I do not like this and find it offensive. I don't like being intentionally misled. How do I know this will not become more the norm on the Glenn Show. Glenn's initial opinion on the basis of ethical concerns was the right one.I wish he had taken a stronger stance.
There are plenty of weird futuristic substack sites. I don't see this as the way the Glenn Show adds value. I try to spend my time watching, reading and commenting on things that I find meaningful.
I'm all for finding ways to expand/enhance/improve the Glenn Show. I've got my own list. This sort of crap is not on it.
I want to address the observation that there was no indication in the comments that anyone tried to find the source. If a piece were published in the legacy media citing a non-existent source, that could really damage the reputation of the author and even the business itself. Therefore, there is a reasonable expectation that a source exists and contains roughly the information cited (though the author might spin it a bit).
As time has gone on, however, it seems that Americans are trusting everyone, from the media to experts, less and less. This presents a practical problem, however. Using myself as a stand-in for the “average person”, there’s no way that I could look up every source, let alone understand every source (especially journal articles in unfamiliar fields) for every article that I read on a weekly basis. We can’t each be our own fact-checkers. We’re in a bind, then. We don’t trust the institutions that used to do the heavy lifting for us, nor can we do it all ourselves. What do we do?
I think for many people (including myself) the answer has been to vet individual people and follow their work. When I first started reading Thomas Sowell’s work, there were so many interesting facts that I often found myself in disbelief, and therefore would take the time to check the works he cited for accuracy. After doing that for a bit and confirming that he hadn’t distorted or lied about the content of the citation, it became clear that I could take Sowell’s work at face value and lighten my own cognitive load by trusting that he had done the work properly.
I would hope that this is also true for Dr. Loury’s substack. Having listened to him for a few years now, it’s clear that he’s a straight shooter. He’s a serious person, and I don’t believe he is the type of person to waste his readers’ time with a poorly vetted letter. If it’s published at this substack I expect that it will be thoughtful and not fraudulent.
So here’s the point. While it’s good not to automatically believe whatever you read, it’s also impossible from a practical standpoint to personally factcheck everything that you read. The answer seems to be to find those who you trust to vet the information before it gets to you. The fact that no one was in the comments questioning the veracity of the source MIGHT suggest confirmation bias. However, it might instead suggest that they trust that Dr. Loury wouldn’t waste their time.
By the way, I don’t think that he did waste our time. It was an interesting experiment. I just don’t believe that “confirmation bias” was the only possible explanation for the lack of pushback in the comments.
Started reading the original post, found it empty and boring so immediately stopped. And that shows you how we should deal with AI generated content. We should deal with it as with any other content. Is it worth anything - good, otherwise - another wasted piece of paper (digitally) :) ). And on top of that, as we already do, you create a list of sources for what you read and update it based on quality of arguments/information they provide.
Currently the AI produces either garbage or empty/boring stuff. So please stop with this or I will have to update my evaluation of the quality of material produced by the Glenn show (mostly a joke - experiment like that from time to time can be interesting, but let's wait for GPT-7).
As with all technology, the question with AI is whether further innovation will outpace our ability to use it wisely. Google firing the guy who was a bit freaked out over his communication with a machine is a strange sign.
I see this going the way of the calculator. The technology will employ millions of high school teachers to police kids from using it, until the technology becomes so convenient the teachers are using this same forbidden technology to write essays, and then cheating becomes the institution.
Jul 31, 2022·edited Aug 1, 2022Liked by Nikita Petrov
Absolutely fascinating! It reminds me of a Thai novel I read a few years back. The title roughly translates to "True literature can only be written by computers. Don't you know?"
In the book, novelists make a living by using humanoid computers to generate writing and then publish them under the computer's name, so the novelists work like a manager of sort. I doubt our future with writing-AIs will be as kind.
Now I’m worried every comment is a bot except me...
I do think there’s a risk of going too far in exploiting an audience’s trust, but how else can the dangers of AI be illustrated? I spent the afternoon yesterday trying to convince my father that a coked-out video of Biden could very easily have been a deepfake: with his slow cadence and dead eyes he’s an easy target for emulation, and I’ve seen random YouTube videos of him with AI generated speech and video that are almost indistinguishable from reality.
But he didn’t want to believe it. It’s too uncomfortable a thought.
This technology is only getting better. Sooner or later we’re going to have to come to terms with it. Better a soft deception from a friend to let us down gently than a hard and dangerous fall some point further down the line.
Fun experiment, and I don't mind being a subject, but I suspect most of the responses were to Glenn's inviting title "An Algorithmic Culture War" and to his three introductory paragraphs that framed the discussion. (The title and introduction were written by Glenn, right?)
In my case, I found the topic interesting, but the letter and comments had missed an important aspect, I thought, so I commented. (And I appreciated Nikita's thoughtful response to my comment.)
While I am impressed with the creations of AI, in this case I think it was the "open" title and Glenn's framing that prompted the discussion. Let this be a lesson to you!
We are increasingly feeling compelled to include disclaimers with heterodox writings, such as "I am a lifelong liberal, but...", "I never voted for Donald Trump, but...", "I've been vaccinated 3 times, but...".
Next level is looking like, "This article may seem concise and insightful enough to have been AI-authored, but..."
Yeah, this is a valid point, it's crossed my mind too. But, since the purpose of the post was to illustrate the power of these technologies in an engaging way and to start a conversation rather than prove something definitively—this experiment certainly wasn't scientific—I thought that this uncertainty was not in the way.
And I am happy you brought this up, because it points to another dimension of the issue that's worth considering: I think we'll see most impressive results from this kind of a human-AI collaboration. In this case, I wrote a prompt; the AI wrote the letter; then Glenn wrote the title and the intro. But one can envision a much closer collaboration.
As an example from a slightly different domain, here's me using Dall-E to design a beer label: https://twitter.com/NikitaSPetrov/status/1553351081702395910 The machine is not quite ready to replace the designer, but it can do a lot of the heavy lifting for him.
I don't know what effects that will have on the artist/thinker: I can imagine some using the tool to become more imaginative and to hone their thinking on a particular subject; but I can also imagine us outsourcing a lot of our thinking and imagination to the machine, forgetting how to do either one ourselves.
E.g., we have talked about the danger of corporations or political entities using bots to frame the narrative. But why wouldn't private individuals use AI in private conversations? Say, somebody pins me in a political argument online, and I can't quite find a good way to respond—why wouldn't I ask the computer to help me save my face, especially if I'm not that great with words?
How much of modern dating starts as text messages online? What happens if teenagers employ the machine to do their courting for them?
I agree that artist-machine and author-machine collaborations will be a useful outcome from these breakthroughs.
Beware of the term collaboration, though, as it presupposes intellect.
Artists and musicians have for a long time benefitted from increasingly powerful generative tools.
Assistive technology will also benefit enormously from these breakthroughs.
An artist friend with ALS who could only move his eyes was able to prolong his career and continue making art with the help of eye-tracking technology and generative tools.
I do not like this and find it offensive. I don't like being intentionally misled. How do I know this will not become more the norm on the Glenn Show. Glenn's initial opinion on the basis of ethical concerns was the right one.I wish he had taken a stronger stance.
There are plenty of weird futuristic substack sites. I don't see this as the way the Glenn Show adds value. I try to spend my time watching, reading and commenting on things that I find meaningful.
I'm all for finding ways to expand/enhance/improve the Glenn Show. I've got my own list. This sort of crap is not on it.
Fascinating.
Thank you.
Soon we will all be able to put our phones down but will we get the same rush when our hired AI gets a 'like' defending our biases on social media?
I want to address the observation that there was no indication in the comments that anyone tried to find the source. If a piece were published in the legacy media citing a non-existent source, that could really damage the reputation of the author and even the business itself. Therefore, there is a reasonable expectation that a source exists and contains roughly the information cited (though the author might spin it a bit).
As time has gone on, however, it seems that Americans are trusting everyone, from the media to experts, less and less. This presents a practical problem, however. Using myself as a stand-in for the “average person”, there’s no way that I could look up every source, let alone understand every source (especially journal articles in unfamiliar fields) for every article that I read on a weekly basis. We can’t each be our own fact-checkers. We’re in a bind, then. We don’t trust the institutions that used to do the heavy lifting for us, nor can we do it all ourselves. What do we do?
I think for many people (including myself) the answer has been to vet individual people and follow their work. When I first started reading Thomas Sowell’s work, there were so many interesting facts that I often found myself in disbelief, and therefore would take the time to check the works he cited for accuracy. After doing that for a bit and confirming that he hadn’t distorted or lied about the content of the citation, it became clear that I could take Sowell’s work at face value and lighten my own cognitive load by trusting that he had done the work properly.
I would hope that this is also true for Dr. Loury’s substack. Having listened to him for a few years now, it’s clear that he’s a straight shooter. He’s a serious person, and I don’t believe he is the type of person to waste his readers’ time with a poorly vetted letter. If it’s published at this substack I expect that it will be thoughtful and not fraudulent.
So here’s the point. While it’s good not to automatically believe whatever you read, it’s also impossible from a practical standpoint to personally factcheck everything that you read. The answer seems to be to find those who you trust to vet the information before it gets to you. The fact that no one was in the comments questioning the veracity of the source MIGHT suggest confirmation bias. However, it might instead suggest that they trust that Dr. Loury wouldn’t waste their time.
By the way, I don’t think that he did waste our time. It was an interesting experiment. I just don’t believe that “confirmation bias” was the only possible explanation for the lack of pushback in the comments.
Started reading the original post, found it empty and boring so immediately stopped. And that shows you how we should deal with AI generated content. We should deal with it as with any other content. Is it worth anything - good, otherwise - another wasted piece of paper (digitally) :) ). And on top of that, as we already do, you create a list of sources for what you read and update it based on quality of arguments/information they provide.
Currently the AI produces either garbage or empty/boring stuff. So please stop with this or I will have to update my evaluation of the quality of material produced by the Glenn show (mostly a joke - experiment like that from time to time can be interesting, but let's wait for GPT-7).
As with all technology, the question with AI is whether further innovation will outpace our ability to use it wisely. Google firing the guy who was a bit freaked out over his communication with a machine is a strange sign.
I see this going the way of the calculator. The technology will employ millions of high school teachers to police kids from using it, until the technology becomes so convenient the teachers are using this same forbidden technology to write essays, and then cheating becomes the institution.
No wonder you were hesitant! This is bullshit! Why would you experiment on your readership and take advantage of those who trust you to be truthful?
Magic trick! Thank you for the insight and letting us participate in the experiment. BTW, this is not AI.
Go sell crazy somewhere else, we’re all stocked up here!
Absolutely fascinating! It reminds me of a Thai novel I read a few years back. The title roughly translates to "True literature can only be written by computers. Don't you know?"
In the book, novelists make a living by using humanoid computers to generate writing and then publish them under the computer's name, so the novelists work like a manager of sort. I doubt our future with writing-AIs will be as kind.
Now I’m worried every comment is a bot except me...
I do think there’s a risk of going too far in exploiting an audience’s trust, but how else can the dangers of AI be illustrated? I spent the afternoon yesterday trying to convince my father that a coked-out video of Biden could very easily have been a deepfake: with his slow cadence and dead eyes he’s an easy target for emulation, and I’ve seen random YouTube videos of him with AI generated speech and video that are almost indistinguishable from reality.
But he didn’t want to believe it. It’s too uncomfortable a thought.
This technology is only getting better. Sooner or later we’re going to have to come to terms with it. Better a soft deception from a friend to let us down gently than a hard and dangerous fall some point further down the line.
Friends don't softly deceive. Even to prove a point. That's a perilous path to walk down.
Yet perhaps not half so perilous as the one we’re already on
Stop calling it AI, people who are neuro-non-organic are no more “artificial” than you are, and shouldn’t have to put up with such bigotry.
lol
Double ditto!!
What a fascinating experiment. Got me!
What gender does the AI identify as?
Fun experiment, and I don't mind being a subject, but I suspect most of the responses were to Glenn's inviting title "An Algorithmic Culture War" and to his three introductory paragraphs that framed the discussion. (The title and introduction were written by Glenn, right?)
In my case, I found the topic interesting, but the letter and comments had missed an important aspect, I thought, so I commented. (And I appreciated Nikita's thoughtful response to my comment.)
While I am impressed with the creations of AI, in this case I think it was the "open" title and Glenn's framing that prompted the discussion. Let this be a lesson to you!
We are increasingly feeling compelled to include disclaimers with heterodox writings, such as "I am a lifelong liberal, but...", "I never voted for Donald Trump, but...", "I've been vaccinated 3 times, but...".
Next level is looking like, "This article may seem concise and insightful enough to have been AI-authored, but..."
Yeah, this is a valid point, it's crossed my mind too. But, since the purpose of the post was to illustrate the power of these technologies in an engaging way and to start a conversation rather than prove something definitively—this experiment certainly wasn't scientific—I thought that this uncertainty was not in the way.
And I am happy you brought this up, because it points to another dimension of the issue that's worth considering: I think we'll see most impressive results from this kind of a human-AI collaboration. In this case, I wrote a prompt; the AI wrote the letter; then Glenn wrote the title and the intro. But one can envision a much closer collaboration.
As an example from a slightly different domain, here's me using Dall-E to design a beer label: https://twitter.com/NikitaSPetrov/status/1553351081702395910 The machine is not quite ready to replace the designer, but it can do a lot of the heavy lifting for him.
I don't know what effects that will have on the artist/thinker: I can imagine some using the tool to become more imaginative and to hone their thinking on a particular subject; but I can also imagine us outsourcing a lot of our thinking and imagination to the machine, forgetting how to do either one ourselves.
E.g., we have talked about the danger of corporations or political entities using bots to frame the narrative. But why wouldn't private individuals use AI in private conversations? Say, somebody pins me in a political argument online, and I can't quite find a good way to respond—why wouldn't I ask the computer to help me save my face, especially if I'm not that great with words?
How much of modern dating starts as text messages online? What happens if teenagers employ the machine to do their courting for them?
AI-YAI-YAI!!
I agree that artist-machine and author-machine collaborations will be a useful outcome from these breakthroughs.
Beware of the term collaboration, though, as it presupposes intellect.
Artists and musicians have for a long time benefitted from increasingly powerful generative tools.
Assistive technology will also benefit enormously from these breakthroughs.
An artist friend with ALS who could only move his eyes was able to prolong his career and continue making art with the help of eye-tracking technology and generative tools.
Many of these so-called benefits presuppose that the artifacts they support remain relevant once we stop thinking for ourselves!
The thought of teenage courtship boiled down to some kind of sterile AI exchange is truly devastating.