So, I think it's worth noting that at least for Gemini images, it's not a matter of juicing the training data and then sitting back and let the model do its thing. Rather, Gemini rewrites your prompt to make it a woke prompt and serves you the output of that. Zev has most of the gory details (https://thezvi.substack.com/p/gemini-has-a-problem) but essentially it adds keywords and instructions that make your prompt sound like it comes from the DEI office at a university trying to produce a wholesome image of a diverse student body for a promotional ad *for every single prompt*, even if you beg for e.g. historical accuracy.
Google doesn't just edit all the prompts, it didn't tell anybody about the editing, won't tell anyone what the rules are or how the editing algorithm works, won't show users "their own" edited prompt that actually got fed into the engine, and doesn't allow users the option or ability to insist upon getting the output that would be generated without these and other thumbs on the scale.
Free-market Libertarians are going to tend to have a knee-jerk reaction to this and say "Private companies must be allowed to do whatever they want! And if people don't like it, they can choose to use a competitor or build their own Google." But a customer can't choose between services when they can't understand how those services are different because the companies themselves are keeping all that secret. Great, now Kafka's K gets to choose between *several* impenetrable, inaccessible, incomprehensible, inscrutable castles.
Everyone is focusing on the racial aspects to all this, which of course are egregious enough, but when one realizes that such things could be applied to literally every possible subject and especially areas of ideological and political controversy, it becomes all the more fraught with sinister possibilities and alarming.
Also, Reinforcement Learning from Human Feedback (RLHF) was conducted after training. It would have given positive reinforcement for "diverse" people in an image and negative reinforcement for all white males. The RLHF was probably not detailed enough to specify that diverse peoples should be historically reasonable.
Arnold, you offer many good insights but maybe should go a little deeper into how LLMs are developed.
Regarding humanoid robots: Some say that the development of aeronautics was delayed for centuries by the belief that the wings of a flying machine had to flap like a bird's wings. It could be that the development of efficient robots is being held back by our quaint belief that they need to look like us. The real benefits of robotics may come not from generalist humanoid-appearing machines, but through super-specialized robots integrated with systems and networks in novel configurations with no counterparts in natural world.
"Some say that the development of aeronautics was delayed for centuries by the belief that the wings of a flying machine had to flap like a bird's wings."
It's a myth.
The theory of aerodynamics had been worked out to levels of accuracy tolerable to design heavier-than-air powered aircraft in the early 18th century, at least 150 years before Kitty Hawk. People knew all about the airfoil shapes needed for wings and propellers for a while, the bottleneck was the engine.
Powered flight arrived the moment there was an internal combustion engine with a power-to-weight ratio sufficient to keep anything in the air. How do we know? Because that's what the Wrights actually accomplished. The engine was the breakthrough that allowed the Wrights to win the race. There weren't any ICU engines in existence (at least in the US) with that high of a power-to-weight ratio, because weight is not as critical a factor for automobiles, and the available ones were all made out of steel. And so they commissioned Charlie Taylor to help them make the "A engine" largely out of copper and aluminum (which was much more rare and expensive a material at the time), to cut the weight considerably. Practically as soon as the engine was ready, they got the Flyer in the air.
As far as the theory of what flight needed, consider that people have looked up since the dawn of humanity to see birds hover in the air on wind currents, from which it isn't hard to infer the basic principles of the idea of lift derived from airflow.
Additionally, while they began as simple "parachute"-type devices to capture wind pressure for propulsion - like the square-rigged mast sails common in the era of classical antiquity - the technology of sails advanced in most places quite long ago to also being "vertical airfoils" (e.g., lateen-type, look at the shape) which is what permits sailing ships to move forward even in headwinds at extreme angles of attack, and of course these are precisely the principles and tech Northern Europeans used to invent the Dutch-type windmill almost one thousand years ago. Why didn't the Romans invent the windmill? Well, they seem to have independently invented the lateen, which is pretty impressive, but they didn't do that until fairly late in the game, around the time of the First Division, and the Western empire was in decline by the time it became common. But who knows.
Greeks definitely knew the principle of windmills, since Heron casually refers to it in his Pneumatica (https://www.loc.gov/resource/rbc0001.2009gen41532/?sp=132&st=image) describing a wind-driven pipe organ. The Greek word Heron uses is ανεμουριον. Whether big windmills were used for grinding flour is impossible to say given the available evidence. I suspect that if they existed and were like the type of windmill I am familiar with, they would have been made out of wood and thus would not have survived except for millstones, which if not reused elsewhere would in any case not have been anything special and might come from other types of powered mill (animal or water).
Heron also invented the vending machine! Roughly around the time of Nero.
Though his machine was just a series of levers which would react to the weight of a coin and temporary lower a pipe or open a valve to pour out a quantity of water or wine. Still, the idea was solid, just needed some work. Apparently it still does, judging from the ones where I work, which still have trouble with little bags of peanuts. "How come they can have an AI score higher than average on an IQ test, but they still can't ... "
Some of the medieval Northern European-type windmills were built with masonry, or on masonry foundations. It's not clear when they might have started to build that way, because many of the early sites were later underwater and many were abandoned as soon as better tech arrived, but I think the archeologists have found some examples.
I was originally going to say that I would figure if the Greeks or Romans used them we would have at least one surviving masonry structure or one surviving description, especially from the Eastern / Byzantine empire with its greater civilizational continuity. But it turns out it is not easy to build one in masonry and also give it the ability to rotate and adjust to wind direction like a weathervane and be able to withstand storms and gusts, and stationary models of any size which couldn't be easily folded up or broken down by a small crew weren't worthwhile undertaking in many locations. Maybe that combined with higher friction for copying and dissemination of innovations (e.g., the hieropolis sawmill) is why Heron's stuff didn't make it past the "cool toy bro" stage.
The typical wind patterns for some select pieces of geography can give several large advantages to the feasibility of using windmills, especially those of the simplest, stationary design. Ideally you want the winds to be (1) blowing all the time, (2) blowing in the same direction, (3) rarely blowing too hard or with too much obstruction from forests or turbulence from, say, mountainous terrain and (4) rarely with the combination of cold and moisture that causes dangerously heavy amounts of ice to accumulate on structures.
The coastal lowlands at the base of the North Sea with steady westerlies warmed by the Gulf Stream are a pretty good spot by those "favorable winds" measures. But I was surprised to learn that this has likely not been true for the whole duration of the multi-century periods relevant to this subject.
Apparently wind patterns there improved 'suddenly' and really did get close to the ideal for simple, early windmills of that type around the time of the medieval warm period, and that timing and location lines up enough with the timing and location of the emergence of those mills that I suspect it's not a mere coincidence. By the time the warm period ended the windmills had been refined and improved enough to handle the more typical and less ideal wind patterns of that area, and it seems most of the expertise of that industry concentrated in that region and had to keep making them better as things went from good to normal then to bad.
Because then came the Little Ice Age which apparently caused all kinds of disruption and havoc in this regard with an especially rough period starting in the late 16th century, around the start of the 80 years war. The Dutch had to invent the "ice yacht" with runners below the hull because the canals turned into ice, and this quickly made it over to communities on the coast of the Baltic Sea which froze over! At any rate there were famines and the new weather was apparently less ideal for mills (perhaps encouraging the Dutch to focus on maritime endeavors and rival the British for control of the seas) until things warmed back up and stabilized again, by which time the industrial revolution had kicked off.
Speaking of "engine breakthroughs as bottlenecks", the Dutch were already pumping water with steam engines as early as 1781, when Jan Hope ordered an improved Newcomen engine from James Watt himself! I guess it makes sense that anything involving air flows would pit the historical schools of technological determinism and Jared Diamond's climatic-geographical determinism against each other, or at least demand a synthetic approach.
One can categorize old windmills by the direction of the axis of rotation. Heron's organ device and the Northern European windmill have the axis pointing in the direction of the wind. But another type is the "panemone" which can have a bunch of vanes, sails, sheets, or panels arranged in a cylinder around an axis perpendicular to the wind, either vertical or horizontal. There is apparently still some controversy - depending on whether you believe the dating from Chinese academics - over whether the Persians or Chinese got to these first, and then who got them from whom.
However, there is more confidence that however the Persians got them, when they did, they quickly spread around the Islamic world all the way from Egypt to India. Just in time for Crusades where some Northern Europeans might have gotten to view them work up close before returning home to the North Sea lowlands during those favorable winds. So ... ???
Wow, I did not know that about wind patterns. That makes a lot of sense. As for Heron, the surviving fragments of Philo of Byzantium's Pneumatica (III BC) as well as relevant fragments of Vitruvius (I BC) [Latin with German translation is the best I could find online: https://archive.org/details/heronsvonalexandhero/page/458/mode/2up] show that Heron most likely compiled earlier work and found funny applications for it rather than inventing stuff independently: Philo's fragment describes principles of pneumatics and experimental devices obviously related to Heron's designed to demonstrate the principles, rather than applications. Vitruvius does not refer to windmills, although he does describe gears converting rotation around a horizontal axis to one around a vertical axis in the section on water-mills. The Antikythera mechanism is evidence that more sophisticated technology existed in the Greco-Roman world than we know from the meager written sources that have reached us. Perhaps now that the Vesuvius Challenge has finally succeeded in reading the carbonized scrolls of Pompeii, we will learn more about it (among other things).
There's little indicating that Biden made Gemini woke. First, the EO is both unenforceable and has not been enforced. Second, a statement that Gemini was made to comply with the EO does not make sense because it has been in development for much longer than the EO has been active. Third, Google has more than enough resources to bat back this unenforceable EO if they wanted to, to the point to which it can indeed expend far more resources than the federal government could to argue this narrow point.
Chau makes a strong claim by saying the government did it, and then makes a second, more defensible pseudo-claim that the EO "contributed" to the outcome. You could also say that Catherine F. Biden contributed to Gemini's failure by giving birth to Joe, but it is similarly a useless argument.
No, the issue with Gemini is more insidious: the government does not need to actually regulate a company that is so full of people who are effectively government agents because those G-men-and-women will hobble the company for you. Why use the legal process if you can get the company to kill itself voluntarily? That's not really a conventional regulatory action limited by all the many due process requirements inherent in it. The issue with this as a regulatory strategy for the government is that you can only expect voluntary super-compliance with unenforceable rules from some companies, but not all of them. Google may give itself Canadian healthcare in the form of hiring a class of saboteur-employees, but eventually you run out of those people, right?
It's funny you say that about Chau; I don't read him often, but every time I do I come away feeling like he was trying to get me to believe he made some claims and backed them up but didn't actually. It always puts me in mind of reading guru nonsense, where it sounds good at first but when you really dig into it you realize they either haven't said anything or are saying something entirely bland but dressing it up as profound. Even allowing for the context limitations of the medium, and the possibility there are ideas he is taking as a given for his audience, it still reads like the work of someone who knows what his audience wants to be true and is playing to that. Vibes over substance.
"The way I think about it, the chatbots have two data sources. One is true data, that is naturally found “in the wild,” so to speak. The other is fake data, supplied by the model builders."
I'd say it's better to analyze the question in the manner of a "integrity breach" incident investigation for some scenario with multiple independent events in a sequence of transfers, like a "chain of custody", "supply chain", or "chain of trust" in cybersecurity / information assurance. There are lots of potential entry points or "men in the middle" and none are mutually exclusive to the others.
Each link in the chain provides another opportunity to mess with the original stuff, or for bad stuff to get in the package.
For example, let's say someone discovers that some bottles of apple juice contain worrisome levels of some toxic organic compound. There are potentially dozens of places it could have come from. Was it in the soil of the grove? Was it in chemicals sprayed on the fruit? Did it get on the fruit during harvesting, shipping, processing? Did someone tamper with the finished product before it hit the shelves? Any combination or all of the above?
In this case you might call something like the Gemini fiasco an "Information Service Epistemic Integrity Breach Incident" and I'll just call that an "EpInt event".
To analyze an EpInt event post-facto, or to try assess the risk of one happening, there is no other place to start than with some diagram of the chain of transfers / entry-points as potential security vulnerabilities.
The first link in the chain comes before Google, which is that the "data in the wild" is hardly "true", "neutral", "unbiased", etc. in any sense, and is already distorted by being weighted with all kinds of unfair or unrepresentative factors, the kind of people who produced disproportionate quantities of that data, how accurate or biased they tended to be, what happens to still be easily and cheaply available online for scraping at the time of collecting training data, and so forth.
Google could then be putting their thumb on that data, weighting it, promoting some over others. They could be "refining" the reinforced learning system to avoid double-plus-ungood-crimethink. They could even review and not just reject but secretly pre-edit the customer's original prompt before it even gets fed into the engine so that it will follow a different command than the one the customer wanted it to. And who even knows how many other links in the chain there are here Google could be intentionally manipulating the kinds of output to deviate from what the Epistemic-Integrity result would have been.
In general it is a mistake to encourage private companies to offer goods and services in a manner so larded up with dozens of layers of intentional obfuscation and non-traceability that is becomes impossible for the public or even parties to lawsuits with discovery orders or government security officials to effectively investigate what went wrong, who did what when to make it go wrong that way (i.e., accountability) and how can it be prevented from going wrong again. This is precisely the kind of situation which encourages people to call for suffocating government regulation, because what is happening behind the scenes that arouses all this agitation and distrust and suspicion is all a giant secret black box that is intentionally designed to be inaccessible and to mislead users and prevent the public from learning what it really going on.
People are going to think that they have little choice but to ask the government to intervene as the only entity with sufficient power and authority to do anything about these complex products of some of the world's richest companies, but of course while the kind of helpless desperation is legitimate, that regulation will all end up horribly as it always does.
I still maintain that purported agreements to sell services that are intentionally designed to be incomprehensible to customers in terms of knowing enough about them to understand what they are actually agreeing to pay for are not legally valid contracts at all. They are illusory and inherently non-warrantable, and companies can't just "terms of service" or "proprietary trade secrets" their way out of that (though of course they and their toadies will pretend they can). Perhaps they don't have to show -all- their cards, but they should be facing steep trade-offs of dramatically increased liability and increased risk of losing big contract breach suits for every card they decide to keep hidden under their sleeves. The current legal regime is tolerating far more hidden cards than is compatible with the kind of Ministry of Truth infosphere-reshaping power these entities can wield. The Epistemic Integrity of the services most people use to get their information is a Public Good, perhaps in the final analysis the most important Public Good there is.
There is a whole lot of context required to make sense of this sentence. Like, what the heck is a phone menu tree? We know it has something to do with customer service based on the prior paragraph / context.
One leap is to realize "phone" is a reference to a telephone call with a customer service agent. Menu and tree are tougher. We're not at a diner or hiking in the woods. These are computer terms. The customer service agent has a computer screen with a menu-based interface, which has a tree structure.
It's like a script the agent can follow but tries to adapt to the customer situation, like a choose-your-own-adventure novel. These can be frustrating to a customer that knows the product well, recognizes an abnormal problem, and is reaching out for expert help. This type of customer can follow the script / menu / tree themselves without involving a telephone or another human.
The customer service automation breakthrough will be when the systems no longer need to throttle customer requests. What do I mean by throttling? Most services I currently have use a phone tree menu that requires multiple interactions and inputs just to get to a menu where an inquiry can be made.
For example, when I call a certain credit card company I am first given my current account balance and statement date. This is completely unhelpful as I already know this information from the phone app! So why does the phone tree do this? I believe it is to throttle customer engagement. It is to make the customer wait a bit longer before getting access to a human or deeper into the phone tree system.
Customer service chatbots have been around a while. I don't use them. I try not to use any customer service - the process is painful. But sometimes a question needs answered and the only way is to call customer service. Will LLM AI handle customer service better? I hope. Will companies implement LLM AI customer service that is helpful? I hope. But it better work. I loathe nothing more than stupid automation and I fault any company that implements such systems.
Re predictive processing: FWIW in the 1990's at least two scholars anticipated the prescriptive processing model of human cognition.
In 1991 philosopher Daniel C. Dennett's book "Consciousness Explained" described perception and dreaming pretty much exactly as described in Scott Alexanders's review of "Surfing Uncertainty." The top-down system proposes as story to what's going on, which the bottom-up (sensory) system confirms or corrects. When we dream, the top-down system keeps proposing a story but the bottom-up system is asleep, responding with white noise instead of coherent yes-no responses. The story goes off the rails pretty quickly.
In 1995 cognitive scientist Douglas R. Hofstadter of Indiana University also posited in his book "Fluid Concepts, Creative Analogies" a top-down plus bottom-up model. He and his graduate students wrote computer code to emulate this process.
In The Experience Machine: How Our Minds Predict and Shape Reality, Andy Clark's 2023 popularization (and updating) of Surfing Uncertainty, he gives a lot of credit to Daniel Dennett. It's a very good book, well worth reading
“It is unfortunate that large language models have emerged at just the point in history where there is a cohort of people for whom social justice as they define it has become the overwhelming priority.” When was there a time when social justice in one form or another was not an overwhelming priority?
I think "as they define it" is doing a lot of work in that sentence. Although I would also say that 20+ years ago this sort of all encompassing "politics is life" business was much less prevalent. The trend of "I could never date someone of the other political party" is relatively new, and I think tracks pretty well with the totalizing nature of modern social discourse. When I was a kid it was surprising when people of different religious faiths married, now it is different political faiths.
I agree that “‘as they define it’ is doing a lot of work in that sentence.” Actually it’s doing far too much work. I think it’s import especially for someone as high profile as Arnold Kling to connect the dots between the social justice worldview and socialism. I detect a bit of complacency here on his part.
This also relates to his understanding of the relationship between socialism and communism. I don’t think Milton Friedman or F.A. Hayek would say “It is unfortunate that [these computer technologies] have emerged at just the point in history where there is a cohort of people for whom social justice as they define it has become the overwhelming priority.” There has always been an overwhelming priority among this cohort.
I’m tempted to agree with you Doctor Hammer, but 20+ years ago the political views around me were already overwhelming. Even 30+ years ago I was overwhelmed by social justice in the San Francisco Bay Area.
And if I at look at my father’s life, I see social justice as an overwhelmingly source of disruption in his life. My dad never said a word about politics during my childhood in El Dorado County, CA, but he graduated from Cal in 1969; he walked through those protest lines on his way to class. He was slated to fly to Vietnam after graduation, but fortunately enough students had dropped out of the UCLA dental program that he was pulled off the waitlist and admitted. Certainly the Vietnam War was overwhelming.
The Civil Rights Movement during that same era which blurred into the Black Power Movement was overwhelmingly for those involved.
The busing program at my 6th grade school—inspired by social justice—that I wrote about yesterday was overwhelmingly.
My grandma told me about the impact of the Great Depression and World Wars on her family. All overwhelming and closely tied to socialism.
Can we and should we disentangle these wars, depressions, political movements from socialism and its child, the social justice ideology? Socialism and the social justice worldview are really no different when we look at the fundamental similarities.
So when has social justice never been an overwhelming source of conflict?
For that matter, we should be highlighting the fundamental similarities between DEI, the social justice worldview, the Black Power Movement, socialism, progressivism, eugenics, fascism, communism, Jim Crow, and chattel slavery. Let’s not be complacent. When our enemies change names we don’t want to start all over again. Let’s show others the similarities between these worldviews and illustrate how the Constitution should protect us against them.
Or maybe things are so good right now, there’s no need to do this? Maybe we’re taking on the oppressed narrative that the social justice warriors are at their peak right now? Help the LLMs. Oppressed.
If we have a different set of facts however we’ll see that the social justice worldview is just another form of socialism. Agreed?
Ah ok, I think I see where you are coming from. I absolutely agree with you that social justice is the child of, or the same thing as, straight socialism and communism. I also agree that for those who are into that it tends to become an all encompassing focal point and often overwhelming source of conflict for those people.
I will say, though, that when I read you were from CA I thought "Oh, that explains it." In much of the rest of the country, at least the mid-East Coast that Arnold and I would be using as a point of reference, being that all in on those things would be considered weird. What passed for normal in CA would have been pretty extreme here abouts. I think it is only recently when what was normal for the Left Coast (and parts of NYC/DC) started becoming normal here. So I can understand Kling's feeling that it was recently that some people started taking it as doctrine that those who don't agree with their version of leftism are evil. Up until recently (last decade or so at least) it would be more normal to take the position of "Well, I don't agree with their politics, but I still gotta work with them." The opportunities for bubble monocultures to develop and then start imposing themselves on the regular world were much more limited.
I want to agree and give a rousing "Hear Hear!" to your point about all this just being the same fight against authoritarian socialism, albeit in a different uniform this time. I would say the big change in culture we have seen in the past few decades has been the general acceptance of socialism and denial of individual liberty and responsibility across the culture. It has always been there, and probably always will be, but the past 30-40 years have seen it go from a niche philosophy with most people only adopting the positions around fear of others' liberty type issues, to being effectively the default, the presumptive null hypothesis in people's minds to be argued against.
“I would say the big change in culture we have seen in the past few decades has been the general acceptance of socialism and denial of individual liberty and responsibility across the culture.” All the more reason to be rigorous with our definitions of social justice and socialism.
Agreed, although I don't think social justice is exactly 1:1 with socialism anymore, at least not in the way that most people think of socialism. Social justice has spun off a bit lately, including things like "climate justice" and other oddities that are not really part of older school socialism. More of a "these are things we want to be able to control people about" and less of the economic redistribution. Granted, socialism devolves quickly into out and out authoritarianism because that is what its goals require, and the non-economic social justice causes do as well, socialism is usually just the economic cause. In fact, I am realizing as I write this that socialism is possibly a subset of social justice, all of which are linked by the general notion "Things ought to be like X, so lets control people to make X the outcome."
Excellent. Thank you. This is progress. And let’s not forget the perfectly acceptable tradition of voluntary socialism. We can view this as: “Things ought to be like X, so let’s go live on a commune and make it work.” In this way voluntary socialism is an ideal and a practice—like a secular and peaceful religion. So, would it be best to think of social justice as a subset of secular religion? Secular religion being a synonym for a worldview, an ideology or a philosophy.
Religions can vary by degree of coercion. 100% voluntary at one end and 100% coercive at the other. Chattel slavery is an example of a 100% coercive religion. “Sam owns Bob.” Voluntary socialism could be 100% voluntary. “You’re free to leave anytime.”
Likewise, religions can vary by degree of dogma or cultishness. We can arrange Christian denominations by degree of dogma. Likewise by degree of censorship. Even elementary school classrooms are not 100% free speech institutions. Students must abide by rules of respect so that students can learn, and teachers can teach. Foul and disrespectful language are censored. In this sense elementary schools operate with a certain degree of dogma. No one questions that the dogma of the classroom is learning. Students are there to learn.
Likewise with discrimination: “We only allow men into our club,” or “We only allow those who sign our HOA contract and pay $1,000.” “Or we discriminate against low SAT scores.” And how does discrimination fit into religion? Discrimination is an action based on a preference. To act in accordance with color-blindness is to act without preference for race.
Is a preference a subset of belief? Or is it the other way around?
A religion is made up of beliefs, preferences, and practices. Religions can be either supernatural or in accordance with the laws of nature (e.g. physics and biology).
Practices can be vary by degree of habit. At one end of the spectrum, we have to “work hard” to accomplish the practice. At the other end, the practice is “completely automatic.”
DEI is a religion with a set of beliefs, preferences, and practices. We want DEI to be completely voluntary, but it has become part of government.
Where do countries and governments enter into this conversation? America is a government described by the U.S. Constitution.
What is unique about governments? They are marked by a high degree of coercion to enforce general beliefs, preferences and practices, yet they also serve to protect the rights of citizens. The Bill of Rights serves to protect the rights of citizens from the U.S. and State Governments.
So a government is a religion that is marked by a high degree of coercion; run by the people and for the people.
In theory, one type of government is a pure socialist government; that is one without any private property. In practice we know this is impossible to realize.
Then I would finish with your ideas on social justice
“social justice isn’t exactly 1:1 with socialism anymore, at least not in the way that most people think of socialism. Social justice has spun off a bit lately, including things like "climate justice" and other oddities that are not really part of older school socialism. More of a "these are things we want to be able to control people about" and less of the economic redistribution. Granted, socialism devolves quickly into out and out authoritarianism because that is what its goals require, and the non-economic social justice causes do as well, socialism is usually just the economic cause. In fact, I am realizing as I write this that socialism is possibly a subset of social justice, all of which are linked by the general notion "Things ought to be like X, so lets control people to make X the outcome."
Yes, a priority for some but not all. Throughout history movements for social reform or expanded rights were often met with harsh resistance (e.g. the outcome of the Peasant Revolt in England, etc.). So far the SJW are enjoying no such treatment. Nor should they.
You had asked,"When was there a time when social justice in one form or another was not an overwhelming priority?" So, to clarify my reply, my point was that social justice was not always a priority for everyone, and in fact that there was often a brutal response, such as in the case of the Peasants Revolt. Today, at least so far, our current crop of SJW types are lucky they are not meeting that kind of treatment. And I'm now reading your other comments in this thread and gaining a bit more insight to where you're coming from.
Got it. Here’s a question to ponder. What role does political force, and force in general, play in dislodging DEI and other secular religions from state funds?
It's all LARPing. LARPing is everywhere, from American youngsters venting about late capitalism to Russian grunts excitedly filming Ukrainian Magura V5 USVs chasing down and sinking _Sergei Kotov_ in the Kerch bay as heavy machine gun tracers from the ship literally fly past their heads. It's all a pretty picture and story until a bomb actually lands on your roof.
So, I think it's worth noting that at least for Gemini images, it's not a matter of juicing the training data and then sitting back and let the model do its thing. Rather, Gemini rewrites your prompt to make it a woke prompt and serves you the output of that. Zev has most of the gory details (https://thezvi.substack.com/p/gemini-has-a-problem) but essentially it adds keywords and instructions that make your prompt sound like it comes from the DEI office at a university trying to produce a wholesome image of a diverse student body for a promotional ad *for every single prompt*, even if you beg for e.g. historical accuracy.
Google doesn't just edit all the prompts, it didn't tell anybody about the editing, won't tell anyone what the rules are or how the editing algorithm works, won't show users "their own" edited prompt that actually got fed into the engine, and doesn't allow users the option or ability to insist upon getting the output that would be generated without these and other thumbs on the scale.
Free-market Libertarians are going to tend to have a knee-jerk reaction to this and say "Private companies must be allowed to do whatever they want! And if people don't like it, they can choose to use a competitor or build their own Google." But a customer can't choose between services when they can't understand how those services are different because the companies themselves are keeping all that secret. Great, now Kafka's K gets to choose between *several* impenetrable, inaccessible, incomprehensible, inscrutable castles.
Everyone is focusing on the racial aspects to all this, which of course are egregious enough, but when one realizes that such things could be applied to literally every possible subject and especially areas of ideological and political controversy, it becomes all the more fraught with sinister possibilities and alarming.
Also, Reinforcement Learning from Human Feedback (RLHF) was conducted after training. It would have given positive reinforcement for "diverse" people in an image and negative reinforcement for all white males. The RLHF was probably not detailed enough to specify that diverse peoples should be historically reasonable.
Arnold, you offer many good insights but maybe should go a little deeper into how LLMs are developed.
Regarding humanoid robots: Some say that the development of aeronautics was delayed for centuries by the belief that the wings of a flying machine had to flap like a bird's wings. It could be that the development of efficient robots is being held back by our quaint belief that they need to look like us. The real benefits of robotics may come not from generalist humanoid-appearing machines, but through super-specialized robots integrated with systems and networks in novel configurations with no counterparts in natural world.
"Some say that the development of aeronautics was delayed for centuries by the belief that the wings of a flying machine had to flap like a bird's wings."
It's a myth.
The theory of aerodynamics had been worked out to levels of accuracy tolerable to design heavier-than-air powered aircraft in the early 18th century, at least 150 years before Kitty Hawk. People knew all about the airfoil shapes needed for wings and propellers for a while, the bottleneck was the engine.
Powered flight arrived the moment there was an internal combustion engine with a power-to-weight ratio sufficient to keep anything in the air. How do we know? Because that's what the Wrights actually accomplished. The engine was the breakthrough that allowed the Wrights to win the race. There weren't any ICU engines in existence (at least in the US) with that high of a power-to-weight ratio, because weight is not as critical a factor for automobiles, and the available ones were all made out of steel. And so they commissioned Charlie Taylor to help them make the "A engine" largely out of copper and aluminum (which was much more rare and expensive a material at the time), to cut the weight considerably. Practically as soon as the engine was ready, they got the Flyer in the air.
As far as the theory of what flight needed, consider that people have looked up since the dawn of humanity to see birds hover in the air on wind currents, from which it isn't hard to infer the basic principles of the idea of lift derived from airflow.
Additionally, while they began as simple "parachute"-type devices to capture wind pressure for propulsion - like the square-rigged mast sails common in the era of classical antiquity - the technology of sails advanced in most places quite long ago to also being "vertical airfoils" (e.g., lateen-type, look at the shape) which is what permits sailing ships to move forward even in headwinds at extreme angles of attack, and of course these are precisely the principles and tech Northern Europeans used to invent the Dutch-type windmill almost one thousand years ago. Why didn't the Romans invent the windmill? Well, they seem to have independently invented the lateen, which is pretty impressive, but they didn't do that until fairly late in the game, around the time of the First Division, and the Western empire was in decline by the time it became common. But who knows.
Thanks. I learned a lot from your reply.
Greeks definitely knew the principle of windmills, since Heron casually refers to it in his Pneumatica (https://www.loc.gov/resource/rbc0001.2009gen41532/?sp=132&st=image) describing a wind-driven pipe organ. The Greek word Heron uses is ανεμουριον. Whether big windmills were used for grinding flour is impossible to say given the available evidence. I suspect that if they existed and were like the type of windmill I am familiar with, they would have been made out of wood and thus would not have survived except for millstones, which if not reused elsewhere would in any case not have been anything special and might come from other types of powered mill (animal or water).
Heron also invented the vending machine! Roughly around the time of Nero.
Though his machine was just a series of levers which would react to the weight of a coin and temporary lower a pipe or open a valve to pour out a quantity of water or wine. Still, the idea was solid, just needed some work. Apparently it still does, judging from the ones where I work, which still have trouble with little bags of peanuts. "How come they can have an AI score higher than average on an IQ test, but they still can't ... "
Some of the medieval Northern European-type windmills were built with masonry, or on masonry foundations. It's not clear when they might have started to build that way, because many of the early sites were later underwater and many were abandoned as soon as better tech arrived, but I think the archeologists have found some examples.
I was originally going to say that I would figure if the Greeks or Romans used them we would have at least one surviving masonry structure or one surviving description, especially from the Eastern / Byzantine empire with its greater civilizational continuity. But it turns out it is not easy to build one in masonry and also give it the ability to rotate and adjust to wind direction like a weathervane and be able to withstand storms and gusts, and stationary models of any size which couldn't be easily folded up or broken down by a small crew weren't worthwhile undertaking in many locations. Maybe that combined with higher friction for copying and dissemination of innovations (e.g., the hieropolis sawmill) is why Heron's stuff didn't make it past the "cool toy bro" stage.
The typical wind patterns for some select pieces of geography can give several large advantages to the feasibility of using windmills, especially those of the simplest, stationary design. Ideally you want the winds to be (1) blowing all the time, (2) blowing in the same direction, (3) rarely blowing too hard or with too much obstruction from forests or turbulence from, say, mountainous terrain and (4) rarely with the combination of cold and moisture that causes dangerously heavy amounts of ice to accumulate on structures.
The coastal lowlands at the base of the North Sea with steady westerlies warmed by the Gulf Stream are a pretty good spot by those "favorable winds" measures. But I was surprised to learn that this has likely not been true for the whole duration of the multi-century periods relevant to this subject.
Apparently wind patterns there improved 'suddenly' and really did get close to the ideal for simple, early windmills of that type around the time of the medieval warm period, and that timing and location lines up enough with the timing and location of the emergence of those mills that I suspect it's not a mere coincidence. By the time the warm period ended the windmills had been refined and improved enough to handle the more typical and less ideal wind patterns of that area, and it seems most of the expertise of that industry concentrated in that region and had to keep making them better as things went from good to normal then to bad.
Because then came the Little Ice Age which apparently caused all kinds of disruption and havoc in this regard with an especially rough period starting in the late 16th century, around the start of the 80 years war. The Dutch had to invent the "ice yacht" with runners below the hull because the canals turned into ice, and this quickly made it over to communities on the coast of the Baltic Sea which froze over! At any rate there were famines and the new weather was apparently less ideal for mills (perhaps encouraging the Dutch to focus on maritime endeavors and rival the British for control of the seas) until things warmed back up and stabilized again, by which time the industrial revolution had kicked off.
Speaking of "engine breakthroughs as bottlenecks", the Dutch were already pumping water with steam engines as early as 1781, when Jan Hope ordered an improved Newcomen engine from James Watt himself! I guess it makes sense that anything involving air flows would pit the historical schools of technological determinism and Jared Diamond's climatic-geographical determinism against each other, or at least demand a synthetic approach.
One can categorize old windmills by the direction of the axis of rotation. Heron's organ device and the Northern European windmill have the axis pointing in the direction of the wind. But another type is the "panemone" which can have a bunch of vanes, sails, sheets, or panels arranged in a cylinder around an axis perpendicular to the wind, either vertical or horizontal. There is apparently still some controversy - depending on whether you believe the dating from Chinese academics - over whether the Persians or Chinese got to these first, and then who got them from whom.
However, there is more confidence that however the Persians got them, when they did, they quickly spread around the Islamic world all the way from Egypt to India. Just in time for Crusades where some Northern Europeans might have gotten to view them work up close before returning home to the North Sea lowlands during those favorable winds. So ... ???
Wow, I did not know that about wind patterns. That makes a lot of sense. As for Heron, the surviving fragments of Philo of Byzantium's Pneumatica (III BC) as well as relevant fragments of Vitruvius (I BC) [Latin with German translation is the best I could find online: https://archive.org/details/heronsvonalexandhero/page/458/mode/2up] show that Heron most likely compiled earlier work and found funny applications for it rather than inventing stuff independently: Philo's fragment describes principles of pneumatics and experimental devices obviously related to Heron's designed to demonstrate the principles, rather than applications. Vitruvius does not refer to windmills, although he does describe gears converting rotation around a horizontal axis to one around a vertical axis in the section on water-mills. The Antikythera mechanism is evidence that more sophisticated technology existed in the Greco-Roman world than we know from the meager written sources that have reached us. Perhaps now that the Vesuvius Challenge has finally succeeded in reading the carbonized scrolls of Pompeii, we will learn more about it (among other things).
There's little indicating that Biden made Gemini woke. First, the EO is both unenforceable and has not been enforced. Second, a statement that Gemini was made to comply with the EO does not make sense because it has been in development for much longer than the EO has been active. Third, Google has more than enough resources to bat back this unenforceable EO if they wanted to, to the point to which it can indeed expend far more resources than the federal government could to argue this narrow point.
Chau makes a strong claim by saying the government did it, and then makes a second, more defensible pseudo-claim that the EO "contributed" to the outcome. You could also say that Catherine F. Biden contributed to Gemini's failure by giving birth to Joe, but it is similarly a useless argument.
No, the issue with Gemini is more insidious: the government does not need to actually regulate a company that is so full of people who are effectively government agents because those G-men-and-women will hobble the company for you. Why use the legal process if you can get the company to kill itself voluntarily? That's not really a conventional regulatory action limited by all the many due process requirements inherent in it. The issue with this as a regulatory strategy for the government is that you can only expect voluntary super-compliance with unenforceable rules from some companies, but not all of them. Google may give itself Canadian healthcare in the form of hiring a class of saboteur-employees, but eventually you run out of those people, right?
It's funny you say that about Chau; I don't read him often, but every time I do I come away feeling like he was trying to get me to believe he made some claims and backed them up but didn't actually. It always puts me in mind of reading guru nonsense, where it sounds good at first but when you really dig into it you realize they either haven't said anything or are saying something entirely bland but dressing it up as profound. Even allowing for the context limitations of the medium, and the possibility there are ideas he is taking as a given for his audience, it still reads like the work of someone who knows what his audience wants to be true and is playing to that. Vibes over substance.
"The way I think about it, the chatbots have two data sources. One is true data, that is naturally found “in the wild,” so to speak. The other is fake data, supplied by the model builders."
I'd say it's better to analyze the question in the manner of a "integrity breach" incident investigation for some scenario with multiple independent events in a sequence of transfers, like a "chain of custody", "supply chain", or "chain of trust" in cybersecurity / information assurance. There are lots of potential entry points or "men in the middle" and none are mutually exclusive to the others.
Each link in the chain provides another opportunity to mess with the original stuff, or for bad stuff to get in the package.
For example, let's say someone discovers that some bottles of apple juice contain worrisome levels of some toxic organic compound. There are potentially dozens of places it could have come from. Was it in the soil of the grove? Was it in chemicals sprayed on the fruit? Did it get on the fruit during harvesting, shipping, processing? Did someone tamper with the finished product before it hit the shelves? Any combination or all of the above?
In this case you might call something like the Gemini fiasco an "Information Service Epistemic Integrity Breach Incident" and I'll just call that an "EpInt event".
To analyze an EpInt event post-facto, or to try assess the risk of one happening, there is no other place to start than with some diagram of the chain of transfers / entry-points as potential security vulnerabilities.
The first link in the chain comes before Google, which is that the "data in the wild" is hardly "true", "neutral", "unbiased", etc. in any sense, and is already distorted by being weighted with all kinds of unfair or unrepresentative factors, the kind of people who produced disproportionate quantities of that data, how accurate or biased they tended to be, what happens to still be easily and cheaply available online for scraping at the time of collecting training data, and so forth.
Google could then be putting their thumb on that data, weighting it, promoting some over others. They could be "refining" the reinforced learning system to avoid double-plus-ungood-crimethink. They could even review and not just reject but secretly pre-edit the customer's original prompt before it even gets fed into the engine so that it will follow a different command than the one the customer wanted it to. And who even knows how many other links in the chain there are here Google could be intentionally manipulating the kinds of output to deviate from what the Epistemic-Integrity result would have been.
In general it is a mistake to encourage private companies to offer goods and services in a manner so larded up with dozens of layers of intentional obfuscation and non-traceability that is becomes impossible for the public or even parties to lawsuits with discovery orders or government security officials to effectively investigate what went wrong, who did what when to make it go wrong that way (i.e., accountability) and how can it be prevented from going wrong again. This is precisely the kind of situation which encourages people to call for suffocating government regulation, because what is happening behind the scenes that arouses all this agitation and distrust and suspicion is all a giant secret black box that is intentionally designed to be inaccessible and to mislead users and prevent the public from learning what it really going on.
People are going to think that they have little choice but to ask the government to intervene as the only entity with sufficient power and authority to do anything about these complex products of some of the world's richest companies, but of course while the kind of helpless desperation is legitimate, that regulation will all end up horribly as it always does.
I still maintain that purported agreements to sell services that are intentionally designed to be incomprehensible to customers in terms of knowing enough about them to understand what they are actually agreeing to pay for are not legally valid contracts at all. They are illusory and inherently non-warrantable, and companies can't just "terms of service" or "proprietary trade secrets" their way out of that (though of course they and their toadies will pretend they can). Perhaps they don't have to show -all- their cards, but they should be facing steep trade-offs of dramatically increased liability and increased risk of losing big contract breach suits for every card they decide to keep hidden under their sleeves. The current legal regime is tolerating far more hidden cards than is compatible with the kind of Ministry of Truth infosphere-reshaping power these entities can wield. The Epistemic Integrity of the services most people use to get their information is a Public Good, perhaps in the final analysis the most important Public Good there is.
Orwell foresaw all of this- the past is being rewritten.
> Die, phone menu trees, die!
There is a whole lot of context required to make sense of this sentence. Like, what the heck is a phone menu tree? We know it has something to do with customer service based on the prior paragraph / context.
One leap is to realize "phone" is a reference to a telephone call with a customer service agent. Menu and tree are tougher. We're not at a diner or hiking in the woods. These are computer terms. The customer service agent has a computer screen with a menu-based interface, which has a tree structure.
It's like a script the agent can follow but tries to adapt to the customer situation, like a choose-your-own-adventure novel. These can be frustrating to a customer that knows the product well, recognizes an abnormal problem, and is reaching out for expert help. This type of customer can follow the script / menu / tree themselves without involving a telephone or another human.
Or wait, is this a reference to: "press 1 for banking, 2 for tech support ..."?
The customer service automation breakthrough will be when the systems no longer need to throttle customer requests. What do I mean by throttling? Most services I currently have use a phone tree menu that requires multiple interactions and inputs just to get to a menu where an inquiry can be made.
For example, when I call a certain credit card company I am first given my current account balance and statement date. This is completely unhelpful as I already know this information from the phone app! So why does the phone tree do this? I believe it is to throttle customer engagement. It is to make the customer wait a bit longer before getting access to a human or deeper into the phone tree system.
Customer service chatbots have been around a while. I don't use them. I try not to use any customer service - the process is painful. But sometimes a question needs answered and the only way is to call customer service. Will LLM AI handle customer service better? I hope. Will companies implement LLM AI customer service that is helpful? I hope. But it better work. I loathe nothing more than stupid automation and I fault any company that implements such systems.
Re predictive processing: FWIW in the 1990's at least two scholars anticipated the prescriptive processing model of human cognition.
In 1991 philosopher Daniel C. Dennett's book "Consciousness Explained" described perception and dreaming pretty much exactly as described in Scott Alexanders's review of "Surfing Uncertainty." The top-down system proposes as story to what's going on, which the bottom-up (sensory) system confirms or corrects. When we dream, the top-down system keeps proposing a story but the bottom-up system is asleep, responding with white noise instead of coherent yes-no responses. The story goes off the rails pretty quickly.
In 1995 cognitive scientist Douglas R. Hofstadter of Indiana University also posited in his book "Fluid Concepts, Creative Analogies" a top-down plus bottom-up model. He and his graduate students wrote computer code to emulate this process.
In The Experience Machine: How Our Minds Predict and Shape Reality, Andy Clark's 2023 popularization (and updating) of Surfing Uncertainty, he gives a lot of credit to Daniel Dennett. It's a very good book, well worth reading
“It is unfortunate that large language models have emerged at just the point in history where there is a cohort of people for whom social justice as they define it has become the overwhelming priority.” When was there a time when social justice in one form or another was not an overwhelming priority?
I think "as they define it" is doing a lot of work in that sentence. Although I would also say that 20+ years ago this sort of all encompassing "politics is life" business was much less prevalent. The trend of "I could never date someone of the other political party" is relatively new, and I think tracks pretty well with the totalizing nature of modern social discourse. When I was a kid it was surprising when people of different religious faiths married, now it is different political faiths.
I agree that “‘as they define it’ is doing a lot of work in that sentence.” Actually it’s doing far too much work. I think it’s import especially for someone as high profile as Arnold Kling to connect the dots between the social justice worldview and socialism. I detect a bit of complacency here on his part.
This also relates to his understanding of the relationship between socialism and communism. I don’t think Milton Friedman or F.A. Hayek would say “It is unfortunate that [these computer technologies] have emerged at just the point in history where there is a cohort of people for whom social justice as they define it has become the overwhelming priority.” There has always been an overwhelming priority among this cohort.
I’m tempted to agree with you Doctor Hammer, but 20+ years ago the political views around me were already overwhelming. Even 30+ years ago I was overwhelmed by social justice in the San Francisco Bay Area.
And if I at look at my father’s life, I see social justice as an overwhelmingly source of disruption in his life. My dad never said a word about politics during my childhood in El Dorado County, CA, but he graduated from Cal in 1969; he walked through those protest lines on his way to class. He was slated to fly to Vietnam after graduation, but fortunately enough students had dropped out of the UCLA dental program that he was pulled off the waitlist and admitted. Certainly the Vietnam War was overwhelming.
The Civil Rights Movement during that same era which blurred into the Black Power Movement was overwhelmingly for those involved.
The busing program at my 6th grade school—inspired by social justice—that I wrote about yesterday was overwhelmingly.
https://open.substack.com/pub/scottgibb/p/busing-programs-and-luxury-beliefs?r=nb3bl&utm_medium=ios
My grandma told me about the impact of the Great Depression and World Wars on her family. All overwhelming and closely tied to socialism.
Can we and should we disentangle these wars, depressions, political movements from socialism and its child, the social justice ideology? Socialism and the social justice worldview are really no different when we look at the fundamental similarities.
So when has social justice never been an overwhelming source of conflict?
For that matter, we should be highlighting the fundamental similarities between DEI, the social justice worldview, the Black Power Movement, socialism, progressivism, eugenics, fascism, communism, Jim Crow, and chattel slavery. Let’s not be complacent. When our enemies change names we don’t want to start all over again. Let’s show others the similarities between these worldviews and illustrate how the Constitution should protect us against them.
Or maybe things are so good right now, there’s no need to do this? Maybe we’re taking on the oppressed narrative that the social justice warriors are at their peak right now? Help the LLMs. Oppressed.
If we have a different set of facts however we’ll see that the social justice worldview is just another form of socialism. Agreed?
Ah ok, I think I see where you are coming from. I absolutely agree with you that social justice is the child of, or the same thing as, straight socialism and communism. I also agree that for those who are into that it tends to become an all encompassing focal point and often overwhelming source of conflict for those people.
I will say, though, that when I read you were from CA I thought "Oh, that explains it." In much of the rest of the country, at least the mid-East Coast that Arnold and I would be using as a point of reference, being that all in on those things would be considered weird. What passed for normal in CA would have been pretty extreme here abouts. I think it is only recently when what was normal for the Left Coast (and parts of NYC/DC) started becoming normal here. So I can understand Kling's feeling that it was recently that some people started taking it as doctrine that those who don't agree with their version of leftism are evil. Up until recently (last decade or so at least) it would be more normal to take the position of "Well, I don't agree with their politics, but I still gotta work with them." The opportunities for bubble monocultures to develop and then start imposing themselves on the regular world were much more limited.
I want to agree and give a rousing "Hear Hear!" to your point about all this just being the same fight against authoritarian socialism, albeit in a different uniform this time. I would say the big change in culture we have seen in the past few decades has been the general acceptance of socialism and denial of individual liberty and responsibility across the culture. It has always been there, and probably always will be, but the past 30-40 years have seen it go from a niche philosophy with most people only adopting the positions around fear of others' liberty type issues, to being effectively the default, the presumptive null hypothesis in people's minds to be argued against.
“I would say the big change in culture we have seen in the past few decades has been the general acceptance of socialism and denial of individual liberty and responsibility across the culture.” All the more reason to be rigorous with our definitions of social justice and socialism.
Agreed, although I don't think social justice is exactly 1:1 with socialism anymore, at least not in the way that most people think of socialism. Social justice has spun off a bit lately, including things like "climate justice" and other oddities that are not really part of older school socialism. More of a "these are things we want to be able to control people about" and less of the economic redistribution. Granted, socialism devolves quickly into out and out authoritarianism because that is what its goals require, and the non-economic social justice causes do as well, socialism is usually just the economic cause. In fact, I am realizing as I write this that socialism is possibly a subset of social justice, all of which are linked by the general notion "Things ought to be like X, so lets control people to make X the outcome."
Excellent. Thank you. This is progress. And let’s not forget the perfectly acceptable tradition of voluntary socialism. We can view this as: “Things ought to be like X, so let’s go live on a commune and make it work.” In this way voluntary socialism is an ideal and a practice—like a secular and peaceful religion. So, would it be best to think of social justice as a subset of secular religion? Secular religion being a synonym for a worldview, an ideology or a philosophy.
Religions can vary by degree of coercion. 100% voluntary at one end and 100% coercive at the other. Chattel slavery is an example of a 100% coercive religion. “Sam owns Bob.” Voluntary socialism could be 100% voluntary. “You’re free to leave anytime.”
Likewise, religions can vary by degree of dogma or cultishness. We can arrange Christian denominations by degree of dogma. Likewise by degree of censorship. Even elementary school classrooms are not 100% free speech institutions. Students must abide by rules of respect so that students can learn, and teachers can teach. Foul and disrespectful language are censored. In this sense elementary schools operate with a certain degree of dogma. No one questions that the dogma of the classroom is learning. Students are there to learn.
Likewise with discrimination: “We only allow men into our club,” or “We only allow those who sign our HOA contract and pay $1,000.” “Or we discriminate against low SAT scores.” And how does discrimination fit into religion? Discrimination is an action based on a preference. To act in accordance with color-blindness is to act without preference for race.
Is a preference a subset of belief? Or is it the other way around?
A religion is made up of beliefs, preferences, and practices. Religions can be either supernatural or in accordance with the laws of nature (e.g. physics and biology).
Practices can be vary by degree of habit. At one end of the spectrum, we have to “work hard” to accomplish the practice. At the other end, the practice is “completely automatic.”
DEI is a religion with a set of beliefs, preferences, and practices. We want DEI to be completely voluntary, but it has become part of government.
Where do countries and governments enter into this conversation? America is a government described by the U.S. Constitution.
What is unique about governments? They are marked by a high degree of coercion to enforce general beliefs, preferences and practices, yet they also serve to protect the rights of citizens. The Bill of Rights serves to protect the rights of citizens from the U.S. and State Governments.
So a government is a religion that is marked by a high degree of coercion; run by the people and for the people.
In theory, one type of government is a pure socialist government; that is one without any private property. In practice we know this is impossible to realize.
Then I would finish with your ideas on social justice
“social justice isn’t exactly 1:1 with socialism anymore, at least not in the way that most people think of socialism. Social justice has spun off a bit lately, including things like "climate justice" and other oddities that are not really part of older school socialism. More of a "these are things we want to be able to control people about" and less of the economic redistribution. Granted, socialism devolves quickly into out and out authoritarianism because that is what its goals require, and the non-economic social justice causes do as well, socialism is usually just the economic cause. In fact, I am realizing as I write this that socialism is possibly a subset of social justice, all of which are linked by the general notion "Things ought to be like X, so lets control people to make X the outcome."
Yes, a priority for some but not all. Throughout history movements for social reform or expanded rights were often met with harsh resistance (e.g. the outcome of the Peasant Revolt in England, etc.). So far the SJW are enjoying no such treatment. Nor should they.
Nor should they? Just a social reform movement?
You had asked,"When was there a time when social justice in one form or another was not an overwhelming priority?" So, to clarify my reply, my point was that social justice was not always a priority for everyone, and in fact that there was often a brutal response, such as in the case of the Peasants Revolt. Today, at least so far, our current crop of SJW types are lucky they are not meeting that kind of treatment. And I'm now reading your other comments in this thread and gaining a bit more insight to where you're coming from.
Got it. Here’s a question to ponder. What role does political force, and force in general, play in dislodging DEI and other secular religions from state funds?
When I see the term "late capitalism", I react the same way I do when I see "stolen election'. I find it very hard to take the writer seriously.
This is quite interesting: https://www.compactmag.com/article/too-late-for-late-capitalism/
Thanks for the pointer to the review in compact magazine. Interesting.
I didn't really know the history of the term, so the first half of that was worth reading.
It seems to be used relexively among the youngsters. They do not choose to act on it in the manner of a flower child, however.
It's all LARPing. LARPing is everywhere, from American youngsters venting about late capitalism to Russian grunts excitedly filming Ukrainian Magura V5 USVs chasing down and sinking _Sergei Kotov_ in the Kerch bay as heavy machine gun tracers from the ship literally fly past their heads. It's all a pretty picture and story until a bomb actually lands on your roof.