The double-edged nature of most technologies seems so obvious that it is hard to believe someone would dispute the need for regulation. When primitive humans first harnessed fire, they quickly learned that, yes, it could heat their home, but without some regulation, it could also burn their home up.
I am virtually 100% sure that AI is capable of terrible things. Social media's machine-learning algorithms were designed to maximize revenue by maximizing engagement. That this turned out to mean maximizing enragement with ancillary radicalization and polarization was an unintended by-product. I'm sure it wouldn't be too difficult to tweak the algorithms to intentionally maximize enragement and polarization, or incline attitudes more favorably toward Trump (as opposed to Biden), or toward China, or against regulation, etc. Our human brains are much more programmable than we like to think, and AIs will rapidly get better at it. If we wait for unmistakable evidence first, we may be screwed ("Oops, looks like AI enabled the destruction of our democracy...").
BTW, the original internet and its protocols were funded by ARPA, which was an arm of the DoD. And I'm pretty sure DoD saw some important defense applications of a command and control network with decentralized routing and control algorithms that gave it high survivability.
Oh, and the internet has certainly been a huge force for good. Being able to communicate and send large amounts of data almost instantly all over the world has enabled all kinds of life changing applications. Don't confuse the internet itself with some of the stupid or useless applications running on it. That would be like questioning the telephone because some guys robocalling company is annoying.
I realize that most internet-hours-spent are probably social, but it's weird to me that someone would see its main value as being a solution to a social problem. The internet archives, distributes, and makes available lots of kinds of information, in general. It has upended multiple industries, vastly increased our ability to know/remember/understand things, made it harder to lie and not be called out, made it possible to solve much harder coordination problems (while creating new ones, of course), and shortened development timelines of lots of other technologies and products.
I think it's easy to forget how much harder it was to know things even 20 years ago, how much more we relied on what a handful of nearby friends and colleagues happened to (mis)remember. It's jarring when I still see echoes of that from, say, people who post requests for medical advice on Facebook and believe whatever responses they happen to get.
If you think the Nuclear Regulatory Commission was "over-burdensome," you should read "Confessions of a Rogue Nuclear Regulator" by Gregory Jazcko, former chairman of the NRC, explaining how the Commission was largely dominated by industry-friendly members, industry-friendly congressmen, and lobbyists.
Zvi Mowshiwitz called this "the dial of progress fallacy" wet Tyler Cowen and AI. The idea that there is a single knob called 'progress' and you either root for it to go up or down, and that in aggregate this is what matters. Perpetrators of this fallacy generally hold that while reality is more complicated and you in general want to take things on a case by case basis, the public will get the most benefit if we shout "Push much harder progress (despite ignoring the nuance)!"
You overlook (1) the benefits of cheap fentanyl and (2) technology’s role in addressing fentanyl abuse. For instance, if you’re blown apart or incinerated, yet are also still alive, you will be very glad for fentanyl abundant. Perhaps the abuse epidemic outweighs the therapeutic impact. But the techno-optimist reply is that, to solve abuse, we need better enforcement and treatment. Those solutions have normatively political components, like more funding for police. Yet also, technology can improve interdiction, if we let it. Technology will also increase economic growth, thereby allowing easier trade-offs to adequately fund enforcement.
"You can’t go down to Walmart and buy a tactical nuclear weapon. They don’t sell ICBMs over the internet. I don’t totally understand the pragmatic concessions conservatives made in deciding there are limits to the second amendment, but I’m glad they made them."
The Second Amendment mentions "arms" not "ordnance." In other words, firearms not artillery. These aren't "pragmatic concessions."
At any rate SCOTUS misinterprets this. It is permission for states to have militias, not individuals to own firearms to protect themselves against the "gommit"
Andreesen’s podcasts appearances (I’m thinking of Sam Harris specifically) after he dropped this article were really embarrassing. I’m told this guy is a visionary but to me sounds like an even more arrogant Elon Musk without the credentials.
I think what Andreesen is going for is an "on net" position. Clearly some tech has some downsides, but he's arguing that the upsides of accelerationist policy are so much greater than the downsides that we would be better off with a damn the torpedoes attitude... on net.
I don't personally agree with this position, but I think the idea that Andreesen is unaware of some of the downsides of technology is itself something of a straw man.
While I generally agree, I wonder how you address the fairly significant number of people who do not endorse technological progress, i.e., Amish/Mennonites, Haridi Judaism, etc. I see these as people who consciously decide to reject in part the modern world. To some extent they're free-riders, but is that too simplistic?
What seems true is that on average technological advancement makes us better off and that we are so bad at predicting what advances will be harmful it's better to just adopt a policy of encouraging the development.
Yes, you are totally correct that the internet has had some really negative consequences...but, ironically, the negatives mostly stem from the things that everyone assumed would make us better off while the things that scared people (job loss, competition from other countries, harms from porn etc etc) mostly didn't materialize as harms. The harms were the result of empowering the public to participate in politics and connecting them with information and giving them the ability to share it.
If you'd asked someone back then to regulate the dangers they'd have mostly regulated the parts that have made lives better and worked to encourage the parts that have made things worse.
I think it's important to distinguish a number of different claims:
1) Its never the case that technical Innovation discovers something that makes us on net worse.
2) It's never the case that we can (as a polity/government) effectively predict what advances will make us worse off and on average such advances make us better off.
3) Its not the case that we can both predict what advances in tech will make us worse off and prevent others from making those advances and we are better off making those advances and regulating applications than trying to prevent them from being made.
I think most techno-optimists believe something much closer to 2 or 3.
Always appreciate a good "Reversed Stupidity Is Not Intelligence" post. Andreesen's position, while neither unexpected nor novel, updates me in favour of Dial-of-Progress* type theories: that some people are afraid of taking our collective feet off the gas because it'll mean backsliding into the metaphorical Dark Ages. It's hard to rediscover how to Do Science and Build Stuff if a critical mass of humanity loses those skills, so the ratchet needs to only ever go one way. Even when there are obvious examples of potentially bad tech, it's important not to get into the habit of ballyhooing them, lest such dismissals become generalized reflexes. (I think I read on this newsletter that Italy has this sort of cultural equilibrium, and it's not great?)
Also, I feel like Debbie Downer should be capitalized as a proper noun. Just like Wendy Wasteful and Holly Hindsight. Proper acknowledgement of names is important!
The point boils down to complaints about the internet/web/social media just being time-wasting and inferior to IRL interactions come from people who have rich options for all of those things. If you grow up in rural North Dakota, you can't substitute in-person salons for reading internet forums; if you grow up in rural India you can't substitute university philosophy classes for YouTube lectures.
To go beyond that, Matt is not substituting watching Kenji make a meatball sub for reading another chapter in Keynes' General Theory, he's just trying to wind down at night. In the 90s he'd be watching a Leno monologue instead. Knowing a little bit more about the mechanics of meatball subs is probably a net gain over Leno's local newspaper headline jokes.
I figured out that this was from back in 2017. Was he directly referring to Matt here? If so, what article? Matt's "The case against Facebook" article was in 2018.
The double-edged nature of most technologies seems so obvious that it is hard to believe someone would dispute the need for regulation. When primitive humans first harnessed fire, they quickly learned that, yes, it could heat their home, but without some regulation, it could also burn their home up.
I am virtually 100% sure that AI is capable of terrible things. Social media's machine-learning algorithms were designed to maximize revenue by maximizing engagement. That this turned out to mean maximizing enragement with ancillary radicalization and polarization was an unintended by-product. I'm sure it wouldn't be too difficult to tweak the algorithms to intentionally maximize enragement and polarization, or incline attitudes more favorably toward Trump (as opposed to Biden), or toward China, or against regulation, etc. Our human brains are much more programmable than we like to think, and AIs will rapidly get better at it. If we wait for unmistakable evidence first, we may be screwed ("Oops, looks like AI enabled the destruction of our democracy...").
BTW, the original internet and its protocols were funded by ARPA, which was an arm of the DoD. And I'm pretty sure DoD saw some important defense applications of a command and control network with decentralized routing and control algorithms that gave it high survivability.
Oh, and the internet has certainly been a huge force for good. Being able to communicate and send large amounts of data almost instantly all over the world has enabled all kinds of life changing applications. Don't confuse the internet itself with some of the stupid or useless applications running on it. That would be like questioning the telephone because some guys robocalling company is annoying.
I realize that most internet-hours-spent are probably social, but it's weird to me that someone would see its main value as being a solution to a social problem. The internet archives, distributes, and makes available lots of kinds of information, in general. It has upended multiple industries, vastly increased our ability to know/remember/understand things, made it harder to lie and not be called out, made it possible to solve much harder coordination problems (while creating new ones, of course), and shortened development timelines of lots of other technologies and products.
I think it's easy to forget how much harder it was to know things even 20 years ago, how much more we relied on what a handful of nearby friends and colleagues happened to (mis)remember. It's jarring when I still see echoes of that from, say, people who post requests for medical advice on Facebook and believe whatever responses they happen to get.
If you think the Nuclear Regulatory Commission was "over-burdensome," you should read "Confessions of a Rogue Nuclear Regulator" by Gregory Jazcko, former chairman of the NRC, explaining how the Commission was largely dominated by industry-friendly members, industry-friendly congressmen, and lobbyists.
Zvi Mowshiwitz called this "the dial of progress fallacy" wet Tyler Cowen and AI. The idea that there is a single knob called 'progress' and you either root for it to go up or down, and that in aggregate this is what matters. Perpetrators of this fallacy generally hold that while reality is more complicated and you in general want to take things on a case by case basis, the public will get the most benefit if we shout "Push much harder progress (despite ignoring the nuance)!"
You overlook (1) the benefits of cheap fentanyl and (2) technology’s role in addressing fentanyl abuse. For instance, if you’re blown apart or incinerated, yet are also still alive, you will be very glad for fentanyl abundant. Perhaps the abuse epidemic outweighs the therapeutic impact. But the techno-optimist reply is that, to solve abuse, we need better enforcement and treatment. Those solutions have normatively political components, like more funding for police. Yet also, technology can improve interdiction, if we let it. Technology will also increase economic growth, thereby allowing easier trade-offs to adequately fund enforcement.
I still cannot read the post./ It send me to and invitation to subscribe, but I AB subscribed What's going on?
Milan! Help! :)
"You can’t go down to Walmart and buy a tactical nuclear weapon. They don’t sell ICBMs over the internet. I don’t totally understand the pragmatic concessions conservatives made in deciding there are limits to the second amendment, but I’m glad they made them."
The Second Amendment mentions "arms" not "ordnance." In other words, firearms not artillery. These aren't "pragmatic concessions."
I hope Boswell explains this. :)
At any rate SCOTUS misinterprets this. It is permission for states to have militias, not individuals to own firearms to protect themselves against the "gommit"
Andreesen’s podcasts appearances (I’m thinking of Sam Harris specifically) after he dropped this article were really embarrassing. I’m told this guy is a visionary but to me sounds like an even more arrogant Elon Musk without the credentials.
I think what Andreesen is going for is an "on net" position. Clearly some tech has some downsides, but he's arguing that the upsides of accelerationist policy are so much greater than the downsides that we would be better off with a damn the torpedoes attitude... on net.
I don't personally agree with this position, but I think the idea that Andreesen is unaware of some of the downsides of technology is itself something of a straw man.
While I generally agree, I wonder how you address the fairly significant number of people who do not endorse technological progress, i.e., Amish/Mennonites, Haridi Judaism, etc. I see these as people who consciously decide to reject in part the modern world. To some extent they're free-riders, but is that too simplistic?
No more than anyone who does not "support" growth and technology yet still benefits.
Your points are why "techno-optimism in general" is just a weird playing field on which to have a meaningful discussion.
What seems true is that on average technological advancement makes us better off and that we are so bad at predicting what advances will be harmful it's better to just adopt a policy of encouraging the development.
Yes, you are totally correct that the internet has had some really negative consequences...but, ironically, the negatives mostly stem from the things that everyone assumed would make us better off while the things that scared people (job loss, competition from other countries, harms from porn etc etc) mostly didn't materialize as harms. The harms were the result of empowering the public to participate in politics and connecting them with information and giving them the ability to share it.
If you'd asked someone back then to regulate the dangers they'd have mostly regulated the parts that have made lives better and worked to encourage the parts that have made things worse.
I think it's important to distinguish a number of different claims:
1) Its never the case that technical Innovation discovers something that makes us on net worse.
2) It's never the case that we can (as a polity/government) effectively predict what advances will make us worse off and on average such advances make us better off.
3) Its not the case that we can both predict what advances in tech will make us worse off and prevent others from making those advances and we are better off making those advances and regulating applications than trying to prevent them from being made.
I think most techno-optimists believe something much closer to 2 or 3.
Always appreciate a good "Reversed Stupidity Is Not Intelligence" post. Andreesen's position, while neither unexpected nor novel, updates me in favour of Dial-of-Progress* type theories: that some people are afraid of taking our collective feet off the gas because it'll mean backsliding into the metaphorical Dark Ages. It's hard to rediscover how to Do Science and Build Stuff if a critical mass of humanity loses those skills, so the ratchet needs to only ever go one way. Even when there are obvious examples of potentially bad tech, it's important not to get into the habit of ballyhooing them, lest such dismissals become generalized reflexes. (I think I read on this newsletter that Italy has this sort of cultural equilibrium, and it's not great?)
Also, I feel like Debbie Downer should be capitalized as a proper noun. Just like Wendy Wasteful and Holly Hindsight. Proper acknowledgement of names is important!
*https://thezvi.substack.com/p/the-dial-of-progress/
Andreesen has a response to Matt on the internet, here's a clip of him being interviewed by Kara Swisher where he lays it out: https://www.youtube.com/watch?v=HOn9_StvCwM
The point boils down to complaints about the internet/web/social media just being time-wasting and inferior to IRL interactions come from people who have rich options for all of those things. If you grow up in rural North Dakota, you can't substitute in-person salons for reading internet forums; if you grow up in rural India you can't substitute university philosophy classes for YouTube lectures.
To go beyond that, Matt is not substituting watching Kenji make a meatball sub for reading another chapter in Keynes' General Theory, he's just trying to wind down at night. In the 90s he'd be watching a Leno monologue instead. Knowing a little bit more about the mechanics of meatball subs is probably a net gain over Leno's local newspaper headline jokes.
I figured out that this was from back in 2017. Was he directly referring to Matt here? If so, what article? Matt's "The case against Facebook" article was in 2018.
Not referring directly to Matt, but to the common criticism of pro-internet takes.
Thanks.
Breaking: decomposing corpse of Mark Andreesen found; website maintained by AI.