I’m not a fan of this rant. soatok is basically getting mad at a bunch of random, disconnected phenomena (mostly) related to the broad technologies of cryptocurrencies and LLMs, which don’t have much to do with each other. I agree with him on some of those points, disagree with him on others, and overall I don’t think that these phenomena taken together say anything in particular about the trajectory of the tech industry, either positive or negative.
In brief I agree that most blockchain projects are probably not useful to anybody in the long term; I think that bitcoin in particular is extremely useful as a way of implementing censorship-resistant money (but it hasn’t been profitable to mine bitcoin on a GPU in a long time, and there’s nothing wrong with that state of affairs); and I don’t think that the specific buzzwords around cryptocurrency are more or less objectionable than buzzwords around any other computer technology.
As far as generative AIs go, I think they’re a cool technology with a lot of potential uses, not all of which I personally will like or care about, but that seems no different than any other previous computer technology. I don’t actually think it’s bad if a blog platform has a button to AI-generate an image, and this is not the reason I don’t use wordpress. I’m not sure how I should feel about Mozilla’s Privacy-Preserving Attribution thing, this seems potentially problematic (although also a potentially-interesting cryptographic protocol) and Mozilla has certainly made plenty of decisions around Firefox I think are bad, but I don’t know what this has to do any kind of AI.
I’m skeptical of intellectual property law as it is and I’m not particularly bothered by current LLMs violating the intellectual property of artists. The real concern for artists is that LLMs can generate certain classes of art for much cheaper than any human’s time is worth, which will probably result in some loss of commission business for artists whose customers turn to an LLM instead, but more often there will just be more art in contexts where previously it wasn’t worth the transaction cost to commission art from an artist at all.
The main thing I wanted to convey today was, “No, you’re not alone, things are getting stupider,” to anyone who wondered if there was a spark of sanity left in the tech sector.
I don’t think the tech industry is getting stupider, or more insane. I think that companies that control the software people use often make product decisions about that software that work against the interest of the users, which is bad, but not novel.
Years ago, I wrote on Medium, but got tired of the constant pressure to monetize my blog, so I decided to pay for a WordPress.com account. I write for myself, after all, and don’t expect any compensation for it.
This is a genuine question - why are you paying for Wordpress at all? Why not just spend a few hours converting your existing blog to a static site, using any of the myriad static site generators that exist, or Ghost if you want to do a Substack-style thing? What value does Wordpress ad to someone who has the skillset to set up a blog themselves?
I sympathize with the author, but I think this approach to talking about it is, frankly, reactionary. We must analyze the political economy of AI tooling and understand its antecedents and where it might be headed.
One place I would suggest is the notion of Intellectual Property. The AI field are essentially opportunists, attempting the same sort of regulatory arbitrage as Airbnb or Uber. They have an interest in negating IP because they need it to train their models. While the specific VCs underwriting AI may benefit from this, VC as a class is wholly dependent on IP to exist. This is an antagonistic contradiction.
The monkey selfie case should be instructive here: a Macaque cannot be an author, because IP law only applies to humans. This means that AI systems cannot be authors either, which means AI output is very dubious when it comes to ownership.
Obligatory note that, speaking as a rationalist-tribe member, to a first approximation nobody in the community is actually interested in the Basilisk and hasn’t been for at least a decade. As far as I can tell, it’s a meme that is exclusively kept alive by our detractors.
Also, “this sure looks like a religion to me” can be - and is - argued about any human social activity. I’m quite happy to see rationality in the company of, say, feminism and climate change.
Finally, of course, it is very much not just rationalists who believe that AI represents an existential risk. We just got there twenty years early.
I’m not sure that the web server is supposed to return a Content-Type header of application/activity+json, and what a web browser is supposed to do with that… unless I’m missing some meta-reference about the title.
I’m not a fan of this rant. soatok is basically getting mad at a bunch of random, disconnected phenomena (mostly) related to the broad technologies of cryptocurrencies and LLMs, which don’t have much to do with each other. I agree with him on some of those points, disagree with him on others, and overall I don’t think that these phenomena taken together say anything in particular about the trajectory of the tech industry, either positive or negative.
In brief I agree that most blockchain projects are probably not useful to anybody in the long term; I think that bitcoin in particular is extremely useful as a way of implementing censorship-resistant money (but it hasn’t been profitable to mine bitcoin on a GPU in a long time, and there’s nothing wrong with that state of affairs); and I don’t think that the specific buzzwords around cryptocurrency are more or less objectionable than buzzwords around any other computer technology.
As far as generative AIs go, I think they’re a cool technology with a lot of potential uses, not all of which I personally will like or care about, but that seems no different than any other previous computer technology. I don’t actually think it’s bad if a blog platform has a button to AI-generate an image, and this is not the reason I don’t use wordpress. I’m not sure how I should feel about Mozilla’s Privacy-Preserving Attribution thing, this seems potentially problematic (although also a potentially-interesting cryptographic protocol) and Mozilla has certainly made plenty of decisions around Firefox I think are bad, but I don’t know what this has to do any kind of AI.
I’m skeptical of intellectual property law as it is and I’m not particularly bothered by current LLMs violating the intellectual property of artists. The real concern for artists is that LLMs can generate certain classes of art for much cheaper than any human’s time is worth, which will probably result in some loss of commission business for artists whose customers turn to an LLM instead, but more often there will just be more art in contexts where previously it wasn’t worth the transaction cost to commission art from an artist at all.
I don’t think the tech industry is getting stupider, or more insane. I think that companies that control the software people use often make product decisions about that software that work against the interest of the users, which is bad, but not novel.
This is a genuine question - why are you paying for Wordpress at all? Why not just spend a few hours converting your existing blog to a static site, using any of the myriad static site generators that exist, or Ghost if you want to do a Substack-style thing? What value does Wordpress ad to someone who has the skillset to set up a blog themselves?
I sympathize with the author, but I think this approach to talking about it is, frankly, reactionary. We must analyze the political economy of AI tooling and understand its antecedents and where it might be headed.
One place I would suggest is the notion of Intellectual Property. The AI field are essentially opportunists, attempting the same sort of regulatory arbitrage as Airbnb or Uber. They have an interest in negating IP because they need it to train their models. While the specific VCs underwriting AI may benefit from this, VC as a class is wholly dependent on IP to exist. This is an antagonistic contradiction.
The monkey selfie case should be instructive here: a Macaque cannot be an author, because IP law only applies to humans. This means that AI systems cannot be authors either, which means AI output is very dubious when it comes to ownership.
Obligatory note that, speaking as a rationalist-tribe member, to a first approximation nobody in the community is actually interested in the Basilisk and hasn’t been for at least a decade. As far as I can tell, it’s a meme that is exclusively kept alive by our detractors.
Also, “this sure looks like a religion to me” can be - and is - argued about any human social activity. I’m quite happy to see rationality in the company of, say, feminism and climate change.
Finally, of course, it is very much not just rationalists who believe that AI represents an existential risk. We just got there twenty years early.
I’m not sure that the web server is supposed to return a Content-Type header of
application/activity+json, and what a web browser is supposed to do with that… unless I’m missing some meta-reference about the title.It’s been working for me since just a few minutes after your comment!