#long rant (tm)

[uboat53]

So, some more thoughts about AI. Obviously I've been testing with local AIs and paying attention to the advances that are going on in the field and… I don't think it's happening. LONG RANT (TM) time.

INTRODUCTION

Okay, what's the thing that all the AI companies are, implicitly or explicitly, promising? AI is going to reach the point where it is more capable at anything than a human being could possibly be and, thus, it will replace every single human being in employment and either bring about a utopia or dystopia.

And here's the thing, it's not going to happen and there several reasons why it won't.

THE DETERMINISM PROBLEM

One of the things I've been researching is how to train an AI. I do weird stuff sometimes and it'd be nice to have an AI specifically tuned/trained for the types of things I do rather than trying to fit a generalist square peg into the round hole of strange things I tend to do when I'm bored or curious. The first thing you learn when studying this is that AI is explicitly non-deterministic.

What does "non-deterministic" mean? Well, "deterministic" means that, if you give it the same input, you'll get the same output every time. Computer code, for example, is deterministic. Non-deterministic, obviously, means the opposite of that. In fact, every training guide I've read is explicit that if you train an AI to the point where it spits out the same answer every single time to a given query, you've actually broken it. You want to back off while it's still a bit fuzzy in order to preserve what makes it AI.

The thing is, that's a problem in an unbelievable number of areas. I mean, sure, there are some fields in which close is good enough or where multiple answers could be correct, but most fields you at least need to be able to give an accurate answer, the same accurate answer, every time! Even 1%-2% off can be completely unacceptable in a huge amount of applications.

Now, this can be mitigated to some degree by the use of systems of AI that check each other but, even then, accuracy is too low for a great deal of common applications currently done by humans or even some done by deterministic software.

THE COMPUTE PROBLEM

One of the things I don't see talked about a lot is that it's taking exponentially more computing power to train and run each generation of AI. There are early generation AIs (circa 2020, 2022, or even sometimes 2024) that can run on your personal computer but, right now, even Moonshot's recently released Kimi K3 AI, which is an order of magnitude lighter weight than Sol, Fable, or Mythos, requires whole server racks to run just a local instance.

Meanwhile, the world is running short of computer parts. I'm sure you've noticed the price of video cards and, lately, memory for computers spiking. NVidia is trying to sell more powerful chips that don't even exist yet and may not even be past the concept stage and every major company has been spending exponentially more money trying to suck up all the chips that exist in an effort to gain a little bit of ground on the competition.

Every chip maker in the world is working at maximum capacity right now and I'm not even confident they'll be able to produce enough for the next generation, much less the one after that. Moreover, the constrained supply and exploding demand is creating the next problem which is…

THE MONEY PROBLEM

The late 2010s and early 2020s were a really interesting time money-wise in that a lot of big companies, tech companies in particular, had enormous piles of cash and nothing to spend it on. There weren't enough promising start-ups to grab and they didn't have enough promising R&D opportunities to funnel money into, so they ended up with huge balance sheets based on their incredibly profitable core businesses.

But here's the thing, if you're a business that doesn't show good potential for growth, no matter how consistent or profitable your core business is, your stock price falls. Most people buy stock on the promise of future growth, not current profitability. Don't get me wrong, profitable companies are great to invest in and can pay really nice dividends but only growth stocks will return above the 7% long-term average of the market.

Because of this, in order to maintain their stock price and keep their investors happy, those tech companies needed to find a place for that cash to grow, something that would show they still had growth potential and were worth a P/E ratio of 15-20 like a growing tech giant and not a P/E ratio of under 10 like a legacy company. AI fit the bill perfectly.

Eventually, though, AI started to cost too much, the piles of cash that were huge less than a decade ago are starting to dwindle and, increasingly, the AI boom is being funded by debt. That's not too big of a problem because, while interest rates are high relative to recent numbers, they're still way low compared to their long-term trend, so you can fund data center build-outs by borrowing money on fairly favorable terms.

The problem is that it looks like inflation is becoming an issue, and when inflation goes up, interest rates go up, and borrowing that was affordable at 4% or 5% might not be at 7%. Up to this point, the companies involved in AI have been borrowing pretty much at the maximum limit of how much money people are willing to give them. Even a small decline in their ability to raise capital or borrow money is going to be enough to tip things over to where they can no longer repay debt as it comes due.

I'm not sure we're there or if/when we ever will get there, but there's a huge risk, particularly given the level of circular financing in the industry, that even a small negative shift in the broader economic conditions could cause the whole thing to topple in on itself. It's honestly a bigger concern than it should be right now.

THE CLOSED SYSTEM PROBLEM

In addition to all of that, one of the biggest problems we have is that neither of the two main frontier labs is public.

Google is a public corporation, it has to release details of what it's doing, what it's spending, what its plans are, and some information about its products. The same is true of Microsoft and Meta. Anthropic and OpenAI are not, which means they can pretty much choose not to release… well, whatever information they want.

Seriously, no one knows anything more about the model weights, the training, the guardrails, the finances, or anything else other than what they've chosen to release publicly. Legally, there are very few requirements for what they have to do and, quite frankly, they haven't actually done all that much. What we know about the American AI models largely comes from the public disclosure of Anthropic and OpenAI's corporate partners and the occassional leak of information to journalists from sources within the company who may or may not have their own axes to grind.

But there is one thing we do hear from them all the time; how dangerous their products are.

Speaking of… stop me if you've heard this one, AI is so powerful it could doom all of humanity. Sam Altman and Dario Amodei rarely miss an opportunity to opine on how their products could be used to destroy humanity and, usually when they're trying to raise another round of funding or trying to distract from an embarrassing news story, you can find them giving interviews to friendly or non-tech savvy reporters about how we're on the cusp of reaching the point where AI, specifically their AI, could end the world as we know it.

How plausible is that? It's unclear, largely because they don't release any actual information. It's fairly clear that their models are very good at hacking and, from general use, we can see that they're also very useful for programming and mathematical proofs, though it remains to be seen if they're cost-effective, but if there's any evidence that any AI is impossibly good at any other task I haven't seen it.

What the closed system does is make it so that their opinions are the only ones on the topic that can be informed by actual facts and, let's face it, they have every incentive to exaggerate the capabilities of the product that they desperately need to sell to the public for hundreds of billions of dollars or face financial ruin. That deserves a bit of detail in its own category so…

THE DOOM PROBLEM

I'll admit the possibility that, at some future point, AI or some other technology may advance to the point that the machine-pocalypse happens, but you have to admit there's a far more likely explanation for all the talk of it today.

You see, AI companies have a problem: they're spending trillions of dollars on a product that's not currently making enough money to pay the interest on all of that debt. More to the point, in order to get to where they have a product that can make enough money to pay the interest on the debt, it looks like they're going to have to spend hundreds of billions, probably even trillions, more!

How do you get someone to give you money in that scenario? Well, you clearly can't sell it with your actual product, but what if your product was potentially so powerful as to wipe out humanity? What if it was even just on the cusp of that level of capability? That's obviously something that could be extremely profitable if put to more constructive uses.

Of course, this kind of talk is a double-edged blade because, as much as greedy investors want to fund the most powerful, capable thing, the public hears that kind of talk and turns against the product. We're already seeing this in local opposition to data centers and national demands for tighter regulation.

More to the point, the fact that AI companies have to resort to this kind of talk, making fantastical claims about their product instead of making more reasonable predictions, tells me that they might be in a little bit of trouble.

WHAT'S NOT A PROBLEM

Look, it's very clear that AI is actually really useful… at certain things.

What it's genuinely good at is statistical analysis. I mean, heck, the whole reason it can mimic human-like speech patterns is because it has statistically analyzed billions of words produced by humans and produced a statistical model that can then create something that sounds like that.

It's also really good at tasks like programming and mathematical proofs for the same reason. If it can absorb a ton of information about something and then run through a training gamut where it can receive quick confirmation on whether it did well or poorly in a repeating cycle, it's likely that it can be trained to do a thing! And AIs trained this way have already produced impressive code and developed several mathematical proofs that have bedeviled humans for decades or even more.

Where it gets really fun is when we can use that to look for patterns in things we've never found patterns in before. After training on known data sets, an AI was produced that could quite reliably find breast cancer in a lab five years before it was detectable by other methods. That's really cool!

WHY THAT'S STILL A PROBLEM

But here's the thing, in order for an AI to be useful in the ways I described above, it generally has to have a reasonable amount of training from someone who knows what they're doing and the results themselves are generally only particularly useful for a similarly skilled individual.

Now, this isn't bad, I think AI, particularly Generative AI, is still going to be extraordinarily useful in science and will power a not insignificant number of breakthroughs, but that's not great if you're an AI company.

You see, if you're an AI company, you're pouring an absolutely absurd amount of money into creating this thing and you desperately need to make that money back at some point. The best way to do this is to have as many potential customers as possible but, if the product is only really useful to someone who is fairly experienced in a given field, that's a much tighter limit on the number of potential customers than a product that anyone can use for anything.

That's a big problem because if they can't get enough money to repay all the money they've borrowed, or at least make the interest payments, the AI companies are going to go under eventually. Getting enough money means making it useful enough to enough people that the collective amount of money they are willing to pay covers the cost of development and operation. The more the amount of money put into development (and operation as the models get more complex) increases, the more you need to remove any limits on the factors involved in making money.

CONCLUSION

I mean, I think it's fairly clear that there are cases where AI can perform certain tasks significantly faster and sometimes even better than a human being. Unfortunately, all of these cases have the same two issues which are (1) they still need to be overseen by someone with significant background/education in the task and (2) none of them have the kind of widespread demand that will allow the AI companies to pay back the now trillions of dollars that have been invested in creating the product. None of what we've seen so far is even close to the "AI will replace all human labor" or "AI will overwhelm humanity and bring about the apocalypse" situations that you hear so many people talking about.

And look, maybe one day we'll have computer programs so advanced that they will be able to match or even outmatch human beings on a wide range of things, but it seems unlikely both that we're on the cusp of it now or that the current AI technology will get there anytime soon. As far as I can tell, the claims in that direction seem to be mostly motivated by a need to raise money, keeping the current companies in business in the hopes that a few more months will allow them to discover a use case that could generate enough revenue to recoup the trillions of dollars they now owe from their efforts so far, rather than any genuine analysis of the product itself and the trend of its capabilities over time.

As usual, let me know if you think I missed anything or if you think I'm wrong about something important, I'm always happy to hear things I haven't thought of.

[uboat53]

Well, with the death of Sen. Lindsey Graham (R-SC), we're treated to yet another round of the sanctimonious games where people who opposed him insist that death is no reason not to criticize and people who supported him declare that to say anything ill of the dead is wrong and should be punished. But how should we actually address the memory of the recently dead?

I have some thoughts on the matter, LONG RANT (TM) time.

PERSONALLY

Look, I've got my moral system and I hold to it; I don't think death is ever positive and I won't celebrate the loss of another person's life.

That said, I think we do have to acknowledge that the effect that some people have on the world is so negative that the world is a better place when they die. How do we square these two ideas?

Well, I think the thing that ties them together is that, yes, this person had a terrible influence on the world, but it wasn't inevitable. It was the result of choices they made and the influence that others and the world at large had on them, but it didn't have to be that way. The same person in a different story can turn out entirely differently.

With that in mind, the death is still a tragedy. Not because it made the world a worse place, but because it's a reminder that they could have made the world better had things been different. Every human life is a spark of potential and, if we can't mourn the actual result of their life, the loss of that potential is still worth remembrance.

THOSE WHO LIKED THE DECEASED

Obviously those who liked the deceased are going to want to remember the positive about them, and why shouldn't they? Even the worst people in history have their positive sides and people who loved and cared about them and it would be foolish to pretend otherwise.

Those people deserve to remember the person they knew. So how do we square that with…

THOSE WHO HATED THE DECEASED

This isn't a problem most of the time, most people simply don't have the level of impact that allows them to be broadly hated and all but the most heinous of us usually have a hard time making more than one or two enemies who genuinely hate us. But for public figures, that number can be much higher and with good reason.

In my opinion, these people fall into three buckets. The first bucket is those who are hated by don't deserve to be. Some people spend their lives trying to make the world better and genuinely do, but are hated by one side or the other largely due to political propaganda campaigns. Martin Luther King Jr., for example, worked to undo racial hatred and oppression. In terms of how it affected people, his work was undeniably good and made the world a better place, but there were many who benefited from that oppression and they worked very hard to make him hated among a swath of Americans.

The second bucket is those who try to make the world better, but don't really succeed. I'm sure many would disagree, but I put George W. Bush Jr. in this category. From his advocacy for AIDs relief in Africa to his efforts to promote equal opportunity in education to his seemingly genuine belief that the Iraq War was a war to spread democracy, I think that he really did want to make every person, even those who didn't support him, better off even if I might disagree with specific actions like No Child Left Behind or the Iraq War itself. I can recognize that people with different perspectives will come up with different ideas of how to make the world better and I can recognize that even some of my own ideas might not actually work, but still acknowledge the intent behind it.

Finally, the third bucket is the people who are hated for well-deserved reasons. Henry Kissinger, for example, spent his entire life advocating for brutal war crimes and the suppression of liberty and democracy around the world. People around the world, from Vietnam, Cambodia, and Laos to Bolivia, Chile, and Argentina to East Timor and Bangladesh and many, many more besides were murdered in huge numbers and left permanently scarred by his direct and intentional actions. I think any reasonable person would agree that both the result and the intention were negative for the majority of people impacted by his actions.

PUTTING IT ALL TOGETHER

So what is the correct way to remember a deceased public figure? Well, I think it depends on who and where you are.

If you were directly harmed by the deceased and it was their intention to harm you, I think however you feel is fully justified and no one should be allowed to tell you otherwise. If you were harmed but the intention wasn't there, I still think you're fully justified to feel how you feel; no one is entitled to forgiveness, especially if they haven't taken any action to atone.

If you weren't directly harmed by the deceased, but they intended to and did cause significant harm, I think it's still justified to criticize them even on the occasion of their death. Dying doesn't make anyone a saint and it is still important to public discourse to make clear that certain actions are beyond the pale. Even if someone didn't intend the harm they caused, I still think criticism is valid, though it should be tempered based on that difference.

What about those who didn't cause harm but are still hated? Honestly, this is a tough one because most of the time the people who hate them have convinced themselves that the person did cause harm in order to justify that hatred. Obviously they'd put themselves into one of the two categories above and insist on their right to criticize. I would say that, in this case, those who supported the deceased need to weigh in with their own counter-arguments. If they truly believe that the hatred is unjustified, they can point to how the deceased actually helped or intended to help those who are angry. If they can only point to situations where the deceased helped their own group by harming those who criticize them, that should provide all the evidence necessary about the validity of their defense and the criticism.

As to those who support, like, or love the deceased? Well, apart from arguing against criticism, their primary responsibility, if they wish to mourn the deceased in peace, is to tune out the public discourse. Unfortunately, this is simply part of the cost of being a public figure and you don't get to dictate the entirety of pubic discourse even if you are grieving. If you want to mourn the deceased in peace with like-minded individuals, you're going to need to shut yourself off from the public for a time.

LINDSEY GRAHAM

So what about the most recent deceased public figure? Well, Graham was one of the leading American voices in favor of restricting the freedom of speech, a leading opponent of privacy laws, particularly in relation to national security surveillance, a proponent of removing legal protections from those accused of terrorism, he attempted to remove the Affordable Care Act's protections for people with pre-existing conditions, he attempted to make all but the earliest abortions illegal nation-wide, and he was a major proponent of the wars in Afghanistan and Iraq as well as a backer of the recent conflicts in Venezuela and Iran and Israel's actions in Gaza. On the other side of things, he was a rare GOP voice pushing back on vaccine conspiracies, a major supporter of animal welfare, a prominent backer of Ukraine, and was one of the few GOP figures pushing to address climate change.

Hey, I never said people were simple, every person has good and bad in them.

Now, depending on how any or all of those issues affected or is affecting you, you may feel differently about him. Personally, I think that the bad outweighs the good and I don't think that the fact of his death should prevent me from saying so. That said, those who think that the good outweighs the bad are also free to argue that I'm wrong.

What I think no one is free to do is demand that any point of view be silenced. Part of being a public figure is that the public must always be free to speak openly about them. The private individual still exists, but those who wish to maintain their focus on that private individual must then make the choice to avoid the public forum.

CONCLUSION

So yeah, that's my take on it. I, personally, won't celebrate any death and, though I accept that there are people for whom the world is better because of their death, I take that as a tragedy of what might have been rather than a cause for celebration.

Still, a public figure is not a private individual and most of those touched by their actions will never know them in person. The occasion of their death is not a reason to silence the public debate and those who wish to mourn or celebrate the private individual in peace must accept that doing so means removing themselves from the public discussion for a time.

[uboat53]

I read something recently about Senator John Fetterman (D-PA) that gave me an insight and the language for something that's kind of been rattling around in my brain for a while. LONG RANT (TM) time.

INTRODUCTION

So there's two things that kind of came together for this thought. First, there's the old truism that what's really different about billionaires as opposed to the rest of us is that they don't experience pushback when they pitch crazy ideas. Being that wealthy means that, unless you deliberately work against it, your immediate social orbit will be filled with people who want to get on your good side in order to access and influence some of that wealth and power. In other words, the sort of people who are going to go along with whatever you come up with.

Second, I read about David Safier. If you're not aware, Safier is a fundraiser for Orthodox Jewish causes who's become unusually close with Sen. John Fetterman (D-PA) to the point where he's regularly sat in on sensitive meetings with foreign dignitaries, attended floor votes with the Senator, and spent entire days in the Senator's office, often with the Senator's staff excluded. Staff and former staff for Fetterman, usually, though not always, speaking anonymously, have spoken about how Fetterman's views on key issues would change significantly after spending time with Safier.

What do these two things have to do with one another? Well, they reveal that one of the unspoken assumptions about the wealthy and powerful is far less true than I think we would wish it to be.

POWER AND INFORMATION

Wealthy and powerful people have a lot of power, it's kind of in the definition. Jeff Bezos, for example, worth over $260 billion at the time of me writing this, can do a lot of things that would be quite literally impossible for most of us. He could buy up entire towns, get meetings with political leaders just about anywhere in the world, and, on a daily basis, makes decisions that affect the livelihoods of millions of people. And it's not just wealth, political power can be much the same. I guarantee you that, if he chose to, Senator Cory Booker (D-NJ) could probably destroy the life of any average person in the country by barely lifting a finger.

On the flip-side, they also have access to a wealth of information. Not only do wealthy and powerful people have access to vast apparatus' dedicated to combing huge reams of data, but they have the resources to drill down and get even more detailed information about specific subjects if they want it. If, for example, you wanted to know the life history of a random guy in the town of Chandragiri outside of Kathmandu in Nepal, most people probably wouldn't even know where or how to start. Someone with substantial wealth and power, though, has access to people and institutions that, with a little bit of effort, could probably find out every detail about the guys' life.

Put those two together and you have what I think is most people's impression of this particular group; that they wield effectively unlimited power guided by effectively unlimited knowledge.

THE HUMAN REALITY

And that's where the John Fetterman story really crystalizes what's wrong with that impression. I mean, it's not entirely wrong, as a U.S. Senator, John Fetterman certainly has access to an incredible wealth of information on just about any issue, but that access means pretty much nothing if he's not using it. Information is only useful in proportion to your ability and willingness to both absorb it and contextualize it and, even in the best of cases, no human being has the ability to absorb the sheer quantity of information that modern technology is capable of gathering and, since no person can have access to all contexts, contextualization of information is dependent on who is around to provide perspective.

All the information in the world about a Nepali person won't help you if you don't have someone with the perspective to understand what all of that means in the context of actual Nepali society. Knowing the full mathematical model of the Theory of General Relativity won't help you in the slightest if you don't have someone with the background to understand what the equations actually mean. Trust me, a physicist, on that one.

In other words, there are two bottlenecks in fully utilizing what wealth and power gives you: what existing framework you have to deeply understand the information and what information you actually get in the first place.

WHERE IT ALL FALLS DOWN

Remember what I said at the beginning about wealthy and powerful people being surrounded by people who won't push back on anything they say and want to influence them? Yeah, that's where the whole "rich people have access to all the information in the world" thing falls apart. These people have access to more information than anyone else in the world but, unless they make a ton of effort to change it, they're surrounded by an apparatus that limits what information actually gets to them and deliberately tilts the context of that information.

Much like Greek gods, the people we're talking about have access to power far beyond what most of us do, but they're still subject to the same limited view as the rest of us humans. They have biases, they have blind spots, and, even more so than the rest of us, they lack the kind of guardrails and filters that are necessary for moving beyond those things.

In other words, there's no reason to invent complex conspiracy theories to explain why wealthy and powerful people do things with their vast power that don't seem to line up with what they say they're trying to accomplish or even that seem to run entirely counter to those things, you simply have to understand that they're not able to fully realize the wealth of information that is theoretically at their disposal and are, instead, acting on a very limited subset of it that is presented in a very specific context by a constellation of people around them who want to direct their actions to their own benefit and then filtered through their existing limited experience.

CONCLUSION

Look, I think we've gone way too far in the direction of treating wealthy and powerful people as gods. I get it, they have access to options and information that most of us can only dream of, but they're still human beings and subject to the same biases, limitations, and just plain fickle nature as the rest of us.

I think this has very broad implications for the role of wealthy and powerful people and how we perceive them in our society, but the most immediate thing I think that people should do is to stop viewing the pronouncements of this group of people as if they come from on high. Understand that they're people with a limited, manufactured, and often deliberately warped, often by others with ulterior motives, perspective and treat their views and actions accordingly.

[uboat53]

Does J.D. Vance believe in God? Look, I'm not usually one to question another person's faith, my position is that a person's faith is just that, personal. But there are some people who use their faith in public debate and, in those cases, I think it's fair to ask the question, and this is no more true than in cases where a glance shows you some deep misrepresentations of that faith. So let's dig in on Vance.

ORDO AMORIS

My questions about Vance's faith started with the Ordo Amoris thing. I'm sure you remember, about a year ago the Vice-President went on Sean Hannity and quoted St. Augustine in order to explain how "deranged" liberals were for having compassion for migrants. Specifically, he said the following:

"There’s this old school—and I think it’s a very Christian concept, by the way—you love your family, and then you love your neighbor, and then you love your community, and then you love your fellow citizens in your own country. And then after that, you can focus [on] and prioritize the rest of the world."

Source

In response to pushback from British academic and politician Rory Stewart, Vance then tweeted out the following:

"Just google “ordo amoris.” Aside from that, the idea that there isn’t a hierarchy of obligations violates basic common sense. Does Rory really think his moral duties to his own children are the same as his duties to a stranger who lives thousands of miles away? Does anyone?"

Source

He then got a bunch more pushback from just about every academic who has studied Augustine (or Aquinas, or Christian doctrine in general) because, well, Christians believe that! It's actually a key point of Christian theology.

What is Vance talking about? Well, the concept of Ordo Amoris, Order of Love, comes from Augustine's book "De doctrina Christiana" or "On Christian Doctrine/Teaching" in which he writes the passage:

"Further, all men are to be loved equally. But since you cannot do good to all, you are to pay special regard to those who, by the accidents of time, or place, or circumstance, are brought into closer connection with you. For, suppose that you had a great deal of some commodity, and felt bound to give it away to somebody who had none, and that it could not be given to more than one person; if two persons presented themselves, neither of whom had either from need or relationship a greater claim upon you than the other, you could do nothing fairer than choose by lot to which you would give what could not be given to both. Just so among men: since you cannot consult for the good of them all, you must take the matter as decided for you by a sort of lot, according as each man happens for the time being to be more closely connected with you."

Source

This seems to be exactly what Vance is talking about, but you'll note it's not as strident as he is. It doesn't, as he seems to imply, imply that the rest of the world should be ignored until those close to you are fully taken care of, it simply says that, in cases where we cannot help everyone all at once, we should help those close to us first. Reading more Augstine pretty much confirms that this is the case. In his book "De Moribus Ecclesiae Catholicae et de Moribus Manichaeorum", for example, Augustine says the following:

"Loving your neighbor as yourself does not even mean loving in the way members of one’s immediate family, brothers, sisters, children, wives, and husbands are loved, or any other kith and kin, or one’s next-door neighbors or fellow citizens. . . . After all, we would not have relationships of that sort . . . if our nature had remained true to the precepts and the image of God."

Source

I've skimmed various writings of Augustine, I don't claim to have done more than that, and even then it's pretty obvious that he's not talking about selfishly hoarding for those close to oneself, but merely addressing the very specific case where two people are in front of you needing your help RIGHT NOW and you can only help one. In fact, if you read the passages immediately surrounding the Ordo Amoris one I quoted above (you can find them in the source link I provided), you'll see that the context provided by them also confirms this particular understanding of the concept.

So if it's that easy to understand, how did Vance get it so wrong?

READING FOR QUOTES

Look, I'm not going to judge here, we've all been there. It's the end of the quarter, you've got three papers due in the next two days and you just need to get them out the door. No time to read and reread each book enough that you can fully comprehend all the nuances and find the best possible quote to illustrate the complexity, you've just got to find one and go, so you just skim the thing, find one that seems like it supports your argument, and you put it in and move on to the next one.

I've done it a dozen or more times and I'm pretty sure almost everyone else has too, my point is that I know what I'm looking at when I see it. How do you pull something off the paper that contradicts not only the overall context of the writing(s) but also the immediate context around it and present that as the representation of the writing? Reading for quotes.

The way you get here is you decide what you think the text is supposed to say and you skim the text until you find a passage that sounds like it can support what you've decided it means. But here's the thing, does that sound like what you do with something you care about?

DO YOU CARE?

I mean, look, maybe it's because I'm an atheist, but I feel like, if I genuinely believed that the eternal fate of my immortal soul was riding on correctly understanding a set of teachings, I would put a bit more effort into it than this. I mean, even as an atheist, I spend a good deal of time and effort trying to understand and shape my own morality even when there's no eternal reward or punishment riding on it, I can't imagine doing otherwise if it were even more consequential.

It's not as if this kind of stuff is supposed to be small potatoes either, if you actually believe in the teachings of the Catholic Church, then you can literally be condemned to an eternity of torment if you get this kind of stuff wrong. It seems like it'd be pretty important to get this right and not just read a bunch of stuff for quotes to confirm whatever you came up with on the fly.

But here's the thing, saying that Vance doesn't care seems a bit too simple. Obviously he does care in some sense because he spends a great deal of time on it. Just listen to him giving interviews, it's very clear that he's at least familiar with a fairly broad range of Catholic doctrine and thought and that he spends a good deal of time reading, thinking, and talking about it.

So why did he make such a basic error in reading Augustine? I think the answer goes a little bit deeper than just whether he cares or not.

WHO SERVES WHO?

This idea kind of came to me a bit more recently when the news broke that Secretary of Defense Pete Hegseth had recited a prayer that was not actually from the Bible, but from the movie Pulp Fiction. The quote was:

"The path of the downed aviator is beset on all sides by the iniquities of the selfish and the tyranny of evil men. Blessed is he who, in the name of camaraderie and duty, shepherd the lost through the valley of darkness, for he is truly his brother’s keeper and the finder of lost children. And I will strike down upon thee with great vengeance and furious anger those who attempt to capture and destroy my brother. And you will know my call sign is Sandy One when I lay my vengeance upon thee, and amen."

Source

Why is the Secretary of Defense, a man who claims to be an extremely devout Christian, quoting movies instead of his own actual sacred texts? Could it be, perhaps, that the sacred texts don't say what he wants them to say? If you're not reading Bible quotes out of context or confusing movie dialog for them out of laziness, what other reason could there be?

And that's where I had the most interesting thought. You see, if I believed that the teachings of the Catholic Church and the Christian faith were literal truth and that the eternal fate of my immortal soul rested on understanding and carrying out those teachings, I would approach them with the goal of understanding them regardless of what they said. I would read them in full, reread them in context, read related passages together, and do this over and over and over again over the course of my lifetime to try to determine what I should be doing so that I can, you know, do it.

On the other hand, if I wanted to justify whatever it was that I wanted to do anyways… If I wanted to make my arguments in a way that would make anyone else look like the bad guy if they disagreed with me… I could do a lot worse than to argue that my beliefs were the justified teachings of the largest religion on Earth.

One way requires you to serve God, the other way makes God serve you.

CONCLUSION

Like I said, I'm not often one for questioning other people's faith, but when a public figure makes their faith the cornerstone of their policy arguments in a way that makes it incredibly clear that they haven't made the slightest effort to actually learn the teachings, then it becomes impossible to deny the simple truth.

J.D. Vance, among many other publicly devout members of the current administration, isn't religious in the sense that he seeks to better understand the teachings of Christianity in order to serve God, he's religious in the sense that he reads the teachings of Christianity in order to figure out how to make God serve him.

If you can think of a better example of religion without true faith, I'd love to hear it.

[uboat53]

Continuing off my recent comments about American corporate profits, I need to think something through. LONG RANT (TM) time?

INTRODUCTION

So here's the thing, pretty much everyone agrees that the biggest problem in the American economy today is that everything is increasingly becoming unaffordable, but no one has figured out what to do about it. We also don't need to pretend that this is a symmetric problem in American politics; Democrats have a lot of ideas to try to address this issue, Republicans have none. Only people who are deluded or lying would think that the party that simply cuts taxes for rich people and repeals protections for workers every time they're in power while doing absolutely nothing else and not even discussing any other actions is going to do (or even wants to do) anything to solve this issue.

Given that, we're looking at the Democratic ideas, but how do they measure up? Well, the Obama administration took pretty strong action to try to address health care costs, but they didn't do a great job revitalizing the job market (I know, I lived the first decade of my career through that period). The Biden administration took much stronger action in terms of putting money in people's pockets and showed pretty conclusively that, the US having a largely consumer driven economy, this creates a pretty hot jobs market. Unfortunately, it also showed that it triggered unacceptable levels of inflation.

Look beyond the recent administration's efforts to rebalance the economy in favor of workers, you can find all sorts of ideas floating around the American left, from the center to the far left. I'm thinking of ideas like strengthening union protections to Medicare for all or free college or any of the numerous other regulations and any stimulus plans for various industries.

Just as important as the ideas, though, is how they are implemented. That's what I want to take a look at now, looking specifically at the ideas that have been implemented so far, what effect they've had, what the relevant factors of the economic situation are, and what that might indicate for future efforts. So let's get to it.

OBAMA

For those who don't remember, Obama inherited a real crapsack of an economy. Between the convention and his inauguration, the economy had contracted 2% and would contract a further 1.5% by his first summer. His first year had a total GDP growth of 0%, it literally went right back to where it had started.

Because of this, in his first year, Obama and the Democratic Congress he had to work with passed a lot (and I do mean a LOT) of legislation not only to try to restore the economy, but to do so in a way that rebalanced it away from high finance and the rich. The American Recovery and Reinvestment Act of 2009, the Credit CARD Act of 2009, the Dodd-Frank bill, the Tax Relief, Unemployment Insurance Reauthorization, and Job Creation Act of 2010, and, of course, the Patient Protection and Affordable Care Act, better known as Obamacare.

Did all of this make a difference? Maybe? Corporate profits were fairly flat through the 2010s despite GDP continuing to increase at a steady pace during this time. Now, I'm not exactly sure where that increase in GDP went, though, because worker's share of the pie didn't increase either. Overdraft fees have been fairly flat since Dodd-Frank kicked in by the mid 10s and health care's share of the economy has been fairly flat since Obamacare, arresting their precipitous increase in the post-WW2 period.

Overall, I would say that Obama's efforts made things better than they would have been, but it's fairly clear that it didn't address the real long-term drivers of our issues.

BIDEN

Like Obama, Biden came to office with a crisis and a Democratic Congress and passed a lot of legislation quickly because of it. The American Rescue Plan, in particular, was designed with the lessons of the Obama presidency in mind. Where Obama had passed a smaller plan in response to the Great Recession with the expectation that they could go back and do more if necessary only to see that idea shot down by Congressional Republicans, the Biden administration pushed for a maximalist bill. The ARS injected a huge amount of money not just into the American economy, but directly into the pockets of the average American.

In 2022, Biden managed a rare bipartisan piece of legislation, the Inflation Reduction Act. Among other things, this legislation attempted to reduce prescription drug prices, increase tax enforcement, and injected a huge chunk of money into energy infrastructure and climate change mitigation.

There's some (admittedly mixed) evidence that this combination of legislation arrested a sharp spike in the poverty rate caused by the effects of the Covid19 pandemic, that it may have started to bend the cost curve for prescription drugs, that it returned people to work after the pandemic shock, and that it caused wages to increase at a faster rate than they had before, but one thing that most certainly came out of it was the largest increase in inflation in two generations.

The causes of the inflation surge are certainly still debated, with the pandemic shock to supply lines, the massive infusion of cash to consumers from the government, and straightforward corporate greed all believed to have played a role even if the extent of each isn't 100% clear. But here's the thing we can see from the data. Take a look at corporate profit data again, notice how it accelerates sharply starting in 2020 after having been flat for nearly a decade? Then take a look at labor share data one again, notice how it drops sharply starting in 2020 after having been flat for a decade?

LESSONS LEARNED?

So here's the thing… after the pandemic-era supply difficulties worked themselves out, the main reason given for inflation and price increases generally was increased labor costs and, if you look again at the hourly earnings chart, you can definitely see that the rate of wage increase post-2020 was higher than the rate pre-2020, but it doesn't match with the data that shows labor declining in its share of the economy or corporate profits increasing dramatically in that same period.

Because, here's the thing, for both wages and corporate profits to be going up, that means that prices had to increase faster than the rate of wage increase. In fact, that's exactly what you see when you look at the labor share graph. That means that, underlying all of this, is the fact that corporations, on average, have enough pricing power to pass more than 100% of their increased costs on to consumers. It sure sounds like what you would expect in a situation where the majority of industries constituted a monopoly/oligopoly or a monopsony/oligopsony (or both!).

Obviously, I'm not an economist, I don't have any relevant degrees or experience, but I don't think this is a particularly difficult case to make. I was able to pretty quickly show the consolidation of a bunch of different industries when I made the case for corporate greed as an inflation driver a few years ago and the situation has not exactly gotten better since then. When you have an industry that is controlled by a single player or just a few major players, you get fairly predictable market behavior, and when the entire economy consists of such consolidated industry, you would expect to see that same behavior.

The other problem is that American voters are getting increasingly impatient. I don't blame them, the last 45 years have been absolutely brutal to the average worker/consumer. Still, what it means is that the room to experiment is not very large. Most experts pretty much agree that the actions the Biden administration was taking would have, over time, shifted the balance of the economy, but no one was willing to give him that time. Whatever actions the next Democratic administration takes are going to need to be targeted and work fast, the political situation does not give the time or space for experimentation anymore.

LESSONS INTO ACTION

So what should they do? I may be wrong, but all of the above seems to point to aggressive anti-trust action.

Think about it, if you can increase wages and even just give people money but all of that just gets eaten up by corporations because they have enough market power to just vacuum up all of that the, if you want your efforts to actually get to workers and consumers, you have to take market power away from big corporations that dominate the various industries of the US economy.

Could this work? Just as importantly, could it work fast enough to make a difference? Well, I'm going to look at, in my opinion, the most effective anti-trust action in the last generation.

In 2013, Sprint started targeting T-Mobile for acquisition. Amid opposition from the Obama administration which felt that this would unacceptably consolidate the US cellular industry, the deal fell apart. In the wake of this, since it no longer had the hope to acquisition, T-Mobile became hyper-competitive. Between 2013 and 2024, it turned itself from barely profitable (it actually had not been profitable before 2013) with the smallest customer base of the national cellular carriers into a behemoth rivalling Verizon and AT&T, the two giants. In fact, in 2020, T-Mobile acquired Sprint in a complete reversal.

One huge effect of this sudden competitive streak is that the average price of a cell phone plan dropped substantially between 2014 and 2017, only rising again after the reconsolidation of the industry in 2020. Just as the theory predicted, increased competition was great for the consumer.

Even more important, though, is how quickly all of this happened. Prices stopped increasing in early 2014, pretty much as soon as the merger ran into trouble, and started dropping in early 2015, as soon as it was conclusively dead. In other words, this is a political focus that would likely have its impacts quickly enough to prevent the rapid losses of power that have almost immediately defanged the last several Democratic administrations. It may, in fact, be the key to unlocking all of the other genuinely good plans to rebalance the economy in favor of workers and consumers for the long-term.

To be fair, the Biden administration did actually move more aggressively than any administration in my lifetime to enforce anti-trust laws, but they did it entirely within the existing framework of laws which were designed to prevent the rise of such a situation rather than address the situation once it already exists; playing whack-a-mole on an otherwise flat field as opposed to mowing a uniformly high field of grass. It seems there's a good case to be made that existing law is not enough and that we need actual legislation to do what is needed.

CONCLUSION

Democratic presidents have been trying as best they can over the last two decades to re-balance the American economy in favor of workers and consumers, but every one of their efforts has run into the stone wall of market power that corporations in the majority of American industries hold. To have the best change of making any progress in the next Democratic administration, we need to make anti-trust administration not just "a" focus, but "the" focus of, not just policy, but legislation right out of the gate.

If it works the way previous anti-trust efforts have, it will likely also have the side-effect of having noticeable economic impacts in time to mitigate the usual mid-term rout driven by voter's frustration that things aren't getting better fast enough. With at least a little bit of luck, it might give Democrats double (or even more!) their usual two years to implement policies that could finally start bending the long-term arc of the American economy back in favor of workers and consumers.

That's my thought at least. Let me know if you think I'm missing something or if I got something wrong in a way that undermines my argument.

[uboat53]

You know, I recently had an epiphany about why I'm neither bullish nor bearish on AI. My social circle is pretty wide and encompasses people who are both euphoric about AI and it's possibilities as well as people who are absolutely terrified of it and everywhere in between. As for myself, I think what it is likely to do will be impressive, but not necessarily a change of direction from the changes that have been taking place my entire life.

Why I think that, though, hasn't been something I could really articulate until a fairly recent exchange of ideas.

LET ME EXPLAIN

I was born in 1985. Room-sized computers had been around for decades, but were pretty much only available to the extraordinarily wealthy or to institutions. When I was 6 years old, we got our first personal computer. When I was 9, we got access to the internet. When I was in middle school, laptops started to become widely available. When I was in high school, cell phones started becoming widely available. When I was in college, smartphones started to come out. Nowadays, I can pull over on the side of the road, tether my laptop to my cell phone, and have a business meeting with a customer in Singapore during which we can share reasonable resolution video of both ourselves and our computer screens and exchange files, then I put it all away when I'm done and start driving again.

That last one isn't theoretical, by the way, it's something I've quite literally done in the last few years that would have been completely impossible in 1985. But what do all of those things have in common? Well, all of them either allow more data to be processed or allow greater access to existing data.

RELEVANT ANECDOTE

In 2009-2012 I worked for Raytheon on an army radar system. At the time it was being upgraded to use this newfangled technology called computers. I know, the army is always on the cutting edge.

The thing I noticed, though, after having worked on it for a bit, was that only some parts of it were being upgraded from analog to digital; the radar dish itself wasn't being touched. In fact, it was the same dish that had been used since the system was first designed in the late 70s.

The upgrade, though, had a huge impact in terms of capability. Where the old analog process/display system could show you one track, it was an artillery detection radar, so it would show origin and projected impact points of the projectiles, and only one side of the track, at a time, the new computerized process/display system allowed the user to observe monitor 15 tracks, and both origin and impact points for all of them at the same time; a 30x increase in capability! How was this possible, though, it was the same dish?

Well, the simple answer is that the dish had already been collecting that data. For half a century it had been collecting orders of magnitude more information than could actually be processed or accessed and, in fact, it still was. If the military had decided then to install even more powerful computation onto that system, there would still have been more data that could have been made available.

In other words, the computers didn't change anything about the data itself, it just allowed us to process and access much more of it than we could before.

PROCESS AND ACCESS

And that's the basic fact of all of the technology that's come out in my lifetime as well as a good deal of what came out before. Inventions like electricity, radar, lasers, and others allowed the collection of much, much more data that in previous centuries, but we haven't really had much of that kind of innovation lately. What we have had is an innovation revolution in how to actually use that data that we already had.

Computers, both the old room-sized ones and the newer personal computers, allowed us to process more data. The internet allowed us to access more data. Cell phones allowed access to more data in more places. Smartphones affected both access and processing. Each step has taken more of the data that was already available and has been for decades and allowed us to process and access more of it than ever before.

Now, to be clear, by processing and accessing more of that information, we've been able to make new discoveries and this has, itself, generated more data, but this generation of new data has come nowhere near approaching the pace at which we have increased our ability to process and access.

SO AI…

AI, as far as I can tell, seems to do more of the same. Most of the improvement seems to be in form of processing, being able to crunch very large sets of data very quickly, but there does seem to be some improvement in access as well. This is impressive but not functionally different from the direction of improvements that I have seen my entire life.

Will things change as a result of AI? I think it's not only possible, but likely. Will AI fundamentally reshape the world? I'll admit I'm doubtful of that. None of the other dramatic improvements in our ability to process and access data have done so. Is it possible that this one is different? Sure. Do I think it's likely? Let's just say I haven't seen convincing evidence that it is.

CONCLUSION

I guess if I have a conclusion, it's this:

The odds that AI will fundamentally alter the world, the economy, or human society in either a catastrophic or utopian direction seem fairly low to me. Knowing what I know about it so far, it seems to me to be mostly a continuing improvement in our ability to both process and access data of the same kind that have defined improvements in technology for most of my life.

Don't get me wrong, those improvements add up, the world that my daughter was born into in 2020 is wildly different than I one I was born into in 1985, but it's not foundationally different. People still work for a living and we still have pretty much the same government. The details may have changed, but the fundamental structure of society and the world is fairly similar.

And that's pretty much my view on it. It's a big change, but not unlike most of the changes I've already lived through. Let me know if you think I got something wrong or if there's something important I'm missing, though, I'm always interested to hear new ideas.

[uboat53]

As you may have guessed, I'm the kind of guy who consumes a lot of content, both on social and traditional media, about politics and political philosophy. As such, I see a lot (and I do mean, A LOT) of really bad opinions and arguments being made. Most of those boil down to fundamental misunderstandings about how things actually work.

So, without further ado, here are a few things that I think would make a massive difference in the current state of political understanding in the United States (and probably other countries as well) if people actually understood them.

POLICE CAN CAUSE A RIOT

Look, what would you do if you were in a place where you felt you had a right to be expressing something you felt you had a right to express and a bunch of people started to threaten you with violence for it? What if you were with a large group of people? What if one of those people started to act on the anger you're all feeling?

Just about any expert in police studies knows that the standard police tactic of meeting demonstrations with a show of force is the fastest and easiest way to change a peaceful protest into a riot. If the average American knew that I think their understanding of public protest and the way it gets portrayed would be very different.

GANGS FORM IN AREAS WITH LITTLE LAW ENFORCEMENT

Imagine you lived in a place where law enforcement wasn't doing its job, what would you do? Would you just tolerate the crime happening all around you or would you band together with friends and neighbors and try to deal with it? To phrase the question is to know the answer.

But here's the thing, what happens when law enforcement does finally show up but, instead of focusing on the petty crime that you formed to try to address, they focus mainly on you? With petty crime still happening and law enforcement not dealing with it, your friends and neighbors are still turning to you for help and protection, you're still needed, only now you're battling both the crime and the police at the same time.

Congratulations, you've become a gang.

And this isn't a new phenomenon, it's why just about every racial/immigrant enclave that's negelected but also persecuted forms some kind of gang or organized crime structure, from African-Americans to Jews, Italians, Irish, and more back in their own days.

The key is under-enforcement, letting petty crime go unaddressed, and over-enforcement, targeting people who organize as a result. If you have those two things, you will have gangs and organized crime. The under-enforcement allows the gang or crime syndicate to foster goodwill and receive protection in the neighborhood by dealing with the petty crime while the over-enforcement turns the neighborhood against the cops and provokes a violent response from the gang/syndicate.

If more people understood that, it would go a long way toward actually addressing the conditions that lead to gangs, gang violence, and organized crime generally rather than doing exactly the things that we know cause their formation.

RAGS TO RICHES ISN'T REAL

In a perfectly fair society, anyone born into any income level would have an equal chance to achieving any other income level. In other words, only 20% of those born into the bottom 20% of income would stay there and the same with those on the top.

Instead, 43% of those born in the bottom 20% stay there, as do 40% of those from the top 20%. The chances of somebody from the bottom 20% making it to the top 20% are only 4% and the chances of the reverse are only 8%.

And here's the thing, if it weren't great, but we still had the best economic mobility, that would be one thing, but the United States actually has one of the worst economic mobility scores in the industrialized world, with only the UK and Italy having lower. In other words, if you're born poor and don't want to stay that way, you're better off living in just any other developed country than in the US.

The fact is that the best way to predict your income as an adult in the United States is to know what zip code you were born in and if more people understood that I think public opinion on most economic policy would look very different.

THE ONLY THING BUSINESS CARES ABOUT IS MAKING MONEY

Look through the news on any given day and you'll see businesses trying to convince you that they care. They're always announcing some charitable donation or a new hiring focus or a new product line that indicates something about some issue that's important to people. And, look, I don't want to belittle that, those kinds of things are important, but they're not important because the business actually cares.

You see, businesses, especially corporations, have one job: to make money for the people who own them.

There are many ways to make money, you can make a quality product, you can make a cheap and easily affordable product, but one of the key ways that businesses have discovered is branding. By being associated with good things, people will want to buy them more. This especially helps when the level of affordability and quality are reasonably similar between different products.

If you're going to bring in new customers, you have have to hit that balance between quality, price, and positive association and, because older people tend to have more solidified preferences, that usually means you're fighting to create a positive association in the minds of young people. When young people were reported to be powering Democratic wins, you saw businesses rush to adopt DEI policies and promote black and female workers, executives, and collaborations. When young people were reported to be powering Republican wins, you saw those same businesses drop DEI policies without a second thought.

Don't get me wrong, it's good and important that businesses do good things, but never fool yourself about why they're doing it and don't ever be afraid to use the power of your wallet, either individually or as part of a group, to push for the good things you'd prefer.

PEACE DOESN'T JUST HAPPEN, WAR DOES

One of the constant conceits of the anti-war movement is that, if they can just stop people from going to war, if they can stop companies building weapons, etc, then we will just have peace. This results from a fundamental misunderstanding of what war is.

You see, war isn't some aberrant outgrowth of individual greed or immorality, war is, to quote 19th century military theorist Carl von Clausewitz, "War is the continuation of policy with other means."

What does that mean? Well, politics (policy), is the process by which we decide how to divide up scarce resources. Those resources may be things like iron or land or people or even abstract ideas like dignity, but however we choose to divide them up and whoever we choose to give them to, the process by which we get there is politics.

(As a side note, this is why I get so annoyed by people who tell me that they don't like to talk about politics. It's an incredibly entitled attitude to be at the point where you are so secure in the resources that are given to you and so uncaring of others in need of resources that you don't even want to be bothered with any of it.)

A lot of the ways that we divide up those resources are peaceful, think things like taxation and redistribution at the national level or international bodies that adjudicate territorial disputes at the international level, but there's nothing that says they have to be. War isn't some aberration of humanity, it's what happens when the peaceful methods of resolving differences cannot produce a settlement that is minimally acceptable to all parties and at least one party thinks that they could get what they want, or at least more of what they want, though violence.

Understanding this would make the anti-war movement in particular more effective and would also give the average person a greater understanding of the kinds of policies that would produce a more peaceful and prosperous society and world.

CONCLUSION

Obviously there are more and, as I think of more, I might write them up as well, but these form a good start. These misunderstandings in particular form the basis of a lot of misunderstandings of criminal justice policy, law enforcement policy, economic policy, foreign relations policy, and just politics in general and I think it would be very good if the average person were made aware of them, even if just in summary form.

Anyways, that's my carping for the time being. As usual, let me know if you think I messed something up or missed something important.

[uboat53]

All right, time to talk about the arguments going on around AI, specifically the disconnects I've noticed in the discussion between those who are against the broader increase in AI and those who support it. Time to dive into a LONG RANT (TM).

INTRODUCTION

One of the interesting things about having lived the life I've lived is that I have friends in just about every niche. I grew up in a small, poor, conservative, and highly religious town, I went to a college whose politics were generally centrist as far as universities go, I worked in the defense industry, in education, and in the broader tech industry, I've lived mostly in California, but in incredibly diverse areas of it, demographically, economically, and politically, I've lived abroad in what is technically a third-world country, and I've traveled quite a bit all over the United States and the world.

At every stage I've picked up friends who I still keep in touch with and I've found that this exposes me to a wide variety of views. In just about every major public debate of the last decade, I don't think there's been one where every point of view hasn't been somehow represented by someone I know. This is just as true about the discussion over the last few years around AI as it has been for everything else and, when I watch the discussion happen, I notice that there is one particular area where the many sides of the AI debate aren't exactly understanding each other.

To put it simply, they're all technically right, and the reason why all of those right answers aren't meshing together into a single right answer has a lot to do with perspective. To understand why and what's going on, I'll need to dig into the past a bit, starting a little over a century ago.

THE PAST

In the middle of the 1800s and particularly by the end, the Industrial Revolution caused massive changes in the United States. You see, at that point, the American economy (and many economies around the world) was transitioning from a primarily agrarian economy to a primarily manufacturing economy and, let me tell you, it wasn't going well.

Huge numbers of people who could no longer find work on farms due to the invention of machines that could do the work of 10 or even 100 people flooded into cities that didn't have nearly the infrastructure to support them. Factory owners used the desperation of those people to earn a living to take massive advantage of them. Wages were too low to lift people out of poverty, shifts were long and constant, workplace safety was nearly nonexistent, and children as young as 5 could work 12 hour shifts 6 days a week for pennies in jobs that would kill and maim them by the scores.

It took decades of work and often brutal confrontation to turn manufacturing jobs from poverty death traps into the "good jobs" that people describe today. If you want to know how brutal, just consider that, if the labor struggle had been a war, it would have been the eleventh most deadly war in US history, and that's just the deaths that can be officially documented in active violence. (It also featured the first time bombs were dropped from airplanes on American soil) It took decades of incremental progress, from laws that protected the right to unionization and strike activity to laws that broke up monopolies, oligopolies, and trusts, to increased taxes on the wealthy that funded a social safety net to boycotts and wildcat strikes and even assassinations of both labor activists and capitalists, for the industrial revolution in America to evolve into broad-based prosperity.

Many people lived their whole lives during this struggle and died before it achieved its results in the 1950s and 60s with the average person seeing a vastly improved standard of living and, even then, the gains were not shared equally, with white men taking a disproportionate share of the income and wealth with far less available to women and minorities.

THE MORE RECENT PAST

In the 1970s and particularly in the 1980s, something changed radically. The broad-based prosperity and shared productivity gains that had finally been achieved and that characterized the 1950s and 60s basically stopped happening. In fact, the change is so stark that you can see it on a huge number of economic graphs.

In that period, a lot of things changed. The government, in particular, largely stopped breaking up monopolies and oligopolies, it stopped taxing the wealthy to pay for things that benefited the non-wealthy, and it dramatically reduced the protection it offered to unions. Technology also changed substantially, with automation and particularly computing allowing a centralization of power and wealth that wasn't really possible before. Computing, in particular, changed the balance of power, with those who were able to crunch large amounts of data being able to easily outmaneuver those who couldn't.

Since that change, incomes have basically stagnated with workers making about the same now, in inflation adjusted terms, as they did in 1980 while productivity has more than doubled. In fact, it gets worse than that when you break it out by education and you find that non-college educated workers, the majority of American workers, have actually lost ground in that time (offset by an increase in income for college educated workers).

I titled this "the more recent past", but it's really not that recent anymore. I'm 41 years old this year and this "recent past" is quite literally more than my entire lifetime. If the pendulum were to start to swing back to the benefit of labor starting immediately, given the pace at which this kind of thing has historically happened, I'd probably be retired long before we reached the kind of broad-based prosperity that existed in the immediate post-WW2 period.

And that leads us to the present where we have the highest levels of both income inequality and wealth inequality since the Gilded Era in the 1920s. Changes in technology have massively advantaged capital over labor, allowing huge increases in production with fewer and fewer human workers in the process. Much like the transition of the Progressive Era, we are seeing tons of workers flood from manufacturing to services with similar results. Some of the laws from that era that are still on the books and enforced, like child labor laws, have prevented things from becoming quite as bad as they were back then, but even there we're seeing efforts to weaken them or repeal them entirely.

All of that history as well as the present is what shapes the debate around AI today.

THE AI DEBATE

So will AI destroy the lives of workers and result in massive concentrations of wealth or will it result in a new era of prosperity and plenty for all? Well, why can't the answer be both?

Seriously, under our current system, the gains from broad deployment of AI, should it reach the extreme heights that its most prominent boosters think it could or not, will probably be captured by the extremely wealthy just like every other gain in productivity over the last half century. If it is as disruptive as seems likely, though, this will likely lead to the kind of labor resistance that was observed in the run up to and through the Industrial Revolution and particularly in the Progressive Era, and could lead to those gains being more widely shared than they will be under our current system.

Again, both of these can be true! The thing is, on what time scale and at what cost? Historically, economic change of this type has come slowly, taking entire lifetimes or even longer to achieve significant results, and they have often come at an extremely high cost in lives and livelihoods.

When the anti-AI person says they want it all destroyed because it will simply lead to more concentration of wealth and income and the pro-AI person responds that this doesn't necessarily have to be true, they're having very different conversations. I once pressed a pro-AI person on this particular disconnect and their response was that the level of concentration that AI could allow would guarantee a Progressive Era style shift that would cause the resulting gains to be broadly shared.

And I don't necessarily disagree with the statement, but I do feel like it completely ignores the speed at which these types of shifts tend to happen. The shift could, in fact, occur AND everyone alive today could live their entire lives without seeing the benefits of it. Both of these could be true statements and I don't think the pro-AI side actually has a good answer as to how that shift might be sped up so that people today would actually benefit from the likely or even speculative productivity gains of AI use.

CONCLUSION

When pro-AI people try to rebut the points of anti-AI people, they're technically correct, but they're not paying attention to the amount of time or the cost their positive points would take to implement.

Right now, our society and economy are configured so that wealthy people who own capital will collect the majority of any productivity gains. Historically, this situation can take a lifetime or more to reverse such that gains in productivity are broadly shared as increased prosperity.

At this point the most likely outcome of AI, assuming it reaches anywhere near the potential its proponents believe it can, would be that it would massively increase productivity at the expense of working people because that is what our current economy does with increased productivity. In the long-term, this increasing consolidation of wealth would likely increase the already existing pressure for social change and that pressure would eventually cause the increases in productivity driven by AI to be broadly shared, making everyone better off. However, given the historical timeline of such things, it's possible, even probable, that several generations may be born and die before that occurs.

So far I haven't seen any of the people who promote AI answer the question of how quickly the possible benefits will be widely spread and that's a much bigger omission than I think they give it credit for.