote lending / borrowing. I know this is not new info for you. I just wanted to be thorough in my response.

So American Capitalism predicts bubbles and acts on those predictions. Thus, intelligent.

You might say the Fed isn't capitalism. Which would be true. But your brain isn't you yet it is the component which makes you intelligent. The Fed is part of America's capitalist system.

You might say that capitalism does not involve government. Which I would say, pure capitalism doesn't involve government, but the active capitalist systems we have today do.

Re: Capitalism is a paperclip maximizer

#94
post #93

Earlier quoted context omitted.

Our disagreement may be definitional - your original response was succinct and clear, but did not make an argument, merely an assertion. A) I see no reason that 'being made up of units which are intelligent' ought to classify a system as intelligent. B) I see no way in which you can usefully talk about anything without intentional behavior as a 'maximizing intelligence'. You can't call 'capitalism' the 'driver of our…

I hope you don't mind, but I'm only going to address one thing right now. If the conversation continues, I will continue to discuss more with you. Lets start with (A). You are correct, the way I described it does not explicitly explain how it is intelligent. Lets adopt the strictest definition of intelligence which is the ability to predict. When I refer to capitalism here, I'll be referring to capitalism implemented…

>If I can show that a capitalist system makes predictions, then I will have proven that it is intelligent.

Your example applies perfectly well to a localized ecosystem. I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent?

That's beside the point though, because your example didn't show 'capitalism' making a prediction, it showed the Fed making a prediction. The federal reserve is a functional organization, an entity with goals and intentional behavior. The Fed is a perfect example of an entity that can demonstrate intelligent behavior

It is indeed 'part of America's capitalist system' - again, you fall back on the assertion that a system is intelligent if the actors in it are intelligent, the only obvious source of our disagreement. By your logic, you'd have to consider a baseball game a functional intelligence, a three-legged race a maximizing entity.

Systems of intelligences are not automatically intelligent. It is possible for them to be so; corporations qualify, governments qualify, unions qualify. In order to have intentional behavior, an entity has to be self-coordinated; it has to be able to decide to perform an action, and then perform that action. A market has no coordination, and has no decision-making power - markets are inherently decentralized in that way.

If you choose to define 'intelligence' such that a thing which cannot make decisions or choices, has no goals or intentions, and has no self-impelled behavior can qualify, then your definition of the word is fundamentally confusing to me, and I'm not really interested in trying to puzzle out what you mean by it.

Re: Capitalism is a paperclip maximizer

#95
post #93

Earlier quoted context omitted.

I hope you don't mind, but I'm only going to address one thing right now. If the conversation continues, I will continue to discuss more with you. Lets start with (A). You are correct, the way I described it does not explicitly explain how it is intelligent. Lets adopt the strictest definition of intelligence which is the ability to predict. When I refer to capitalism here, I'll be referring to capitalism implemented…

>If I can show that a capitalist system makes predictions, then I will have proven that it is intelligent. Your example applies perfectly well to a localized ecosystem. I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent?…

Nice, you're back.

> I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent?

So I am assuming you are claiming to be part of the ecosystem?

> That's beside the point though, because your example didn't show 'capitalism' making a prediction, it showed the Fed making a prediction.

Not besides the point, it attacks a necessary part of my argument.

Anyways, you have made some good points. I'll give it some thought and get back to you.

Re: Capitalism is a paperclip maximizer

#96
post #95

Earlier quoted context omitted.

>If I can show that a capitalist system makes predictions, then I will have proven that it is intelligent. Your example applies perfectly well to a localized ecosystem. I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent?…

Nice, you're back. > I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent? So I am assuming you are claiming to be part of the ecosystem? > That's beside the point though, because your example didn't show 'capitalism' maki…

>So I am assuming you are claiming to be part of the ecosystem?

In general, I am part of many systems (eco- and otherwise), but no; that section of reply was an argument against the idea that being able to 'predict' the future was a reasonable way to define intelligence. It doesn't really change the argument if you stick me on the island though - I behave intelligently, and that affects the behavior of the ecosystem, but that doesn't cause the ecosystem be 'intelligent'.

It's interesting to have a debate this slowly, I feel like it's unusually civil :-)

Re: Capitalism is a paperclip maximizer

#97
post #95

Earlier quoted context omitted.

Nice, you're back. > I can for example show that a surge in the population of rabbits on a small island 'predicts' a surge of the population of their main predator. That's the island ecosystem predicting a population bubble and then acts on those predictions. Intelligent? So I am assuming you are claiming to be part of the ecosystem? > That's beside the point though, because your example didn't show 'capitalism' maki…

>So I am assuming you are claiming to be part of the ecosystem? In general, I am part of many systems (eco- and otherwise), but no; that section of reply was an argument against the idea that being able to 'predict' the future was a reasonable way to define intelligence. It doesn't really change the argument if you stick me on the island though - I behave intelligently, and that affects the behavior of the ecosystem,…

Ya, I've been enjoying it :) I was worried you hadn't based upon your previous comment. I'm sorry I haven't returned to this yet, but will soon I promise.

I think we should try to find out where we have agreement and where we begin to diverge. That should make the remainder of this discussion tighter.

Re: Capitalism is a paperclip maximizer

#98
post #97

Earlier quoted context omitted.

>So I am assuming you are claiming to be part of the ecosystem? In general, I am part of many systems (eco- and otherwise), but no; that section of reply was an argument against the idea that being able to 'predict' the future was a reasonable way to define intelligence. It doesn't really change the argument if you stick me on the island though - I behave intelligently, and that affects the behavior of the ecosystem,…

Ya, I've been enjoying it :) I was worried you hadn't based upon your previous comment. I'm sorry I haven't returned to this yet, but will soon I promise. I think we should try to find out where we have agreement and where we begin to diverge. That should make the remainder of this discussion tighter.

I believe our divergence is somewhere under the definition of 'intelligence'. The word is very ill-defined, to the point that we are reduced to describing it functionally in many cases, as with the Turing test.

>Human intelligence is the intellectual capacity of humans, which is characterized by perception, consciousness, self-awareness, and volition.

That's a very normal definition of intelligence. AI is defined (among many other way, naturally) as "the study of intelligent agents"; 'agency' is a requirement for an artificial intelligence.

Since we started with the context of 'paperclip maximizers', I've been talking about intelligence in that context - in order to be studied as an intelligence, a thing must have 'agency', the ability to intentionally act. A thing that doesn't have agency can still have behavior, and that behavior can still be studied, but the behavior is emergent, not intentional - the system does not have agency unless it is capable of having goals, and of making decisions to achieve or progress toward them.

In particular, capitalism isn't 'trying to maximize capital' - that's an effect it's (supposedly) having, but not an intentional one. It's a pretty clear emergent effect - if many of the actors in a system are trying to maximize their personal capital, then the system turning out to maximize net capital should surprise nobody. It's very equivalent to calling a diffusion chamber a 'maximizer' of entropy - a diffusion chamber does maximize entropy, but by calling it a 'maximizer' (in the context of 'paperclip maximizer'), you would ascribe agency to the chamber - it's not trying to maximize entropy, that's just something that it happens to do.

Re: Capitalism is a paperclip maximizer

#99
post #3

Capital doesn't maximize capital; it maximizes productivity. In absence of controls, capital naturally flows to productivity in order to be maximized by the holders of capital; but only because productivity creates wealth, not the capital itself.

> "only because productivity creates wealth" Does it? For whom does productivity create wealth for? Productivity is about getting set results based on the smallest set of resources (time, money, materials, etc...). It does not include anything about wealth, nor where that wealth is distributed to, or whether resources could have generated more wealth used in a different way. It's actually possible for increased produ…

Productivity is the creation of goods and services. Goods and services are the building blocks of wealth: one who has more goods and services is more wealthy. It is impossible to create more wealth without increased productivity.

Now, to your point, the fruits of productivity could be unfairly captured by non-producers, in which case wealth is being misappropriated from those who created it to those who have the power through whatever form to capture it (i.e., taxes, feudal land ownership, etc).

Post reply on HN