Comment by woolion | Hacker News Reader

Comment by woolion | Hacker News Reader

> We assert that artificial intelligence is a natural evolution of human tools developed throughout history to facilitate the creation, organization, and dissemination of ideas, and argue that it is paramount that the development and application of AI remain fundamentally human-centered.

While this is a noble goal, it seems obvious that this isn't how it usually goes. For instance, "free market" is often used as a dogma against companies that are actively harmful to society, as "globalization" might be. An unstoppable force, so any form of opposition is "luddite behavior". Another one is easier transport and remote communication, that generally broke down the social fabric. Or social media wreaking havoc among teen's minds. From there, it's easy to see why the technological system might be seen as an inherent evil. In 1872's Erewhon, Butler already described the technological system as a force that human society could contain as soon as it tolerated it. There are already many companies persecuting their employees for not using AI enough, even when the employee's response is that the quality of its output is not good enough for the work at hand, rather than any ideological reason.

I'm neither optimistic nor pessimistic about the changes that AI might bring, but hoping it to become "human-centered" seems almost as optimistic as hoping for "humane wars".

> "free market" is often used as a dogma against companies that are actively harmful to society

This is a predominantly America-specific piece of propaganda, and it's pretty recent.

Adam Smith's ideas are primarily arguments against mercantilism (e.g. things like using tariffs to wield self-interested state power), something he showed to be against the common good. The "invisible hand" concept is used to show how self-interested action can, under conditions of *competitive markets*, lead to unintentional alignment with the common good.

Obviously that's a significant departure from the way it's commonly used today, where Thiel's book has influenced so many entrepreneurs into believing Monopolies are Good.

But the history of this is very Cold War-influenced, where "free markets" were politically positioned as alternatives to the USSR's "planned economy", and slowly pushed to depart further and further from Adam Smith's original argument about moral philosophy.

Economic behavior is inherently game theoretic - agents take various actions and get some positive/negative reward as a result. Whether an agent's reward is positive or negative and of what magnitude, depends on the strategies employed by all agents. If some agents adopt new strategies, the reward calculus for everyone involved can completely change [1].

Over the past few centuries, countless new economic structures and strategies have been discovered and practiced. The rewards for the same action today and in the past can be completely different due to this.

So to me, if someone claimed more than a few decades ago that certain economic strategies and structures are good or bad, its simply not worth listening to them, unless someone reconfirms that the old finding still holds with the latest range of strategies. In that case, the credit and citation goes to that new someone, not the ghosts of the past.

[1] A good interactive demo https://ncase.me/trust/

> if someone claimed more than a few decades ago that certain economic strategies and structures are good or bad

As you point out, it is all game theory.

But things that arrange for the game to be more beneficial to everyone, that align our interests more, deserve to be called "good", regardless of their inability to universally do so.

The latter would be an impossible bar for anything.

Where I find things frustrating, is when someone thinks because something is "good", it somehow becomes "enough". (Think, capitalized versions of different economic schools of thought.)

> Economic behavior is inherently game theoretic.

Game theory is just math. As with any math, the calculations can all work out, but that says nothing of how it reflects nature. All you can say is that if the axioms are all true, then this is the necessary outcome. Look for string theory as a cautionary tale here.

Game theory assumes rational systems. But we have over 6 decades of behavior science which contradicts that fundamental assumption. Economic behavior is not necessarily rational, and subsequently it is not inherently game theoretic. You will find plenty of dogmatic, idealistic, superstitious, counter productive, etc. behaviors in an economy. You need psychology, and not just math, to describe the economics which happens in the real world.

Game theory definitely does not require rational agents. Game theory says there are agents with certain specified strategies. Whether a strategy is rational from the underlying theory of value the agent adopts is completely separate matter. For example, its very common to study agents who always do one action no matter what others do or what the reward function is. Hard to call such actors rational, but that does not stop as from studying them.

When I said rational I meant it in this way. That rational agents will perform in a way in which maximizes their utility (or reward function). In psychology we call this theory homo economicus[1]. Regardless of theories of value, perceived rewards, and utilities, human beings have biases, prejudice, superstition, dogma, etc. etc. In real economies these non-strategic behaviors are consistently observed. This is why I say that economics are not just game theories, they are psychology, sociology, religion, as much as they are zero-sum games between actors.

> Hard to call such actors rational, but that does not stop as from studying them.

This is precisely why I object to your first post. “Shut up and do the math” has not done the wonders which string theorists had hoped. Game theory is a perfectly fine way to mathematics, and to study certain strategies, but it tells us nothing about the nature of economies in the real world.

https://en.wikipedia.org/wiki/Homo_economicus

> arguments against mercantilism

It has been funny to watch the rise of "China is beating us" rhetoric against the steady backdrop of "mercantilism is obsolete/bad" dogma, because the elephant in the room is that China has been running a textbook mercantilist playbook.

Looks to my uneducated eye like China has been enforcing competitive markets internally, with consistent economic policies. Meanwhile the US has stopped enforcing antitrust altogether and keeps violently changing its mind about economic policy every four years.

And then externally Chinese policy is oriented around suppressing the value of its currency, which is basically a monopolist's tactic--artificially lowered prices in order to crowd out competition.

I think that's mercantilist-ish, but kind of a modern version?

It's definitely the opposite of what the US does, the currency is the world reserve and therefore drives the price of the dollar above what it would be without trade, which I guess makes exporting from the US much more difficult?

Anyone who is an expert in global economics please correct me here :)

I mean they did execute a wealthy banker a couple years ago. So I think the mercantile class occupies a different place in society there than in America

> Thiel's book has influenced so many entrepreneurs into believing Monopolies are Good.

Haven't read his book, but the idea that monopolies are good isn't typically made in a vacuum, it's made relative to alternatives, most often "ham-fisted government intervention". It's easier to take down a badly behaving monopoly than to change government, so believing monopolies are better than the alternatives seems like a decent heuristic.

>Haven't read his book, but the idea that monopolies are good isn't typically made in a vacuum, it's made relative to alternatives, most often "ham-fisted government intervention". It's easier to take down a badly behaving monopoly than to change government, so believing monopolies are better than the alternatives seems like a decent heuristic.

What? How is the first alternative poor government instead of multipolar competing companies? When was the last time a Monopoly was actually broken up in the US? ATT/Bell 50 years ago? lol

Good for whom, exactly?

This seems like a classic straw man argument. Plutocratic oligarchs have been making the argument that private monopolies are better than representative democracy at basically anything for decades.

This seems like a classic straw man argument. Plutocratic oligarchs have been making the argument that private monopolies are better than representative democracy at basically any societal function for decades without any actual data.

How would a bad monopoly be likely to be taken down if not by government intervention?

It eventually becomes so big and inefficient that it gets overtaken by new competitors.

A Monopoly implies an organization powerful enough to stop competition. Seems like this solution that relies on competitors is fatally flawed. If there are enough competitors to meaningfully compete then there isn't a monopoly.

You can only truly stop competition by government intervention.

When an organization gets big enough it is indistinguishable from government.

No organization can ever rival a real government like the US due to the latter's monopolization on the use of force.

Unfortunately there isn't also a requirement on not being a complete idotic psycho.

Monoply on force is meaningless, if you shoot your head off with it, which is what is happening with the US atm...

Criminally stupid is the trump all card, pun intended.

Insert better horse/car analogy here

A monopoly comes with serious moats, otherwise it wouldn’t be one. It can stay big and inefficient for decades.

Not if they hire good to go and literally kill the competition.

Open source vs. Microsoft is a great example.

The government is a rare example of an extremely strong monopoly and not just a duopoly or a company holding significant marketshare. And yet people never seem to criticise it on those grounds despite it suffering from all of the same problems that corporate monopolies are accused of.

In democracies people can vote in control of the government to benefit themselves, unlike private monopolies. Corrupt autocracies get much criticism.

The amount of political control individuals have in a democracy is much less than people think. I get to vote every few years for a small preselected handful of representatives, none of whom represent me or my views, and that vote is usually meaningless anyways. In my political life none of the parties I have voted for have won an election. So tell me how much control of the government do I have? How much do any of us have? Especially when the government itself often goes against the wishes of their voters anyways. The reason democracy is preferable to a dictatorship is because it prevents the government from doing things which are widely unpopular, like committing acts of genocide and political violence. But that's it really.

Remember, governments have a monopoly in areas of life that impact everybody constantly. Private monopolies do not, in every case you can simply ignore them and not purchase their products, and by doing so you exert more control over them than you do casting a meaningless vote every few years.

The idea of democracy being will of the people is pure fantasy. It's useful to prevent governments from doing really bad things, that's it. It doesn't even matter who you vote for, broadly speaking. Just that you are able to exert that pressure.

Not even sure if AI was ever "human-centered" . DARPA funded a large amount of AI reseearch from even the 1960s. The DART tool used in the Gulf War made back all of their previous investments. [1]

[1] https://en.wikipedia.org/wiki/Dynamic_Analysis_and_Replannin...

I mean were computers themselves ever human-centered? What’s the difference between AI and a computer?

AI are algorithms generally and a computer is something executing a program .

Let me be more nuanced. What’s the difference between AI and a computer program? In the context above

Alright AI above was more like DARPA funded research programs that resulted in algorithms that allowed for the logistics program to be implemented where the department of war (aka Department of Defense at the time) to have the first gulf war be more efficient to the point that they did more than break even on the grants they gave out.

I just re-read my comment (too late for edits) and there are a number of typos (including missing "not") that significantly degraded the syntax, but the point kind of came across anyway.

Globalization was great for poor countries, not so much developed economies.

No it wasn't. Look at Joseph Stiglitz (Globalization and Its Discontents) and Ha-Joon Chang (Bad Samaritans, Kicking Away the Ladder) for counter-examples.

This isn't correct. The deal is that the poor countries get development and increased employment, and the rich countries get lower prices. Generally speaking both types of countries get richer.

That some workers lost their jobs is a symptom of any change. I don't know why people always get upset people losing their jobs. It's like death, if no one died relatively few people would be born. If you resist job losses you reduce overall employment and economic development.

Are you serious? People get upset about losing jobs because they need jobs to pay their bills. Further, we often build our life identities around work; if you're a good car mechanic or a successful restaurant owner, you're proud of that. It's a part of you.

Having to repeatedly restart your career is risky, painful, and demoralizing. I have no problem seeing why people don't like that and why it can lead to populist backlash or even violent revolutions (as it did in the past).

By the way, to address your closing comment: people don't like dying either and tend to get upset when others die?

Society is better if we sacrifice one horse and buggy driver job for two engineering jobs. The drivers suffer from that, but the net win for society is so plainly obvious that it's a better investment to retrain the driver or just pay the off rather than support a job that dying anyways.

> Society is better if we sacrifice one horse and buggy driver job for two engineering jobs.

That's a "statistic" you're pulling out of your butt, and it's doing a lot of work. No one ever knows if something like that will actually happen.

It could actually turn out that AI sacrifices 100 engineering jobs for 10 low-level service or prostitution jobs and a crap-ton of wealth to those already rich.

> The drivers suffer from that, but the net win for society is so plainly obvious that it's a better investment to retrain the driver or just pay the off rather than support a job that dying anyways.

But what actually happens is our free-market society doesn't give a shit. No meaningful retraining happens, no meaningful effort goes into cushioning the blow for the "horse and buggy driver." Our society (or more accurately, the elites in charge) go tell those harmed to fuck off and deal with it.

> It could actually turn out that AI sacrifices 100 engineering jobs for 10 low-level service or prostitution jobs and a crap-ton of wealth to those already rich.

That's where wealth redistribution (Taxation) comes in. The USA is not good at progressive taxation, but everyone could be better off if it were implemented properly.

> The USA is not good at progressive taxation

The top 10 percent of incomes pay 76% of all income taxes, the top 5% pays around 55% of all income tax.

I would say it’s pretty progressive.

> The top 10 percent of incomes pay 76% of all income taxes, the top 5% pays around 55% of all income tax.

> I would say it’s pretty progressive.

That's a lesson in misleading statistics. Inequality has risen to absurd levels, and just keeps getting worse. It's not "pretty progressive."

First, you are thinking of people whose main source of wealth are assets, not income.

Also, I don't think inequality is a problem, it's better that than everyone being equal, because as history shows, even then, there are some that are more equal than others, and others are just poor.

> But what actually happens is our free-market society doesn't give a shit.

Maybe in the US, but in other countries those things actually happen. It's a political issue, not a moral issue with technology.

I don't think the point is that the transition isn't difficult. It is that there is an overall benefit that outweighs the challenges of the transition.

The sad part is that industrializing societies have not been very good at reconciling the benefits with the costs. The benefits first go to a select few and have seeped out to the masses slowly. Railroads in the US are a good example. The wealth accumulated by the Vanderbilts, Hills and Harrimans, did not get redistributed in any kind of equitable manner. However, everyday people did eventually gain a lot of benefit form of those railroads through economic expansion. (None of which address the loss of the native Americans, whose losses should also be part of the equation.)

My impression is that the transition is such an open-ended process that you can’t really call it that. It’s unclear if and when the challenges will be overcome.

> I don't think the point is that the transition isn't difficult. It is that there is an overall benefit that outweighs the challenges of the transition.

In the abstract, sure. But not when you're on the receiving end. It's like with NIMBYism: we all roll our eyes at NIMBYs until it's actually our own backyard. You're not going to convince a coal miner that they're better off learning to code. You're not going to convince a software engineer that they're better off in the mines.

You're missing my point. Job losses are a fact of life, just like death. Why should people get upset about the fact that someone might lose their job, or die? It's not amoral. These things happen constantly to millions of people, we'd we worn out. We happily send young healthy people to their death fighting in wars so we don't have to. I don't see people weeping because the armed forces exist.

Or is this just some sort of PC bullshit, that we can't talk about this sort of progress without carefully lamenting job losses? If you're not useful doing a job, why should you be employed in it? That's the bottom line.

It would be ok if people lost jobs, and the products and services become cheap correspondingly.

But it would be a problem if people lost jobs and all the product and services keep costing the same as before, and get costlier over time as before...

Oh come now - globalizations was great at the regional level.

It was not that great for sub groups within developed nations.

The original thesis believed that people would be retrained into other equally well paying roles.

Turns out people can’t retrain into new domains, and led to under employment.

Why not, because they're too old to learn, or because the support infrastructure is not there? I believe most people are capable of continued learning, but they might need help (financial etc.) to make the transition.

Even with support infrastructure/money it didn’t work.

You aren’t going to transition into the same level of experience in a new industry.

If it’s jumping into tech/code, for example; even with the best resources, it’s a slog to get back to similar levels of renumeration.

Add in the fact that you have obligations, bills and dependents?

Theres never going to be enough money to keep people afloat while they change domains mid-life.

Check out the programs to retrain miners. Aside from fraud, there were also unrealistic promises.

However, leaving aside leakages - there weren’t that many entry level jobs for code in those regions in the first place.

The cultural differences (trades, physical labour vs code, abstract problems, sedentary work) were sizable barriers to success.

Reply on news.ycombinator.com