One funny/sad thing that I know might happen goes like this:
A decade ago I wrote a couple Zombie stories to express myself. Noone read them…
So as an efficient machine AI wont scrape these stories, it wont find the pen-name anywhere on the internet, so it’ll just say, “I’ll just repost them verbatim and change the names” and noone will know, of course unless it becomes a great AI novel licensed by Netflix and then I bring out the receipts…
Well there will be problems and there will be solutions. Proof -of-human is already a thing in Blockchain space. If dating apps get inundated with chatbots new dating apps will evolve emphasizng real in person interactions.
If llm can write good enough fiction by themselves and stable diffusion good enough art. Well so these things become abundant. I see no problem with that.
We don't have GAIs yet. When we do things might be different ( as humans themselves might become obsolete). But with current crop of tools? Llm is just another one in the stack.
I guess I think about ChatGPT as less “creating more content” and more “navigating content better.” I use ChatGPT instead of Google now for exactly the reasons you mention (everything on Google is SEO content with nothing useful in it.) But I fact check sources of course 🤓
I have been wondering though if we’ll just go back to paper to cut out the noise!
Yeah I'm not saying there aren't uses, but the way ChatGPT is hyped and promoted is definitely about creating text. Writing emails, essays, cover letters, etc. for people.
I'd love to see more focus on using LLMs to winnow down text though. If I'm a researcher, having chatgpt (accurately) tell me which 10 of 1,000 articles to read would be a big help.
Mar 18, 2023·edited Mar 18, 2023Liked by Lincoln Michel
Nice addition to the current stream of AI-related articles. I can never argue *against* a little optimism; but, there is one thing no one seems to be considering. You say, for example, "An odd thing I’ve noticed about the chatbot hype is that almost every use case works for select individuals but backfires with mass adaptation," as though the real goal of all of this is mass adoption of AI as a tool.
But, this duplicates the same mistake that everyone has made for every new media-technology. We always think that these things (TV, Computers, The Internet, Social Media Algorithms) are being rolled out to *help us*, to be tools *for us*. But, historically, there is really very little evidence for that...and least not over time. In fact, Media Reform and Media Literacy scholars have always repeatedly pointed out that we, in fact, are the product.
And, wrt, to social media and AI, *we* are the currently the tools, *not* the AI. Our input and our data are being used to help them build a product that, eventually, will be used ON the general populace, not BY the general populace. —Oh, sure, there will be aspects of it from which we'll derive some limited benefit (like any drug). But just as social media stopped really being about networking and was quickly turned into nothing but surveillance capitalism, so, too, will AI become whatever Musk and Thiel and his crowd want it to be. I, for one, find that worrying.
I think you're quite right and this is well said. FWIW, I don't think this AI stuff is being rolled out for us in the sense of for our benefit. I meant more than they seem to expect us to pay and/or use them and for a lot of the "use cases" I'm not sure that will happen.
But you're quite right that surveillance capitalism and targeted advertising, etc. might be more the endgame.
This feels exactly correct. I've been thinking for a while that the next phase of the internet is a return to curated link directories and forum culture....ie, 2001 all over again? Who would have guessed. That's an especially astute point about how every problem AI solves, cannot be solved at scale. It's such a short-sighted solution.
It's a very Kantian point, relative to... well, I don't know that the tech world is utilitarian, because that would imply any real philosophical grounding that's not a post-facto justification for doing what they would have done anyway.
“So perhaps the A.I. revolution will just shift how this is all handled and produce nothing except even more money for a few tech companies.” Maybe it’s time we send copies of Vonnegut’s “Player Piano” to the googlers and twits and metapeoples.
At what point did everyone forget what the 'Internet' is? It's not the text, images, audio and videos we and the AIs ingest and wax lyrical about. It's the stuff that transmits the data which represents those things. The Internet was designed at its outset to be highly resilient. Decades later it's still doing its job. So either propose a breakage mechanism or stop talking about things being (potentially) broken because it makes for better FUD. Otherwise the counterfactual and distressingly unimaginative have already won.
I think I was pretty explicit I was discussing the current social media age of the internet where large social platforms like Facebook, Twitter, and TikTok dominate and not talking about the protocols of the internet. The internet, in this sense, has gone through different forms and will go through others. Thus all the talk of "Web 3" in the tech world.
Saying that the internet isn't what people think of as the internet but is just data protocols seems like saying that the US political system isn't the two dominant parties but is the protocols for picking electors and counting votes.
They call it “The Information Highway”, It needs billions in maintenance, you always have to create more of it, it’s riddled with traffic, and don’t even get me started of what happens when it snows hard…
Social platforms are a subset of the totality of internet data being transmitted. Increasing their volume won't "break the internet", neither will AIs sat on cloud instances. Web 3 had predecessors. I'm merely encouraging you to use them as descriptors.
I do hope they break the internet or rather than the internet social media as it exists now, I believe tabula rasa is the only way to save the web's social sphere.
One funny/sad thing that I know might happen goes like this:
A decade ago I wrote a couple Zombie stories to express myself. Noone read them…
So as an efficient machine AI wont scrape these stories, it wont find the pen-name anywhere on the internet, so it’ll just say, “I’ll just repost them verbatim and change the names” and noone will know, of course unless it becomes a great AI novel licensed by Netflix and then I bring out the receipts…
Well there will be problems and there will be solutions. Proof -of-human is already a thing in Blockchain space. If dating apps get inundated with chatbots new dating apps will evolve emphasizng real in person interactions.
If llm can write good enough fiction by themselves and stable diffusion good enough art. Well so these things become abundant. I see no problem with that.
We don't have GAIs yet. When we do things might be different ( as humans themselves might become obsolete). But with current crop of tools? Llm is just another one in the stack.
Hey! For some of us, the jokes and idiocy are the truth!
I guess I think about ChatGPT as less “creating more content” and more “navigating content better.” I use ChatGPT instead of Google now for exactly the reasons you mention (everything on Google is SEO content with nothing useful in it.) But I fact check sources of course 🤓
I have been wondering though if we’ll just go back to paper to cut out the noise!
Yeah I'm not saying there aren't uses, but the way ChatGPT is hyped and promoted is definitely about creating text. Writing emails, essays, cover letters, etc. for people.
I'd love to see more focus on using LLMs to winnow down text though. If I'm a researcher, having chatgpt (accurately) tell me which 10 of 1,000 articles to read would be a big help.
Nice addition to the current stream of AI-related articles. I can never argue *against* a little optimism; but, there is one thing no one seems to be considering. You say, for example, "An odd thing I’ve noticed about the chatbot hype is that almost every use case works for select individuals but backfires with mass adaptation," as though the real goal of all of this is mass adoption of AI as a tool.
But, this duplicates the same mistake that everyone has made for every new media-technology. We always think that these things (TV, Computers, The Internet, Social Media Algorithms) are being rolled out to *help us*, to be tools *for us*. But, historically, there is really very little evidence for that...and least not over time. In fact, Media Reform and Media Literacy scholars have always repeatedly pointed out that we, in fact, are the product.
And, wrt, to social media and AI, *we* are the currently the tools, *not* the AI. Our input and our data are being used to help them build a product that, eventually, will be used ON the general populace, not BY the general populace. —Oh, sure, there will be aspects of it from which we'll derive some limited benefit (like any drug). But just as social media stopped really being about networking and was quickly turned into nothing but surveillance capitalism, so, too, will AI become whatever Musk and Thiel and his crowd want it to be. I, for one, find that worrying.
I think you're quite right and this is well said. FWIW, I don't think this AI stuff is being rolled out for us in the sense of for our benefit. I meant more than they seem to expect us to pay and/or use them and for a lot of the "use cases" I'm not sure that will happen.
But you're quite right that surveillance capitalism and targeted advertising, etc. might be more the endgame.
Yes that’s exactly what I use it for! Research! It’s so helpful to just have it line up all the stats and sources I need.
This feels exactly correct. I've been thinking for a while that the next phase of the internet is a return to curated link directories and forum culture....ie, 2001 all over again? Who would have guessed. That's an especially astute point about how every problem AI solves, cannot be solved at scale. It's such a short-sighted solution.
It's a very Kantian point, relative to... well, I don't know that the tech world is utilitarian, because that would imply any real philosophical grounding that's not a post-facto justification for doing what they would have done anyway.
“So perhaps the A.I. revolution will just shift how this is all handled and produce nothing except even more money for a few tech companies.” Maybe it’s time we send copies of Vonnegut’s “Player Piano” to the googlers and twits and metapeoples.
At what point did everyone forget what the 'Internet' is? It's not the text, images, audio and videos we and the AIs ingest and wax lyrical about. It's the stuff that transmits the data which represents those things. The Internet was designed at its outset to be highly resilient. Decades later it's still doing its job. So either propose a breakage mechanism or stop talking about things being (potentially) broken because it makes for better FUD. Otherwise the counterfactual and distressingly unimaginative have already won.
I think I was pretty explicit I was discussing the current social media age of the internet where large social platforms like Facebook, Twitter, and TikTok dominate and not talking about the protocols of the internet. The internet, in this sense, has gone through different forms and will go through others. Thus all the talk of "Web 3" in the tech world.
Saying that the internet isn't what people think of as the internet but is just data protocols seems like saying that the US political system isn't the two dominant parties but is the protocols for picking electors and counting votes.
They call it “The Information Highway”, It needs billions in maintenance, you always have to create more of it, it’s riddled with traffic, and don’t even get me started of what happens when it snows hard…
Social platforms are a subset of the totality of internet data being transmitted. Increasing their volume won't "break the internet", neither will AIs sat on cloud instances. Web 3 had predecessors. I'm merely encouraging you to use them as descriptors.
I do feel we're reverting back to a web 1.5. I don't think anyone's been satisfied with Web 2.0
I do hope they break the internet or rather than the internet social media as it exists now, I believe tabula rasa is the only way to save the web's social sphere.