Very convincing. Perhaps I'd remove the part about the second law of thermodynamics, because it's not applicable to open systems (as we humans are): a given open system (an animal, a bacteria, a human) can be entropy-negative at the expense of the environment. But this is a technical detail (I can't get rid of my nerdy nature) and doesn't undermine the article's point.
Alberto , what is your position on Marc Andreessen’s excellent progressive article “why AI will save the world” that is pulling the hood off headlines like this one?
I think it's a good article that eloquently argues in favor of his particular position. I don't think there's any conflict between that article and mine. Marc says:
"Now, it is certainly the case that many new technologies have led to bad outcomes – often the same technologies that have been otherwise enormously beneficial to our welfare. So it’s not that the mere existence of a moral panic means there is nothing to be concerned about."
And I think my article falls in the category of "there are things to be concerned about" here. He might disagree, though. I don't know.
I've thought (and written) a lot about how LLMs in particular are changing the nature of human knowledge. The TL;DR is that blocks of ideal text are being used to represent nuanced ideas in a way that's good, but also in a way that will change the nature of the way language itself is used. We won't make words from letters any more; we'll skip straight to making paragraphs and longer stuff from preset sentence fragments.
Just, you know, to throw something even weirder at this! Not that we needed it.
Thanks for sharing Alberto. Which role do you think 'new' social networks like Substack (opt-in subscription feed) play on this? Is creating high-quality content worth it when your readers pay for it?
I believe Substack is comfortably sitting at the very top in terms of quality writing. I've been on Medium for years (the closest platform in structure and goals, IMO) and the quality is simply incomparable nowadays.
It's hard to get people to pay, of course, much harder than growing a free newsletter, but if you manage to succeed, having an engaging and interested audience is worth every minute of effort.
I sense that your article is very timely and unfortunately quite accurate. I expect we are either in a golden era or that we have just passed through one without realizing the gravitas of it. Actually, laziness is a quality of many biological organisms if you think about it in the macro scale. It was made apparent to me when I learned about sharks not going out of their way to find properly vetted or identified meal, when they might more easily bite a candidate that’s closer just to see if it was the proper contender after all. Going the extra distance, it was explained coherently, would require additional energy and that it wasn’t really in an organism’s interest to act beyond certain thresholds of inefficiently. However, I suspect a lot of humans abuse the privilege.
There are industrious and creative humans, and of course many have refined tastes that will try to not tolerate shit content in any form. While we would all like to think that we might be that type, I’m also sure we all have our limits and would also cave in under the right conditions. My brother is a gourmet chef of impeccable standards, yet I’m sure he has his guilty pleasures like any human. Nonetheless I suspect there’s going to be certain bifurcations in our species, with some being far more shit tolerant than others. Some will positively thrive in a world of shit.
I realize it’s kind of an idealized picture, but I often think of characters from Star Trek: TNG, in that none of them are just trying to make a buck. In fact capitalism in that system doesn’t even exist in their culture anymore. So, they are all striving for self-improvement and mastery of their tasks. That does seem like the ideal. In fact there was one interesting episode plot in which they discovered some cryogenically frozen humans from the 20th century. One of them was incredibly fixated on his investments, and was unimaginably disappointed to learn that those weren’t really of any consequence anymore. He couldn’t quite get his head around that. He had dreams of compounded value, yet not everything goes to 20th century plan.
I think a lower half of the bifurcation may manifest in something I regard as the “Doritos
Islands”, in which humans are allowed to squander their lives (perhaps using their UBI), and do not much of anything except existing and being entertained. I envision there must be certain stipulations such as giving up your right to vote right and you’re right to breed. At least I think that’s a fair part of the bargain. When you don’t bring anything to the existential table than you don’t get to leave a legacy. It may serve to answer the profound questions: How much NASCAR can you actually watch? How many Doritos can you actually eat? You know, the important questions.
"Writing, if you love it, is a pleasure. If you don’t, it’s hard, slow, and tiring. Language models solve that." Love that particular inference!
But the internet nowadays seems to care about the content more (some prefer quantity, and some quality of it) and not care mostly about the 'style' of the writer. This is evident from the fact that people do engage with AI written article, even when it lacks the apparent 'style' of the Author. And they seem pretty okay with it, at least for now.
Later on, people would (hopefully) get sick of AI-written content.
If a great AI flood were, would you agree that it would increase the value of the writers who 'actually write' their pieces, and not merely just 'generate' them; just like how during the age of digital art, people who can actually draw on canvas are respected more?
Great and insightful post, as always. I wonder if, following McLuhan's 4th law of media (the flip: when brought to the limit technology tends to flip into reverse) this may not lead to an unexpected revival in the power of traditional non-algorithmic gate keepers. When people are bombarded by gargantuan waves of bullshit overwhelming their limited attention they may end up turning to more trusted, human of sources of writing and text curating. Don't you think.
Also, now is also a good time to say I've been enjoying your writing immensely. Thank you for your work.
Great article! It couldn't have come at a better time. The truth is, keeping a close watch on the accuracy of information has always been an arduous task. We've all been guilty of it because the whole purpose of expressing our opinions is to make our point heard. However, what makes things even more perilous are the hallucinations created by these AI algorithms. It's uncertain whether humans are equipped to be the ultimate guardians in this battle. The challenge is real, and it's time we address it head-on.
I don't disagree with the point its contributing to - and I think your overall case is strong - but I've never been a fan of characterising the natural human instinct to preserve precious energy as meaning we are all inherently 'lazy'. One of the reasons I find Kahneman's stuff tiresome that he leans on this framing so heavily. Is unwarranted moralistic misanthropy really, and of course ironically rather lazy itself
Very convincing. Perhaps I'd remove the part about the second law of thermodynamics, because it's not applicable to open systems (as we humans are): a given open system (an animal, a bacteria, a human) can be entropy-negative at the expense of the environment. But this is a technical detail (I can't get rid of my nerdy nature) and doesn't undermine the article's point.
I know haha but liked the intro to human laziness. But yeah, it's slightly scientifically inaccurate - some artistic freedom ;)
Alberto , what is your position on Marc Andreessen’s excellent progressive article “why AI will save the world” that is pulling the hood off headlines like this one?
I think it's a good article that eloquently argues in favor of his particular position. I don't think there's any conflict between that article and mine. Marc says:
"Now, it is certainly the case that many new technologies have led to bad outcomes – often the same technologies that have been otherwise enormously beneficial to our welfare. So it’s not that the mere existence of a moral panic means there is nothing to be concerned about."
And I think my article falls in the category of "there are things to be concerned about" here. He might disagree, though. I don't know.
This is well written.
I've thought (and written) a lot about how LLMs in particular are changing the nature of human knowledge. The TL;DR is that blocks of ideal text are being used to represent nuanced ideas in a way that's good, but also in a way that will change the nature of the way language itself is used. We won't make words from letters any more; we'll skip straight to making paragraphs and longer stuff from preset sentence fragments.
Just, you know, to throw something even weirder at this! Not that we needed it.
Thanks for sharing Alberto. Which role do you think 'new' social networks like Substack (opt-in subscription feed) play on this? Is creating high-quality content worth it when your readers pay for it?
Important question!
I believe Substack is comfortably sitting at the very top in terms of quality writing. I've been on Medium for years (the closest platform in structure and goals, IMO) and the quality is simply incomparable nowadays.
It's hard to get people to pay, of course, much harder than growing a free newsletter, but if you manage to succeed, having an engaging and interested audience is worth every minute of effort.
I sense that your article is very timely and unfortunately quite accurate. I expect we are either in a golden era or that we have just passed through one without realizing the gravitas of it. Actually, laziness is a quality of many biological organisms if you think about it in the macro scale. It was made apparent to me when I learned about sharks not going out of their way to find properly vetted or identified meal, when they might more easily bite a candidate that’s closer just to see if it was the proper contender after all. Going the extra distance, it was explained coherently, would require additional energy and that it wasn’t really in an organism’s interest to act beyond certain thresholds of inefficiently. However, I suspect a lot of humans abuse the privilege.
There are industrious and creative humans, and of course many have refined tastes that will try to not tolerate shit content in any form. While we would all like to think that we might be that type, I’m also sure we all have our limits and would also cave in under the right conditions. My brother is a gourmet chef of impeccable standards, yet I’m sure he has his guilty pleasures like any human. Nonetheless I suspect there’s going to be certain bifurcations in our species, with some being far more shit tolerant than others. Some will positively thrive in a world of shit.
I realize it’s kind of an idealized picture, but I often think of characters from Star Trek: TNG, in that none of them are just trying to make a buck. In fact capitalism in that system doesn’t even exist in their culture anymore. So, they are all striving for self-improvement and mastery of their tasks. That does seem like the ideal. In fact there was one interesting episode plot in which they discovered some cryogenically frozen humans from the 20th century. One of them was incredibly fixated on his investments, and was unimaginably disappointed to learn that those weren’t really of any consequence anymore. He couldn’t quite get his head around that. He had dreams of compounded value, yet not everything goes to 20th century plan.
I think a lower half of the bifurcation may manifest in something I regard as the “Doritos
Islands”, in which humans are allowed to squander their lives (perhaps using their UBI), and do not much of anything except existing and being entertained. I envision there must be certain stipulations such as giving up your right to vote right and you’re right to breed. At least I think that’s a fair part of the bargain. When you don’t bring anything to the existential table than you don’t get to leave a legacy. It may serve to answer the profound questions: How much NASCAR can you actually watch? How many Doritos can you actually eat? You know, the important questions.
"Writing, if you love it, is a pleasure. If you don’t, it’s hard, slow, and tiring. Language models solve that." Love that particular inference!
But the internet nowadays seems to care about the content more (some prefer quantity, and some quality of it) and not care mostly about the 'style' of the writer. This is evident from the fact that people do engage with AI written article, even when it lacks the apparent 'style' of the Author. And they seem pretty okay with it, at least for now.
Later on, people would (hopefully) get sick of AI-written content.
If a great AI flood were, would you agree that it would increase the value of the writers who 'actually write' their pieces, and not merely just 'generate' them; just like how during the age of digital art, people who can actually draw on canvas are respected more?
Do you have an example of where people engage with an article that was obviously written by AI? I'd be curious to see that.
People who can draw on canvas are certainly respected more, but do they earn more money? I'm guessing no (unfortunately).
The flood is coming, definitely
Great and insightful post, as always. I wonder if, following McLuhan's 4th law of media (the flip: when brought to the limit technology tends to flip into reverse) this may not lead to an unexpected revival in the power of traditional non-algorithmic gate keepers. When people are bombarded by gargantuan waves of bullshit overwhelming their limited attention they may end up turning to more trusted, human of sources of writing and text curating. Don't you think.
Also, now is also a good time to say I've been enjoying your writing immensely. Thank you for your work.
Appreciate it Ido! And I agree. I think we will eventually turn to more trusted sources. At least those of use who are aware of this happening at all.
I see a role for Google and Bing here. Perhaps an upvote system, similar to Reddit, on search results to filter out AI garbage.
Isn’t that a losing battle? Valuable human time used to filter out low quality cheaply generated and voluminous AI content
Great article! It couldn't have come at a better time. The truth is, keeping a close watch on the accuracy of information has always been an arduous task. We've all been guilty of it because the whole purpose of expressing our opinions is to make our point heard. However, what makes things even more perilous are the hallucinations created by these AI algorithms. It's uncertain whether humans are equipped to be the ultimate guardians in this battle. The challenge is real, and it's time we address it head-on.
The internet has been dying for quite some time well before generative AI came along.
Muchas gracias.
Valioso aporte. Ya me imagino las próximas elecciones presidenciales de cada país. Si con las redes sociales era una MIERDA, ahora será una IAerda.
I don't disagree with the point its contributing to - and I think your overall case is strong - but I've never been a fan of characterising the natural human instinct to preserve precious energy as meaning we are all inherently 'lazy'. One of the reasons I find Kahneman's stuff tiresome that he leans on this framing so heavily. Is unwarranted moralistic misanthropy really, and of course ironically rather lazy itself