I like this article. I do think we need to study and prepare for existential level (low likelihood, high consequence) risk better than we do. One added consideration: The types of people who innovate and invent (say, entrepreneurial types) are fundamentally different in character than regulation types, there is very little overlap. So innovators are always ahead of the curve. and the regulation types always play catch up. It was ever thus.
"We only learn from blood" , was my attempt at a dramatic turn of phrase, in the lecture on the Titanic that got me "internet famous" for a few days in 1997, when the NYT was giving people good links to "Titanic" material on the new Internet thingy with the movie out.
It got me flown cross-continent to, as I told my wife "Lecture the US Navy on marine architecture", at least in that a little science conference sponsored by the Navy picked me as the general-meeting presentation for all.
My point came from how many disasters had already shown the Titanic disaster to be quite possible, and indeed, how earlier ships had been safer, safety factors chopped out one by one to save money, for sixty years...until a sufficiently-LARGE disaster shook everybody up, and new regulations came overnight. But it took a lot of blood to make people notice. The "Atlantic" disaster, 500 dead, just not big enough!
All still collected at my old Unix-user-group web site, 25 years later:
the original short essay(#1) is the quick read, but #3, the longer lecture with graphics, talks the most about the "wait until people die" topic that Dan raises today.
My amendment, is that ENOUGH people have to die. ( On gun regs, the US is still getting there.)
Dan touches on that, though, with the same topic I highlight in the longer lecture: the newspapers went wild. The "Titanic" stayed on the front pages with agony stories about individual survival for weeks. THEN there was a big commission of inquiry. And it really counted that the Titanic killed *celebrities*.
There is no rational reason to regulate AI. People are terrible at predicting outcomes of new tech, regulators even more so. Because life has become so easy in the West, we have adopted the mentality that any danger, any risk is unacceptable and must be minimized through the use of government force. Yet every major advance since the industrial revolution happened in unregulated or pre-regulated industries. In general, regulation means stagnation.
Risk is part of life. The proper response to the tombstone mentality is to reject the zero-risk mentality and panicked, monomaniacal focus on single risks (e.g., COVID) that end up causing more destruction than the problems they are trying to fix.
I have a lot of sympathy for the sentiment, if not the conclusion, of your comment. We are indeed bad at predicting tech outcomes, over-regulation can indeed produce stagnation, and hypersensitivity to risk can lead to very bad decisions (as I explore in Risk). These are very real countervailing concerns.
But I question your sweeping conclusions. The future harms and benefits of new technologies are not completely unknowable. Sometimes they are quite predictable. See, for example, https://dgardner.substack.com/p/the-techno-libertarian-faith . And I'm not sure on what basis you can conclude that "in general, regulation means stagnation." Really? No innovation in wireless services? No innovation in pharmaceuticals? Aviation, cars, radio and television: If I look at where these became regulated, I don't see innovation ending. Not remotely. And there are so many other examples I can think of top of mind. So what did you have in mind? How would you support your conclusion?
Progress and stagnation are relative terms. My thesis stated more precisely is that the extent to which an industry is regulated is the extent to which progress in that industry is slowed or stifled. And within an industry, the less regulated areas are those where we see the most progress.
To take one of your examples, great progress has indeed been made in the pharma industry over the past decades. But it still takes 10+ years and $1+ billion to bring a new drug to market due in large part to regulation. And some of the greatest advances have occurred in anti-cancer drugs because when the choice is certain death versus possible help from a new treatment, regulators are more willing to allow that new treatment.
To take another example, wireless services are indeed regulated by the FCC, though much such "regulation" amounts to enforcement of property rights over frequency bands. But the core software and hardware technologies behind smartphones and apps are mostly not regulated -- and that's where the great progress has occurred, resulting in us all having amazingly powerful computers in our pockets.
I can't say whether the stagnation of the auto industry in America in the 1970s-1980s and loss of market share to Japan was caused more by regulation as such or other government interference. But no one would deny that it happened, and IMO is continuing even now within the major manufacturers. It's only in the upstarts such as Tesla using new tech and non-unionized workforces (thus avoiding union-related regulations) that we see much progress today.
But notably, you're not including "harms not done" in your definition of progress. Take aviation, which is severely regulated. Over the past 60 years, aviation safety has improved spectacularly, spurred on largely by the extremely high demands of regulation. Surely that counts for something?
Again, I don't disagree with your point. But I think there are tradeoffs in almost every case and it's wrong to ignore either gains or losses. To go back to the pharmaceuticals example: You're surely right that regulation has slowed innovation, but have you read the history of medicine before regulation? It's terrifying. We may not be innovating as rapidly as we could be had we never regulated, but we're also sickening and killing people at a much lower rate. And I think we should call that progress, too.
I think you are wrong on vaccines where better regulated countries with more trust in government had better covid outcomes. It is telling that you don’t even entertain that might be incompetence of businessmen that led to US automotive industry decline.
You also miss two really obvious counter examples.
Imagine how well driving works in a country with no traffic regulations.
You can also compare outcomes for countries with tighter gun control regulations and see whether the is any case for a free market approach working well there.
If you view traffic laws and restrictions on private gun ownership as "regulations" then you are using the term in a different way than I am. My argument refers specifically to economic regulations in the form of restrictions on business conduct and innovation. Not every law or government restriction on freedom counts as a regulation in this sense.
This is a really interesting point I hadn't thought of: What do people hear when the word "regulation" is used? It's potentially an extremely broad term, which could allow for quite a bit of confusion in discussions like this.
Commissions are political theatre designed to obscure and vindicate the incompetence and malice of the government. Consider the so-called investigation of the use of the Emergency Measures Act. Anyone who paid attention recognized that the facts didn’t fit the conclusion. Spare us more theatre.
Or, perhaps there’s truth to this: there is none so blind as he who will not see. Preston Manning stated reality: “No government initiated a commission without knowing the conclusion ahead of time.” Theatre it is!
Well, with respect to the good Mr. Manning, his point is real but exaggerated. And more importantly, governments that create commissions to study things expecting certain results are sometimes surprised by what the commission reports.
A favourite example of mine: In the early 1970s, Richard Nixon convened a panel of rock-ribbed Republicans to study illicit drug policy fully expecting them to endorse the war on drugs and call for more. Instead, after carefully study, they changed their minds and called for legalization. It was a huge embarrassment for Nixon.
I like this article. I do think we need to study and prepare for existential level (low likelihood, high consequence) risk better than we do. One added consideration: The types of people who innovate and invent (say, entrepreneurial types) are fundamentally different in character than regulation types, there is very little overlap. So innovators are always ahead of the curve. and the regulation types always play catch up. It was ever thus.
"We only learn from blood" , was my attempt at a dramatic turn of phrase, in the lecture on the Titanic that got me "internet famous" for a few days in 1997, when the NYT was giving people good links to "Titanic" material on the new Internet thingy with the movie out.
It got me flown cross-continent to, as I told my wife "Lecture the US Navy on marine architecture", at least in that a little science conference sponsored by the Navy picked me as the general-meeting presentation for all.
My point came from how many disasters had already shown the Titanic disaster to be quite possible, and indeed, how earlier ships had been safer, safety factors chopped out one by one to save money, for sixty years...until a sufficiently-LARGE disaster shook everybody up, and new regulations came overnight. But it took a lot of blood to make people notice. The "Atlantic" disaster, 500 dead, just not big enough!
All still collected at my old Unix-user-group web site, 25 years later:
https://www.cuug.ab.ca/branderr/titanic/
the original short essay(#1) is the quick read, but #3, the longer lecture with graphics, talks the most about the "wait until people die" topic that Dan raises today.
My amendment, is that ENOUGH people have to die. ( On gun regs, the US is still getting there.)
Dan touches on that, though, with the same topic I highlight in the longer lecture: the newspapers went wild. The "Titanic" stayed on the front pages with agony stories about individual survival for weeks. THEN there was a big commission of inquiry. And it really counted that the Titanic killed *celebrities*.
There is no rational reason to regulate AI. People are terrible at predicting outcomes of new tech, regulators even more so. Because life has become so easy in the West, we have adopted the mentality that any danger, any risk is unacceptable and must be minimized through the use of government force. Yet every major advance since the industrial revolution happened in unregulated or pre-regulated industries. In general, regulation means stagnation.
Risk is part of life. The proper response to the tombstone mentality is to reject the zero-risk mentality and panicked, monomaniacal focus on single risks (e.g., COVID) that end up causing more destruction than the problems they are trying to fix.
I have a lot of sympathy for the sentiment, if not the conclusion, of your comment. We are indeed bad at predicting tech outcomes, over-regulation can indeed produce stagnation, and hypersensitivity to risk can lead to very bad decisions (as I explore in Risk). These are very real countervailing concerns.
But I question your sweeping conclusions. The future harms and benefits of new technologies are not completely unknowable. Sometimes they are quite predictable. See, for example, https://dgardner.substack.com/p/the-techno-libertarian-faith . And I'm not sure on what basis you can conclude that "in general, regulation means stagnation." Really? No innovation in wireless services? No innovation in pharmaceuticals? Aviation, cars, radio and television: If I look at where these became regulated, I don't see innovation ending. Not remotely. And there are so many other examples I can think of top of mind. So what did you have in mind? How would you support your conclusion?
Progress and stagnation are relative terms. My thesis stated more precisely is that the extent to which an industry is regulated is the extent to which progress in that industry is slowed or stifled. And within an industry, the less regulated areas are those where we see the most progress.
To take one of your examples, great progress has indeed been made in the pharma industry over the past decades. But it still takes 10+ years and $1+ billion to bring a new drug to market due in large part to regulation. And some of the greatest advances have occurred in anti-cancer drugs because when the choice is certain death versus possible help from a new treatment, regulators are more willing to allow that new treatment.
To take another example, wireless services are indeed regulated by the FCC, though much such "regulation" amounts to enforcement of property rights over frequency bands. But the core software and hardware technologies behind smartphones and apps are mostly not regulated -- and that's where the great progress has occurred, resulting in us all having amazingly powerful computers in our pockets.
I can't say whether the stagnation of the auto industry in America in the 1970s-1980s and loss of market share to Japan was caused more by regulation as such or other government interference. But no one would deny that it happened, and IMO is continuing even now within the major manufacturers. It's only in the upstarts such as Tesla using new tech and non-unionized workforces (thus avoiding union-related regulations) that we see much progress today.
But notably, you're not including "harms not done" in your definition of progress. Take aviation, which is severely regulated. Over the past 60 years, aviation safety has improved spectacularly, spurred on largely by the extremely high demands of regulation. Surely that counts for something?
Again, I don't disagree with your point. But I think there are tradeoffs in almost every case and it's wrong to ignore either gains or losses. To go back to the pharmaceuticals example: You're surely right that regulation has slowed innovation, but have you read the history of medicine before regulation? It's terrifying. We may not be innovating as rapidly as we could be had we never regulated, but we're also sickening and killing people at a much lower rate. And I think we should call that progress, too.
I think you are wrong on vaccines where better regulated countries with more trust in government had better covid outcomes. It is telling that you don’t even entertain that might be incompetence of businessmen that led to US automotive industry decline.
You also miss two really obvious counter examples.
Imagine how well driving works in a country with no traffic regulations.
You can also compare outcomes for countries with tighter gun control regulations and see whether the is any case for a free market approach working well there.
If you view traffic laws and restrictions on private gun ownership as "regulations" then you are using the term in a different way than I am. My argument refers specifically to economic regulations in the form of restrictions on business conduct and innovation. Not every law or government restriction on freedom counts as a regulation in this sense.
This is a really interesting point I hadn't thought of: What do people hear when the word "regulation" is used? It's potentially an extremely broad term, which could allow for quite a bit of confusion in discussions like this.
Another commentary on regulation https://t.co/PdDFT5fxQM?ssr=true
Commissions are political theatre designed to obscure and vindicate the incompetence and malice of the government. Consider the so-called investigation of the use of the Emergency Measures Act. Anyone who paid attention recognized that the facts didn’t fit the conclusion. Spare us more theatre.
I know several people including myself who paid attention and didn’t recognize that so you are quite wrong here.
Or, perhaps there’s truth to this: there is none so blind as he who will not see. Preston Manning stated reality: “No government initiated a commission without knowing the conclusion ahead of time.” Theatre it is!
Well, with respect to the good Mr. Manning, his point is real but exaggerated. And more importantly, governments that create commissions to study things expecting certain results are sometimes surprised by what the commission reports.
A favourite example of mine: In the early 1970s, Richard Nixon convened a panel of rock-ribbed Republicans to study illicit drug policy fully expecting them to endorse the war on drugs and call for more. Instead, after carefully study, they changed their minds and called for legalization. It was a huge embarrassment for Nixon.