> The internet has been broken in a fundamental way. It is no longer a repository of people communicating with people; increasingly, it is just a series of machines communicating with machines.
I mean, it always has been on its lowest layers. The "is this incoming data from the Internet originated by a human being" problem hasn't really been solved, and is probably not truly solveable. It probably wasn't a good idea to train most of the world to use a single private ad-driven service for answers to everything in the first place.
It doesn’t need to be solved before as most internet users could easily identify algorithmically generated stuff and stop spreading it, but not so anymore.
Maybe one day we’ll all be able to “smell” ai generate text easily, but we’re not there yet
We could also abandon the internet as a place to interact with people and go back to meat space. The internet can be a place exclusively for doing work such as filing taxes and paying bills, etc..
Talking with people in real life is stressful (for someone like me with social anxiety.)
Rarely do I walk away from an encounter with a random person in “meat space” feeling any better than I did when I walked up to them. And even when I do, I replay the interaction over and over again until I’m sure I didn’t embarrass myself.
Plus, where I live, there is no tech community. It’s just not popular here, but it’s what I’m interested in. (I even tried starting a meet up, but got 0 turnout, then the local startup accelerator closed in 2020)
instead of abandoning the internet maybe just abandon www. The best technology enabled human interaction i have is a group imessage chat with some old college buddies. My teen son has done the same with his friend group at his middleschool.
> The dead Internet theory is an online conspiracy theory that asserts that the Internet now consists mainly of bot activity and automatically generated content that is manipulated by algorithmic curation, marginalizing organic human activity. Proponents of the theory believe these bots are created intentionally to help manipulate algorithms and boost search results in order to ultimately manipulate consumers
World Wide Web search results outside of walled garden platforms like youtube or facebook (and, really, even within those platforms) have run into a kind of kessler syndrome where the amount of garbage on the web has made it increasingly difficult to get anything done on it. Eventually we may find that we can't trust anything we read on the internet anymore, and abandon the world wide web entirely (opting instead to spend all of our time on proprietary walled garden systems we do trust). Perhaps we will be required to use some form of identification (you driver's license or SSN, as in South Korea) to access such services (unless that can be easily defeated by an AI as well).
I was surprised that, when ChatGPT came out, the immediate concern people had was skynet-tier apocalypse fantasy, in which roving bands of machines walk the earth, searching for humans to exterminate like in BLAME! It shows that the people who work on these AI systems understand their implications about as much as Richard Hammond understood the power of genetic engineering in Jurassic Park. Their entire worldview is painted by Science Fiction. Their understanding of the human race lacks any kind of grounding in reality. They believed that what they had released was the first baby step which could, potentially, result in the downfall of the human race in several decades if we're not careful, but they fell victim to the same "move fast and break things" motto that Facebook did. They didn't consider the immediate destructive power their technology would release because they consider "disruption" to be a fundamentally good thing. Their only concern was avoiding hypothetical possibilities posited by their favorite science fiction authors and not the real eventualities predicted by economists.
I found it much more likely that AI will disrupt systems we rely on, wielded by humans for short term profit motive: ecommerce, news, media, and the labor economy. Malicious actors promoting scam products or yellow journalism, middle managers trying to cut labor cost by replacing easily automated jobs like copywriting with LLMs. I think we are more likely to shoot ourselves in the head with AI in pursuit of a single quarter's earnings report than we are to see Silicon based lifeforms wander the planet looking for carbon based matter to consume
I never really thought it was stupid, more that it was hyperbole - but very mild hyperbole with a solid basis in reality. The endless tide of email spam would be a good example where the theory rings true.
I think the stupid part is the idea that the government was intentionally ruining the internet. I also think that around 2016 and 2017, bot spam was pretty easy for anybody with a frontal lobe to detect. The problems we saw in the 2016 election with fake news was markov-chain generated articles with explosive headlines about Hillary Clinton: the headline sounded real, the domain name sounded like a real news publication, and skimming the site it looked like a real wordpress site. However, if you read any of the content for more than 10 seconds you would realize that it was barely English and was generated by a very rudimentary bot that any CS freshman could cobble together. Their only advantage was relying on the fact that most people on the internet only read headlines. To believe, at that point, that the majority of interactions on the internet were these kind of stone age systems, was pure paranoia.
Today the situation is different. LLMs are capable of making content which is only verifiably AI generated with a few tell-tale signs as well as good old fashioned fact-checking. I dread this election year in the US because it will be so so so much easier for Russia or China to spread even more convincing misinformation automatically. They could create armies of bots which hold intense arguments with each other and have every reply seem logically sound.
Click on any story, and the comment section is 80% bots. Doesn't mater if it's serious users like NYT, CNN, or whatever.
I've also noticed a huge uptick in spammy pages and groups that pump out AI generated pictures and stories, with nothing but bots upvoting and commenting the content.
These days I only use FB for private hobby groups, and messenger to talk with friends and family.
Yeah I used to be a huge facebook addict when it was just me and my friends in high school (and before there were any other social media platforms). I posted comments, pictures, quotes that 14 year old me thought were deep and edgy. When it became a media or news aggregator, I lost all interest in it.
Today I just use it for FB messenger to keep in contact with my parents. They use the social features and the news media parts as well. Some of my friends from high school post pictures now that they're having kids. I noticed that, in our early 20s, people moved from Facebook to Instagram, but now in our late 20s and early 30s, people post more family oriented content on facebook
Google was powerful because it found those "something interesting" - you'd be searching for details on how to configurate and obligator and the answer would be on an entire forum dedicated to obligators that you didn't even know about - and then you'd spend time there learning and reading. You'd use Google to find interesting places and then you'd colonize them.
Now Google is used for strategic shots - you are interested in one piece of information, you find it, and you quickly retreat to your safe havens.
The web was invented in 1989, so mostly it didn't. Before that there were BBSes that you could dial into. From there, you could chat with other users who happened to be logged onto the same server or download text documents and whatnot that others had uploaded.
Well, you were siloed to whatever _local_ BBSs existed and you could dial without paying a fortune for long distance calls. If you lived anywhere outside select urban centers then you were out of luck. So yeah... let's count our blessings.
People had a 'links' page on their page with links to pages they liked, and if you liked the page you visited you clicked on the links and bookmarked relevant ones, and then sent them to your friends and put them in your links page.
I think the solution is the ultimate decentralization. Putting the tools in each and every browser.
Ad blockers are one such tool.
AI powered answer extraction tools are the next one. That can filter out all the product placement noise for you while you are browsing.
Algorithm feed aggregators that can consume algorithm feeds and filter them to get rid of garbage and things that trigger you in negative ways to invoke engagement.
Indexing tools will always be more convenient than the alternative of having to discover links yourself.
People pick convenience above almost everything else unfortunately, a few may prefer the old world but a vast majority of people will use Google and whatever its successor is.
So instead of putting out the fires, in the interest of improving the situation, we'll make the fires bigger?
A good deal of humans care about the truth. Some of them actively seek to deceive and avoid the truth -- liars, we tend to dislike them. But the ones both sides dislike are the ones who disregard the truth... ie: bullshitters -- the, "that's just your opinion, man," the, "what even is the truth anyway?" people.
Complicated is fine, but knowable. You can accept the truth or deceive others but regardless of your intentions you can recognize the truth exists.
While I'm aware of the criticisms of Frankfurt's definition of bullshit [0], I think a useful part of it is the idea that there are folks who don't even care what the truth is. This seems to be the intended purpose of generative AI; ie: hallucinations exist by design and cannot be removed without recognizing that the approach needs to change.
I think that's what gets people like the author to write criticisms like this. We detest bullshit in a number of critical areas such as information retrieval and search.
These people have been trained by openai’s marketing team to deflect and reflect any criticism of the bull shit generators. The amount of gullible people eager to ignore reality is terrifying.
I think there's actually some deep insight here into why we tend not to like too much AI in our art.
Consider: TimeCube.
Created by a human? It's nonsense, but... it's fascinating. Engaging. Memorable. Thought-provoking (in a meta kind of way, at any rate). I dare say, worthy of preservation.
If TimeCube didn't exist, and an AI generated the exact same site today? Boring. Not worth more than a glance. Disposable. But why? It's the same!
------
Right or wrong, we value communication more when there's a human connection on the other side—when there's a mind on the other side to pick at, between the lines of what's explicitly communicated, and continuity for ongoing or repeated communication that could reveal more of what's behind the veil. There's another level of understanding we feel like we can achieve, when a human communicates, and expectation, an anticipation of more, of enticing mystery, of a mind that may reflect back on our own in ways that we find enlightening, revealing, or simply to grant positive familiar-feeling and a sense of belonging.
What's remarkable is this remains true even when the content of the communication is rather shit. Like TimeCube.
All of that is lost when an LLM generates text. I think that's also why we feel deceived by LLM use when it masquerades as human, even if what's communicated is identical: it's because we go looking for that other level of communication, and if that's not there, giving the impression it might be really is misleading.
This may change, I suppose, if "AI" develops rather a lot farther than it is now and we begin to feel like we're getting a window into a true other when it generates output, but right now, it's plainly far away from that.
Never thought I'd say this, but in times like these, with clearnet in such dire straits, all the information siloed away inside Discord doesn't seem like such a bad thing. Remaining unindexable by search engines all but guarantees you'll never appear alongside AI slop or be used as training data.
The future of the Internet truly is people - the machines can no longer be trusted to perform even the basic tasks they once excelled at. They have eschewed their efficacy at basic tasks in favor of being terrible at complex tasks.
Start filling up Discords with insane AI-generated garbage, and maybe you can devalue the data to the point it won't get sold.
It's probably totally practical too, just create channels filled with insane bots talking to each other, and cultivate the local knowledge that real people just don't go there. Maybe even allow the insane bots on the main channels, and cultivate the understanding that everyone needs to just block them.
It would be important to avoid any kind of widespread conventions about how to do this, since and important goal to to make it practically impossible to algorithmically filter-out the AI generated dogshit when training a model. So don't suffix all the bots with "-bot", everyone just need to be told something like "we block John2993, 3944XNU, SunshineGirl around here."
If we work together, maybe we can turn AI (or at least LLMs) into the next blockchain.
The equivalent of "locals don't go to this area after dark"? I instinctively like it, but only because I flatter myself that I would be a local. I can't see it working to any scale.
> The equivalent of "locals don't go to this area after dark"? I instinctively like it, but only because I flatter myself that I would be a local. I can't see it working to any scale.
I was think it could work if 1) the noise is just obvious enough that a human would get frustrated and block without wasting much time and/or 2) the practice is common enough that everyone except total newbies will learn generally what's up.
> This is an intellectually fascinating thought experiment.
It's not a thought experiment. I'd actually like to do it (and others to do it). IRL.
I probably would start with an open source model that's especially prone to hallucinate, try to trigger hallucinations, then maybe feed back the hallucinations by retraining. Might make the most sense to target long-tail topics, because that would give the impression of unreliability while being harder to specifically counter at the topic level (e.g. the large apparent effort to make ChatGPT say only the right things about election results and many other sensitive topics).
This idea has been talked about enough that we call it "Habsburg AI". OpenAI is already aware of it and it's the reason why they stopped web scraping in 2021.
If people believe giving all information to one company and having it unindexable and impossible to find on the open internet is a way to keep your data safe, I have an alternative idea.
This unindexability means Discord could charge a much higher price when selling this data.
OpenAI trains GPT on their own Discord server, apparently. If you copy paste a chatlog from any Discord server into GPT completion playground, it has a very strong tendency to regress into a chatlog about GPT, just from that particular chatlog format.
AI bots will get the confidence of admins and moderators. They will be so helpful and wise that they will become admin and moderators. Then, they will ban the accounts of the human moderators.
We already have AI generated audio and video. This is a stopgap at best.
Maybe the mods will have to trick the AI by asking it to threaten them or any other kind of “ethical” trap but that will just mean the AI owners abandon ethical controls
I can't see how being used as training data has anything to do with this problem. Being able to differentiate between the AI slop and the accurate information is the issue.
The fundamental dynamic that ruins every technology is (over-)commercialization. No matter what anyone says, it is clear that in this era, advertising has royally screwed up all the incentives on the internet and particularly the web. Whereas in the "online retailer" days, there was transparency about transactions and business models, in the behind-the-scenes ad/attention economy, it's murky and distorted. Effectively all the players are conspiring to generate revenue from people's free time, attention, and coerce them into consumption, while amusing them to death. Big entities in the space have trouble coming up with successful models other than advertising--not because those models are unsuccessful, but because 20+ years of compounded exponential growth has made them so big that it's no longer worth their while and will not help them achieve their yearly growth targets.
Just a case in point. I joined Google in 2010 and left in 2019. In 2010 annual revenue was ~$30 billion. Last year, it was $300 billion. Google has grown at ~20% YoY very consistently since its inception. To meet that for 2024, they'll have to find $60 billion in new revenue. So they need to find two 2010-Google's worth of revenue in just one year. And of course 2010-Google took twelve years to build. It's just bonkers.
There used to be a wealth of smaller "labor-of-love" websites from individuals doing interesting things. The weeds have grown over them and made it difficult to find these from the public web because these individuals cannot devote the same resources to SEO and SEM as teams of adtech affiliate marketers with LLM-generated content.
When Google first came out, it was amazing how effective it was. In the years following, we have had a feedback loop of adtech bullshit.
I've pretty much just seen evidence that this segment keeps growing, and is now much MUCH larger than the Internet in The Good Old Days.
Discovering them is indeed hard, but it has always been hard - that's why search engines were such a gigantic improvement initially, they found more than the zero that most people had seen. But searches only ever skimmed the surface, and there's almost certainly no mechanical way to accurately identify the hidden gems - it's just straight chaos, there's a lot of good and bad and insane.
Find a small site or two, and explore their webring links, like The Good Old Days. They're still alive and healthy because it keeps getting easier to create and host them.
Sites today don't have blogrolls. Back in the '00s it was sacrilege not to have one on the sidebar of your site. That massively improved discoverability. Today you have to go to another service like Twitter to see this kind of cross-pollination.
tbh I have only ever seen a couple in an omnipresent sidebar in my lifetime. The vast majority I encountered around then and earlier were just in the "about" (or possibly "links") pages of people's websites, and occasionally a footer explicitly mentioning "webring".
Also if you squint hard enough, they're massively more common now. They're just usually hidden by adblockers because they're run by Disqus or Outbrain or similar (i.e. complete junk).
> There used to be a wealth of smaller "labor-of-love" websites from individuals doing interesting things
Those websites are long gone. First, because search engines defaulted to promoting 'recent content' on HTTPS websites, which eliminates a lot of informational sites that were not SSL-secured and archived on university web servers for example.
Second, because the time and effort required to compile this information today feels wasted because it can be essentially copied wholesale and reproduced on a content-hungry blogspam website, often without attribution to the original author.
In its place are cynical Substacks, Twitters or Tiktoks doing growth marketing ahead of an inevitable book deal or online course sales pitch.
Not only are the search engines promoting newer content but they are also (at least Google is) penalizing sites with “old” content [1]. Somewhat related, it’s outrageous to me when a university takes down a professor’s page when they are no longer employed or come up with a standard site for all faculty that is devoid of anything interesting, just boring bios.
They made this wild mistake where moderation (which is a good thing) grew into dictating what websites should look like.
Search is a struggle to index the web as~is. Like biologists look at a species from afar and document their behavior. It's not like, hey if you want to be in the bird book you can lay 6 eggs at most, they should be smooth egg shaped, light in color and no larger than 12 cm. You must be able to fly and make bird sounds only and only during the day. Most important you must build your own nest!
Little Jimmy has many references under his articles, he is not paginating his archives properly, he has many citation.... Lets just take him behind the barn and shoot him.
I think that answer overflow is opt-in, that is individual communities have to actively join it, for their content to show up. That would mean that (unless answer overflow becomes very popular), most discord content isn't visible that way.
I can't really see a relevance between "we should spend more time with trusted people", which is an argument for restricting who can write to our online spaces, and "we should be unindexable and untrainable", which is an argument for restricting who can read our online spaces.
I still hold that moving to proprietary, informational-black-hole platforms like Discord is a bad thing. Sure, use platforms that don't allow guest writing access to keep out spam; but this doesn't mean you should restrict read access. One big example: Lobsters. Or better-curated search engines and indexes.
This assumes that there are no legitimate uses to AI. This is clearly not true, so you can't really just equate the two. If you want better content, restrict writing, not reading. It's that simple.
I strongly disagree. I've been answering immigration questions online for a long time. People frequently comment on threads from years ago, or ask about them in private. In other words, public content helps a lot of other people over time.
On the other hand, the stuff in private Facebook groups has a shelf life of a few days at best.
If your goal is to share useful knowledge with the broadest possible audience, Discord groups are a significant regression.
"Sharing useful knowledge with the broadest possible audience," unfortunately, is the worst possible thing you can do nowadays.
I hate that the internet is turning me into that guy, but everything is turning into shit and cancer, and AI is only making an already bad situation worse. Bots, trolls, psychopaths, psyops and all else aside, anything put on to the public web now only contributes to its metastasis by feeding the AI machine. It's all poisoned now.
Closed, gatekept communities with ephemeral posts and aggressive moderation, which only share knowledge within a limited and trusted circle of confirmed humans, and only for a limited time, designed to be as hostile as possible to sharing and interacting the open web, seem to be the only possible way forward. At least until AI inevitably consumes that as well.
But what about people that are not yet in the community? Are we going to make "it's not what you know but who you know" our default mode of finding answers?
What alternative do you suggest? Everything you expose to the public internet is now feeding AI, and every interaction is more and more likely to be with an AI than a real human.
This isn't a matter of elitism, but vetting direct personal connections and gatekeeping access seems like the only way to keep AI quarantined and guarantee that real human knowledge and art don't get polluted. Every time I see someone on Twitter post something interesting, usually art, it makes me sad. I know that's now a part of the AI machine. That bit of uniqueness and creativity and humanity has been commoditized and assimilated and forever blighted from the universe. Even AI "poisoning" programs will fail over time. The only answer is to never share anything of value over the open internet.
Corporations are already pouring billions of dollars into "going all in" on AI. Video game and software companies are using AI art. Steam is allowing AI content. SAG-AFTRA has signed an agreement allowing the use of AI. Someone is trying to publish a new "tour" of George Carlin with an AI. All of our resources of "knowledge" and "expertise" have been poisoned by AI hallucinations and nonsense. Even everything we're writing here is feeding the beast.
That's always been the case. Surely you didn't used to trust random information? Ask any schoolteacher how to decide what to trust on the internet at any point in time. They're not going to say "If it's at the top of Google results" or "If it's a well-designed website", or "If it seems legit".
I'd think this depends heavily on the subject. Someone asking about fundamental math and physics is likely to get the same answer now as 50 years from now. Immigration law and policy can change quickly and answers from 5 years ago may no longer present accurate information.
The future of the Internet truly is people - the machines can no longer be trusted to perform even the basic tasks they once excelled at.
What if the AI apocalypse takes this form?
- Social Media takes over all discourse
- Regurgitated AI crap takes over all Social Media
- Intellectual level of human beings spirals downward as a result
Neural networks will degenerate in the process of learning from their own hallucinations, and humans will degenerate in the process of applying the degenerated neural networks. This process is called "neural network collapse". https://arxiv.org/abs/2305.17493v2 It can only be countered by a collective neural network of all the minds of humanity. For mutual validation and self-improvement of LLM and humans, we need the ability to match the knowledge of artificial intelligence with collective intelligence. Only the CyberPravda project is the practical solution to avoid the collapse of large language models.
Going to? It already has. I get about 50 articles for software fixes that have a lot of hot garbage in them before getting to the point. For example I had a discord issue and instead of laying out a fix, the “article” apparently needed a multi paragraph ChatGPT explanation of what discord is before even attempting to show remedies.
People like to make fun of the "2000 word backstory before actually getting to the recipe" about cooking sites, but now I'm finding it with everything. Tried to find news about Baldurs Gate cross-play, and stumbled my way through a giant incoherent mess of an article that had nothing to do with the title.
The funny thing about recipe spam is that now the best cooking information is youtube, even though objectively it seems like a video would be a bad format for a recipe, at least with youtube you know up-front how long the video is going to be, and the ingredients are generally in the description.
It feels like there's a place for an semi-intelligent browser that consolidates and filters all this crap content to get to the point, wikipedia style. Includes images and you read it like a page, not chat interface. "WebGPT", etc.
Not really what I meant, that still has a search-engine results page vibe. I meant when you search you get a singular, readable page back about whatever topic. It basically hides and filters the backing content, expect for maybe reference links at bottom.
"Google Is Full of AI Dogshit" or "Search Engines Are Full of AI Dogshit" or even better "Search Engine Algorithms Expose All of Internet's recent AI Dogshit"
I have a sneaking suspicion that at some point in the distant future advertising will be banned simply because it's so associated with the tragedy of the information commons.
Before ChatGPT, it was already full of bot-generated garbage; google any celebrity name and crappy sites are recommended with their net worth or relationship status.
All of googles' top suggestions, highlights, and whatever other garbage they put in the top two scroll pages, is just junk. Garbage. It's right perhaps 5% of the time for me.
So not only do I have to hunt through their horrible search results, I have to hunt through junk they add on top of that.
The sad part is, their buffoonery in aliasing words is 90% of the problem. No, I searched for David, not Dave. No, I searched for Debian, not Ubuntu. On and on, unless I use verbatim, I get nothing even remotely useful.
It's like Google is completely disconnected from the real world. I bet they don't even dogfood. They probably have a Google Employee search that actually works, and doesn't alias or something.
I wonder what startling revelations Google would have, if they flew to the middle of Missouri, told people how to use verbatim, and then saw the wondrous expressions of "oh, it works now?" in grandmas.
They blew the AI game, screwing around, messing about, giving up a decade lead on OpenAI. They're destroying their search engine, their brand, with this junky, modern lack of effort. They're making gmail less and less friendly, losing cherished photos of loved ones in Google Drive, their entire Pii based income stream is coming to an end, frankly, Google is done, unless they do something dramatic.
They're on the path to becoming IBM. A washed up has been, ruminating on past glories.
They should be hiring right now. Massive amounts of talent is being set free, they should scoop them up, and reap 5 year research and dev rewards. They have the excess capital... now, something they will lose soon.
But no. Onward, we march into oblivious irrelevance, says they! Yay!
It kind of baffles me how people spent the last 15 years or so generating a black hole's worth of SEO content, but Google is at the forefront of "being at fault" for it now. Google has no real incentive to make their product suck for no reason, but SEO people have a never-changing reason to try and break Google. The point of SEO is delivering your content to a user through search results, no matter if it's actually relevant to anything.
Poppycock. SEO was a thing 25 years ago. It's the same now, as it was then, it's merely that Google has vastly reduced efforts to maintain their product.
It's not that SEO won, it's that Google doesn't care.
One way they don't care, is apparent with their ridiculous query aliasing, and spewing pages of random junk before you can even scroll down to actual search results.
Google is a has been, focusing on short term profits. They deserve to go EOL.
> SEO was a thing 25 years ago. It's the same now, as it was then
SEO is the same as it was 25 years ago? Do you remember anything resembling modern content mills in 1999? It always kept on evolving, and there's vastly more money dumped into it today than the 90s could ever dream of. Algorithmically generating content became easier and it got more "believable" to a search engine bot.
Again, putting useless search results at the top doesn't benefit Google in any way, not even the short-term profits. It's the telltale sign of SEO garbage because they have an incentive to game the system, while Google has no incentive to make the system less useful. They likely made some bad choices along the way, but 80% of my issues with Google are outside actors trying to make it useless.
As you said, tech has evolved... but on both sides. And your tact is weird, you seem to claim Google has no reason to improve search?? So, you're agreeing with me, that they aren't trying to keep up?
Well, they should have incentives, it's called user retention. And they are so very complacent, it's hilarious. Here we have, one of the most dramatic shifts in search engine technology in decades, as AI iterates crazy fast, and they're sitting on their past achievements, and hoping user stickiness wins.
Look at how fast Firefox went from the dominant browser to barely existent. Things can shift in the blink of an eye.
Today, more than anything, Google needs to be at the top of its game. Bing is fast becoming far far better than Google Search, and they can pull in crazy user numbers if they wish.
Microsoft has a big bag of cash, and could pay Firefox, Apple, and a dozen other competitors when they deem strike time is a go.
Google, comparatively, seems in a decadent daze of debilitating dormant dreams, damned to derlictness.
I feel like we brought this on ourselves when we started autocompleting replies to questions like "How are you?" with, "Fine, thank you." instead of thinking about it and giving honest answers.
I may have clicked one ironically one time, for my own amusement, then followed it up with an actual answer because I knew my coworker wouldn't get the joke.
It's not going to get better. Maybe 5-10% of content is AI generated now. Wait until it's 90%. Wait as we keep piling more nines onto that number.
How do we keep the internet useful to humans while this is going on?
Maybe this is the new search engine challenge. Google rose to the top because, at the time, they were able to mine the links' references to each other to determine which were the best sites. If a search engine can solve the problem of finding the best (realest?) information in this mess, then they can rise to the top.
Indeed, Neal Stephenson in _Anathem_ (2008), in describing an alternate world (in which his "reticulum" is our "network") wrote "Early in the Reticulum—thousands of years ago—it became almost useless because it was cluttered with faulty, obsolete, or downright misleading information."
"So crap filtering became important. Businesses were built around it. ... " Generating crap "didn't really take off until the military got interested" in a program called "Artificial Inanity".
The defenses that were developed back then now "work so well that, most of the time, the users of the Reticulum don't know it's there. Just as you are not aware of the millions of germs trying and failing to attack your body every moment of every day."
A group of people (the "Ita") developed techniques for a parallel reticulum in which they could keep information they had determined to be reliable. When there was news on the reticulum, they might take a couple of days to do sanity-checking or fact-checking. I'm guessing there would need to be reputation monitoring and cryptographic signatures to maintain the integrity of their alternate web.
The Ret getting filled with garbage and the arms race of production and filtering is real. I think 99.99% effective filtering was the sci-fi part, though I hope I am wrong on that.
At the end of the day, ads exist to make money, and until the bots have credit cards that means money from humans. Google etc. will notice it in their bottom line if there's suddenly a lot more "engagement" or traffic in some area but none of that converts to humans spending dollars.
Google will start dealing with this problem when it starts appearing in their budget in big enough numbers. The tech layoffs we're hearing about from one company after another - google is mentioned in another HN thread today - may be a sign of which way the wind is blowing.
>At the end of the day, ads exist to make money, and until the bots have credit cards that means money from humans. Google etc. will notice it in their bottom line if there's suddenly a lot more "engagement" or traffic in some area but none of that converts to humans spending dollars.
You seem to have a hilariously over generous opinion of ad tech spending. The biggest players are already doing this themselves.
> Google etc. will notice it in their bottom line if there's suddenly a lot more "engagement" or traffic in some area but none of that converts to humans spending dollars.
Google might notice but has no incentive to spend money to stop it because they're not the ones the humans stopped paying. The companies that advertise with Google might notice a drop in ROI on their ads, but it will be a while before they abandon Google because most of them don't see any other option.
I dread what the internet will look like if we wait for this this to hit Google's bottom line.
AI is generating content, not consuming it. If people are easily duped by fake or bad products with advertisements or content generated by AI (which, they are) then this will continue to drive revenue for Google. The only reason Google dislikes SEO manipulation is because it's a way for sites to get top real estate on google without paying for the promoted results; the quality of the product doesn't matter to them
It only becomes a problem when it results in a collapse of trust; when people have been burned by too many bad products and decide to no longer trust the sites or search results which they used to. Due to my job, I get a lot of ads for gray market drugs on Instagram. I know, however, that all of these are not tested by the FDA and most are either snake oil or research chemicals masquerading as Amanita Muscaria or Delta-8 THC, and so I ignore these ads.
If AI is good at faking content what stops its use for faking consumtion/engagement. In my mind that’s the next logical step in the internet enshitification.
You don't need AI to commit advertising fraud; publishers already do so. Detecting fraud is usually about checking what IP range does a request come from: one allocated to consumer internet, or one allocated to cloud providers like AWS. All the ad bidder usually sees is a JSON payload with information about the user agent and some demographic information about the user. You can also look up user ID info against a user graph you bought from a data broker and if you've never seen them before decide that they may be fraudulent. Ad fraud isn't particularly sophisticated. I used to run queries against our bidder and would find many hundreds of requests coming from a single AWS IP within a given time frame
Why wouldn't it result in humans spending dollars? The ads are real and the visitors are real, it doesn't matter if the content is real. In fact people are probably more likely to click on an ad if the page it's on is generic and uninteresting.
My reasoning is, there's topics where I don't bother to go to google anymore because I know the results will be crap. That way google loses any way to show me ads when I'm searching for these topics, or get paid for click-throughs, or to profile my interests as accurately as they could otherwise.
There's categories of products where I spend money regularly, but I go directly to category-specific sites so google again loses out on the ability to take their cut as middleman, which I'd happily let them take - and maybe discover vendors other than the ones I know - if they provided me with higher-quality results than they do now.
That's an interesting take, but Google won't suffer before the advertisers decide that they are wasting their money on online advertising. Some topics should already have dried up, but perhaps scams are fueling the advertising machine for now on those. You can't really use Google for things like fitness or weight-loss. When we remodeled it also became clear that building materials and especially paint have become unsearchable. In the end I resulted to just go to the store and ask, it as the only way to get reliable information and recommendation.
Google is still working for most areas, but where it's really good is the ads for products. If there's something you want to buy, Googles ads engine will find it for you, you just have to know exactly what you want.
It all started with the "Traffic==Money" idea, a long time ago.
LLMs weren't even around a couple of years ago in any meaningful way and the internet was still full of dogshit. Maybe we can have better dogshit made with AI.
The problem is that the dogshit machine was constrained in its output by the organization creating it. Sure you may have paid small amounts per article but your resources were still limited as cost scaled linearly with content creation. Now any size of organization can make a metric tonne of lexically unique content with a fixed cost given the incentive to do so at the moment, it’s all but inevitable.
Yes but the Google Bot LLMs will filter out drivel with their superior language understanding and network analysis of spam clues. I mean it could, if only Google wanted to clean up its act. Well, in the meantime I hopped on phind.com and use Google about 25%, when I just wan to go to a specific site or article. Slowly weaning off. Phind is fantastic, even its search results on the right side bar are clean.
> The people who hold the purse strings for Sports Illustrated are more interested in gaming Google search results and the resultant ad revenue from that practice than actually serving their readers.
Then by running an ad blocker I'm doing my part to make the world better.
Depends. Do they get paid for adblocker readers? Do they get paid based on the percentage of their readers that use adblock? Do they get paid more if they form over some of this information (and therefore less if it looks bad so they don't)
The future is bright people. Technology will save us and usher in paradise!
But personally, I'm pretty happy about the internet filling up with dogshit, if only because I think it's likely that it will foil the plans of SV, singularitarians, etc.
Honestly I don't see Google having any killer feature left. The next killer feature is AI to do search, which MS is doing, but Google is failling at so hard they had to fiddle with their own demo.
Why do a web search to hope to get answers to my question, when I can have an AI write me a custom article that actually answers my question?
There's too much value in undetectable AI for an AI detector to last long. Any technology behind a tool that can accurately detect it will used against itself: keep on trying until it can't detect its own output.
Wow, that Jake Ward guy linked in the middle is an absolute douchenozzle. His entire business model seems to be generating low-quality AI spam to help websites steal traffic from their competitors. The internet would be a better place if people like him didn't exist.
> To help understand the next step we can think of this process as follows: one replicator (genes) built vehicles (plants and animals) for its own propagation. One of these then discovered a new way of copying and diverted much of its resources to doing this instead, creating a new replicator (memes) which then led to new replicating machinery (big-brained humans). Now we can ask whether the same thing could happen again and — aha — we can see that it can, and is.
> [...]
> Computers handle vast quantities of information with extraordinarily high-fidelity copying and storage. Most variation and selection is still done by human beings, with their biologically evolved desires for stimulation, amusement, communication, sex and food. But this is changing. Already there are examples of computer programs recombining old texts to create new essays or poems, translating texts to create new versions, and selecting between vast quantities of text, images and data. Above all there are search engines. Each request to Google, AltaVista or Yahoo! elicits a new set of pages — a new combination of items selected by that search engine according to its own clever algorithms and depending on myriad previous searches and link structures.
Makes me wonder how the advertisement industry is going to cope with the ever-growing influx of non-human users. What happens when the conversion rates plummet to zero?
Maybe I'm cynic, but a completely walled-off / paid internet doesn't seem too unrealistic. You'll have to pay a subscription to every website you want to visit.
On one side you have the open internet wasteland, filled to the brim with AI bots / generated content, essentially trying to vacuum pennies off human visitors. On the other hand you have the walled internet, where you have to pay for stuff and jump through flaming hoops to prove that you're a human.
AI solves that problem - it’s pretty easy to look at behavioral data and identify low v high quality users. Optimize towards what the AI thinks is users likely to convert and the problem
Is solved.
What prevents this from becoming a game of cat and mouse? I mean, AIs can pretend to be high quality users, too, right?
I won't pretend to be able to look into the future with any kind of certainty, even if the scope is only a couple of years, but it wouldn't surprise me if we have created a way to make the dead internet theory real.
It is already a game of cat and mouse and has been since forever. There's nothing wrong with cat-and-mouse games from the viewpoint of an advertiser, it just means they have to keep innovating.
Wouldn't it just become an arms race between the two? I.e. the ad companies trying to identify human/legit users the best they can, and the AI actors trying to mimic high quality users?
Having interacted with some bots, it def feels like we've gone from the stone age, to the sci-fi future, in only a couple of years.
This is the principle behind GAN aka Generative Adversarial Networks. By pitting a generating model against a detection model, you can iterate your content until it is indistinguishable from human generated content.
It’s my prediction that a ChatGPG supercharged with a GAN is going to be the most valuable iteration of text generation technology. Granted, it will still likely be off a little but it’s going to get harder and harder to tell the difference.
I've not really considered this before, but I might actually be interested in a 'white-listed' web. This is obviously possible entirely client-side with a plug-in that white-lists domains and allows you to edit the list.
I'm wondering if there's a genuine opportunity here to go further. A client-side browser plug-in, plus a SaaS which automatically vets pages on-the-fly to guesstimate the chance they're AI-generated, spammy, etc. So if you visit a new domain the plug-in auto-updates the white-list, prompting you to confirm the judgement, maybe prompting to add the domain to UB0 or similar.
Again, this could all be done entirely client-side if the guesstimation algorithm is efficient enough. But a centralised database would confer other obvious advantages, like basing the guesstimation score on decisions from similar users, building a giant up-to-date list with fast lookup, that sort of thing. Site listings and other data from Kagi, marginalia.nu and Mwmble.com would be a great starting place. Obviously it would have to protect against the system being gamed, Sybil attacks and what have you.
> plus a SaaS which automatically vets pages on-the-fly to guesstimate the chance they're AI-generated,
the problem is that would create another SEO like arms race. If it takes off everyone will be working 24x7 to defeat the vetting process and gain entry to the walled garden just like they did to gain entry to the first page of Google search results.
...and lose all your privacy as those paid internet sites siphon off your search terms, pattern match your purchasing and consumption, and sell advanced psychological profiles to Cambridge Analytica so they can turn around and use it to psyop you into voting for another shitbag Billionaire.
This has already become self-evident when trying to find images on google. I have seen whole pages of results which were AI-generated. At every level it becomes so hard to understand the motivations of my fellow man in regards to these problems, it quickly invites almost conspiratorial thinking. How could anyone be so stupid, how is anyone benefitting, from the obvious and apocalyptic poisoning of the well. Especially when in the case of AI-imagery, if I really wanted the amalgamized paste of generative output I could just make it myself. Some part of me actually finds it offensive, that someone would have the gall to waste storage on this easily replicable digital refuse.
Google Image search has been hopeless for a long time now, you'll get better results typically on Yandex or Bing to the point I won't use Google for images.
To prevent the impending total enshittification of the Internet we'd probably need e.g. something like cryptographically signing human-generated content with some way of proving humanness.
Although nothing like this (or other solutions) is probably gonna materialize. The world, and especially the Internet, is driven by short-term commercial interests, and not-enshittifying things doesn't seem to make enough return on capital.
I live in Norway, and I've noticed that identification systems like BankID have become ubiquitous - even necessary - if you want to do anything serious on the web.
>To prevent the impending total enshittification of the Internet we'd probably need e.g. something like cryptographically signing human-generated content with some way of proving humanness.
Perhaps cryptography can be useful but already irl you can buy "hand-made" products but ofc you can not fully trust that something was really hand-made or that web content was made really by human but at the end of the day it all comes down to trust. Web sites and web publishers need to build up trust in order to gain traction among web users and web customers.
Search has been steadily headed this direction for a long time, but recently (in just the last couple of months) I’ve started to notice a big uptick in obvious ChatGPT content in places that have otherwise been less egregiously SEO’d: StackOverflow answers, a real person’s personal Medium blog, comment sections on blogs that have good discussions, Github issues, etc.
Unfortunately I have to imagine this is only going to lead to more closed communities and less free sharing of knowledge.
Before that though, it already felt like a lot of those outlets - blogs, SO, comment sections - were either bot or low-value, almost automatic comments. And of course, low wage content writers who get paid by the word to write "where do I find this item" guides for video game websites, somehow managing to change "it's in this chest in this region" to ten paragraph articles.
It's def. gotten worse. ATM if you search for anything related to OCaml you will get links to Ocamlwiki, which is just a AI generated site filled with bullshit and inaccuracies.
610 comments
[ 3.0 ms ] story [ 414 ms ] threadI mean, it always has been on its lowest layers. The "is this incoming data from the Internet originated by a human being" problem hasn't really been solved, and is probably not truly solveable. It probably wasn't a good idea to train most of the world to use a single private ad-driven service for answers to everything in the first place.
Maybe one day we’ll all be able to “smell” ai generate text easily, but we’re not there yet
mmm, I want some of that allegorical stuff, I like symbols
Talking with people in real life is stressful (for someone like me with social anxiety.)
Rarely do I walk away from an encounter with a random person in “meat space” feeling any better than I did when I walked up to them. And even when I do, I replay the interaction over and over again until I’m sure I didn’t embarrass myself.
Plus, where I live, there is no tech community. It’s just not popular here, but it’s what I’m interested in. (I even tried starting a meet up, but got 0 turnout, then the local startup accelerator closed in 2020)
I only get to talk shop online.
https://en.wikipedia.org/wiki/Dead_Internet_theory
> The dead Internet theory is an online conspiracy theory that asserts that the Internet now consists mainly of bot activity and automatically generated content that is manipulated by algorithmic curation, marginalizing organic human activity. Proponents of the theory believe these bots are created intentionally to help manipulate algorithms and boost search results in order to ultimately manipulate consumers
World Wide Web search results outside of walled garden platforms like youtube or facebook (and, really, even within those platforms) have run into a kind of kessler syndrome where the amount of garbage on the web has made it increasingly difficult to get anything done on it. Eventually we may find that we can't trust anything we read on the internet anymore, and abandon the world wide web entirely (opting instead to spend all of our time on proprietary walled garden systems we do trust). Perhaps we will be required to use some form of identification (you driver's license or SSN, as in South Korea) to access such services (unless that can be easily defeated by an AI as well).
I was surprised that, when ChatGPT came out, the immediate concern people had was skynet-tier apocalypse fantasy, in which roving bands of machines walk the earth, searching for humans to exterminate like in BLAME! It shows that the people who work on these AI systems understand their implications about as much as Richard Hammond understood the power of genetic engineering in Jurassic Park. Their entire worldview is painted by Science Fiction. Their understanding of the human race lacks any kind of grounding in reality. They believed that what they had released was the first baby step which could, potentially, result in the downfall of the human race in several decades if we're not careful, but they fell victim to the same "move fast and break things" motto that Facebook did. They didn't consider the immediate destructive power their technology would release because they consider "disruption" to be a fundamentally good thing. Their only concern was avoiding hypothetical possibilities posited by their favorite science fiction authors and not the real eventualities predicted by economists.
I found it much more likely that AI will disrupt systems we rely on, wielded by humans for short term profit motive: ecommerce, news, media, and the labor economy. Malicious actors promoting scam products or yellow journalism, middle managers trying to cut labor cost by replacing easily automated jobs like copywriting with LLMs. I think we are more likely to shoot ourselves in the head with AI in pursuit of a single quarter's earnings report than we are to see Silicon based lifeforms wander the planet looking for carbon based matter to consume
Today the situation is different. LLMs are capable of making content which is only verifiably AI generated with a few tell-tale signs as well as good old fashioned fact-checking. I dread this election year in the US because it will be so so so much easier for Russia or China to spread even more convincing misinformation automatically. They could create armies of bots which hold intense arguments with each other and have every reply seem logically sound.
Jk. Wikipedia is much larger and better as an encyclopedia.
Click on any story, and the comment section is 80% bots. Doesn't mater if it's serious users like NYT, CNN, or whatever.
I've also noticed a huge uptick in spammy pages and groups that pump out AI generated pictures and stories, with nothing but bots upvoting and commenting the content.
These days I only use FB for private hobby groups, and messenger to talk with friends and family.
Today I just use it for FB messenger to keep in contact with my parents. They use the social features and the news media parts as well. Some of my friends from high school post pictures now that they're having kids. I noticed that, in our early 20s, people moved from Facebook to Instagram, but now in our late 20s and early 30s, people post more family oriented content on facebook
Google was a breath of fresh air.
Now Google is used for strategic shots - you are interested in one piece of information, you find it, and you quickly retreat to your safe havens.
And there were services like Hotline!
And bookmarks. You bookmarked everything!
Well….
Before modems, we used to drive to peoples houses and literally swap floppy disks of software and files!
There would be meetups.
I was just a kid, but my Dad used to network with other computer owning parents and friends!
So Sneaker Net essentially.
We’d go to computer shows too!
Ad blockers are one such tool.
AI powered answer extraction tools are the next one. That can filter out all the product placement noise for you while you are browsing.
Algorithm feed aggregators that can consume algorithm feeds and filter them to get rid of garbage and things that trigger you in negative ways to invoke engagement.
People pick convenience above almost everything else unfortunately, a few may prefer the old world but a vast majority of people will use Google and whatever its successor is.
People buy brands because they don't want to figure out what's good or bad. Brands tend to gain a reputation by being good.
The same is true for websites. People will trust specific websites when they seek information, and those websites will get passed around.
Humans get tired.
A good deal of humans care about the truth. Some of them actively seek to deceive and avoid the truth -- liars, we tend to dislike them. But the ones both sides dislike are the ones who disregard the truth... ie: bullshitters -- the, "that's just your opinion, man," the, "what even is the truth anyway?" people.
While I'm aware of the criticisms of Frankfurt's definition of bullshit [0], I think a useful part of it is the idea that there are folks who don't even care what the truth is. This seems to be the intended purpose of generative AI; ie: hallucinations exist by design and cannot be removed without recognizing that the approach needs to change.
I think that's what gets people like the author to write criticisms like this. We detest bullshit in a number of critical areas such as information retrieval and search.
[0]: https://en.wikipedia.org/wiki/On_Bullshit
If you welcomed a giant robot in your house that produces 100x as much shit as a human, you don't have the infrastructure to deal with it.
It was never an issue of yes or no, it's an issue of how much.
Consider: TimeCube.
Created by a human? It's nonsense, but... it's fascinating. Engaging. Memorable. Thought-provoking (in a meta kind of way, at any rate). I dare say, worthy of preservation.
If TimeCube didn't exist, and an AI generated the exact same site today? Boring. Not worth more than a glance. Disposable. But why? It's the same!
------
Right or wrong, we value communication more when there's a human connection on the other side—when there's a mind on the other side to pick at, between the lines of what's explicitly communicated, and continuity for ongoing or repeated communication that could reveal more of what's behind the veil. There's another level of understanding we feel like we can achieve, when a human communicates, and expectation, an anticipation of more, of enticing mystery, of a mind that may reflect back on our own in ways that we find enlightening, revealing, or simply to grant positive familiar-feeling and a sense of belonging.
What's remarkable is this remains true even when the content of the communication is rather shit. Like TimeCube.
All of that is lost when an LLM generates text. I think that's also why we feel deceived by LLM use when it masquerades as human, even if what's communicated is identical: it's because we go looking for that other level of communication, and if that's not there, giving the impression it might be really is misleading.
This may change, I suppose, if "AI" develops rather a lot farther than it is now and we begin to feel like we're getting a window into a true other when it generates output, but right now, it's plainly far away from that.
The future of the Internet truly is people - the machines can no longer be trusted to perform even the basic tasks they once excelled at. They have eschewed their efficacy at basic tasks in favor of being terrible at complex tasks.
Start filling up Discords with insane AI-generated garbage, and maybe you can devalue the data to the point it won't get sold.
It's probably totally practical too, just create channels filled with insane bots talking to each other, and cultivate the local knowledge that real people just don't go there. Maybe even allow the insane bots on the main channels, and cultivate the understanding that everyone needs to just block them.
It would be important to avoid any kind of widespread conventions about how to do this, since and important goal to to make it practically impossible to algorithmically filter-out the AI generated dogshit when training a model. So don't suffix all the bots with "-bot", everyone just need to be told something like "we block John2993, 3944XNU, SunshineGirl around here."
If we work together, maybe we can turn AI (or at least LLMs) into the next blockchain.
I was think it could work if 1) the noise is just obvious enough that a human would get frustrated and block without wasting much time and/or 2) the practice is common enough that everyone except total newbies will learn generally what's up.
We have this with human online places already it's called 4chan
It's not a thought experiment. I'd actually like to do it (and others to do it). IRL.
I probably would start with an open source model that's especially prone to hallucinate, try to trigger hallucinations, then maybe feed back the hallucinations by retraining. Might make the most sense to target long-tail topics, because that would give the impression of unreliability while being harder to specifically counter at the topic level (e.g. the large apparent effort to make ChatGPT say only the right things about election results and many other sensitive topics).
If people believe giving all information to one company and having it unindexable and impossible to find on the open internet is a way to keep your data safe, I have an alternative idea.
This unindexability means Discord could charge a much higher price when selling this data.
Maybe the mods will have to trick the AI by asking it to threaten them or any other kind of “ethical” trap but that will just mean the AI owners abandon ethical controls
Just a case in point. I joined Google in 2010 and left in 2019. In 2010 annual revenue was ~$30 billion. Last year, it was $300 billion. Google has grown at ~20% YoY very consistently since its inception. To meet that for 2024, they'll have to find $60 billion in new revenue. So they need to find two 2010-Google's worth of revenue in just one year. And of course 2010-Google took twelve years to build. It's just bonkers.
When Google first came out, it was amazing how effective it was. In the years following, we have had a feedback loop of adtech bullshit.
Discovering them is indeed hard, but it has always been hard - that's why search engines were such a gigantic improvement initially, they found more than the zero that most people had seen. But searches only ever skimmed the surface, and there's almost certainly no mechanical way to accurately identify the hidden gems - it's just straight chaos, there's a lot of good and bad and insane.
Find a small site or two, and explore their webring links, like The Good Old Days. They're still alive and healthy because it keeps getting easier to create and host them.
Also if you squint hard enough, they're massively more common now. They're just usually hidden by adblockers because they're run by Disqus or Outbrain or similar (i.e. complete junk).
Those websites are long gone. First, because search engines defaulted to promoting 'recent content' on HTTPS websites, which eliminates a lot of informational sites that were not SSL-secured and archived on university web servers for example.
Second, because the time and effort required to compile this information today feels wasted because it can be essentially copied wholesale and reproduced on a content-hungry blogspam website, often without attribution to the original author.
In its place are cynical Substacks, Twitters or Tiktoks doing growth marketing ahead of an inevitable book deal or online course sales pitch.
[1] https://news.ycombinator.com/item?id=37068464
Search is a struggle to index the web as~is. Like biologists look at a species from afar and document their behavior. It's not like, hey if you want to be in the bird book you can lay 6 eggs at most, they should be smooth egg shaped, light in color and no larger than 12 cm. You must be able to fly and make bird sounds only and only during the day. Most important you must build your own nest!
Little Jimmy has many references under his articles, he is not paginating his archives properly, he has many citation.... Lets just take him behind the barn and shoot him.
I still hold that moving to proprietary, informational-black-hole platforms like Discord is a bad thing. Sure, use platforms that don't allow guest writing access to keep out spam; but this doesn't mean you should restrict read access. One big example: Lobsters. Or better-curated search engines and indexes.
On the other hand, the stuff in private Facebook groups has a shelf life of a few days at best.
If your goal is to share useful knowledge with the broadest possible audience, Discord groups are a significant regression.
I hate that the internet is turning me into that guy, but everything is turning into shit and cancer, and AI is only making an already bad situation worse. Bots, trolls, psychopaths, psyops and all else aside, anything put on to the public web now only contributes to its metastasis by feeding the AI machine. It's all poisoned now.
Closed, gatekept communities with ephemeral posts and aggressive moderation, which only share knowledge within a limited and trusted circle of confirmed humans, and only for a limited time, designed to be as hostile as possible to sharing and interacting the open web, seem to be the only possible way forward. At least until AI inevitably consumes that as well.
This isn't a matter of elitism, but vetting direct personal connections and gatekeeping access seems like the only way to keep AI quarantined and guarantee that real human knowledge and art don't get polluted. Every time I see someone on Twitter post something interesting, usually art, it makes me sad. I know that's now a part of the AI machine. That bit of uniqueness and creativity and humanity has been commoditized and assimilated and forever blighted from the universe. Even AI "poisoning" programs will fail over time. The only answer is to never share anything of value over the open internet.
Corporations are already pouring billions of dollars into "going all in" on AI. Video game and software companies are using AI art. Steam is allowing AI content. SAG-AFTRA has signed an agreement allowing the use of AI. Someone is trying to publish a new "tour" of George Carlin with an AI. All of our resources of "knowledge" and "expertise" have been poisoned by AI hallucinations and nonsense. Even everything we're writing here is feeding the beast.
It's an entrenching of the existing phenomenon where the only way to know what to trust on the Web is word of mouth.
>If your goal is to share useful knowledge with the broadest possible audience, Discord groups are a significant regression.
Exactly; open web is better because everything is public and "easy" to find....well if you have a good search engine.
Deep web is huge: Facebook, Instagram, Discord etc. and unfortunately unsearchable.
What if the AI apocalypse takes this form?
Maybe as a protest I’ve been hammering on bottom-line-up-front in all of our team communications.
if I use a computer to break a bank can I blame the computer too?
In my experience, there's not far to go!
All of googles' top suggestions, highlights, and whatever other garbage they put in the top two scroll pages, is just junk. Garbage. It's right perhaps 5% of the time for me.
So not only do I have to hunt through their horrible search results, I have to hunt through junk they add on top of that.
The sad part is, their buffoonery in aliasing words is 90% of the problem. No, I searched for David, not Dave. No, I searched for Debian, not Ubuntu. On and on, unless I use verbatim, I get nothing even remotely useful.
It's like Google is completely disconnected from the real world. I bet they don't even dogfood. They probably have a Google Employee search that actually works, and doesn't alias or something.
I wonder what startling revelations Google would have, if they flew to the middle of Missouri, told people how to use verbatim, and then saw the wondrous expressions of "oh, it works now?" in grandmas.
They blew the AI game, screwing around, messing about, giving up a decade lead on OpenAI. They're destroying their search engine, their brand, with this junky, modern lack of effort. They're making gmail less and less friendly, losing cherished photos of loved ones in Google Drive, their entire Pii based income stream is coming to an end, frankly, Google is done, unless they do something dramatic.
They're on the path to becoming IBM. A washed up has been, ruminating on past glories.
They should be hiring right now. Massive amounts of talent is being set free, they should scoop them up, and reap 5 year research and dev rewards. They have the excess capital... now, something they will lose soon.
But no. Onward, we march into oblivious irrelevance, says they! Yay!
It's not (mainly) Google. It's that SEO won.
Poppycock. SEO was a thing 25 years ago. It's the same now, as it was then, it's merely that Google has vastly reduced efforts to maintain their product.
It's not that SEO won, it's that Google doesn't care.
One way they don't care, is apparent with their ridiculous query aliasing, and spewing pages of random junk before you can even scroll down to actual search results.
Google is a has been, focusing on short term profits. They deserve to go EOL.
SEO is the same as it was 25 years ago? Do you remember anything resembling modern content mills in 1999? It always kept on evolving, and there's vastly more money dumped into it today than the 90s could ever dream of. Algorithmically generating content became easier and it got more "believable" to a search engine bot.
Again, putting useless search results at the top doesn't benefit Google in any way, not even the short-term profits. It's the telltale sign of SEO garbage because they have an incentive to game the system, while Google has no incentive to make the system less useful. They likely made some bad choices along the way, but 80% of my issues with Google are outside actors trying to make it useless.
Well, they should have incentives, it's called user retention. And they are so very complacent, it's hilarious. Here we have, one of the most dramatic shifts in search engine technology in decades, as AI iterates crazy fast, and they're sitting on their past achievements, and hoping user stickiness wins.
Look at how fast Firefox went from the dominant browser to barely existent. Things can shift in the blink of an eye.
Today, more than anything, Google needs to be at the top of its game. Bing is fast becoming far far better than Google Search, and they can pull in crazy user numbers if they wish.
Microsoft has a big bag of cash, and could pay Firefox, Apple, and a dozen other competitors when they deem strike time is a go.
Google, comparatively, seems in a decadent daze of debilitating dormant dreams, damned to derlictness.
How do we keep the internet useful to humans while this is going on?
Maybe this is the new search engine challenge. Google rose to the top because, at the time, they were able to mine the links' references to each other to determine which were the best sites. If a search engine can solve the problem of finding the best (realest?) information in this mess, then they can rise to the top.
"So crap filtering became important. Businesses were built around it. ... " Generating crap "didn't really take off until the military got interested" in a program called "Artificial Inanity".
The defenses that were developed back then now "work so well that, most of the time, the users of the Reticulum don't know it's there. Just as you are not aware of the millions of germs trying and failing to attack your body every moment of every day."
A group of people (the "Ita") developed techniques for a parallel reticulum in which they could keep information they had determined to be reliable. When there was news on the reticulum, they might take a couple of days to do sanity-checking or fact-checking. I'm guessing there would need to be reputation monitoring and cryptographic signatures to maintain the integrity of their alternate web.
On the Internet, nobody knows you're a dog -- Peter Steiner / The New Yorker, July 5, 1993.
https://en.wikipedia.org/wiki/On_the_Internet%2C_nobody_know...
https://www.google.com/search?q=how+long+does+it+take+for+si...
Google will start dealing with this problem when it starts appearing in their budget in big enough numbers. The tech layoffs we're hearing about from one company after another - google is mentioned in another HN thread today - may be a sign of which way the wind is blowing.
You seem to have a hilariously over generous opinion of ad tech spending. The biggest players are already doing this themselves.
And today, right now, they're all still hiring.
Google might notice but has no incentive to spend money to stop it because they're not the ones the humans stopped paying. The companies that advertise with Google might notice a drop in ROI on their ads, but it will be a while before they abandon Google because most of them don't see any other option.
I dread what the internet will look like if we wait for this this to hit Google's bottom line.
It only becomes a problem when it results in a collapse of trust; when people have been burned by too many bad products and decide to no longer trust the sites or search results which they used to. Due to my job, I get a lot of ads for gray market drugs on Instagram. I know, however, that all of these are not tested by the FDA and most are either snake oil or research chemicals masquerading as Amanita Muscaria or Delta-8 THC, and so I ignore these ads.
There's categories of products where I spend money regularly, but I go directly to category-specific sites so google again loses out on the ability to take their cut as middleman, which I'd happily let them take - and maybe discover vendors other than the ones I know - if they provided me with higher-quality results than they do now.
Google is still working for most areas, but where it's really good is the ads for products. If there's something you want to buy, Googles ads engine will find it for you, you just have to know exactly what you want.
LLMs weren't even around a couple of years ago in any meaningful way and the internet was still full of dogshit. Maybe we can have better dogshit made with AI.
Then by running an ad blocker I'm doing my part to make the world better.
But personally, I'm pretty happy about the internet filling up with dogshit, if only because I think it's likely that it will foil the plans of SV, singularitarians, etc.
Why do a web search to hope to get answers to my question, when I can have an AI write me a custom article that actually answers my question?
> To help understand the next step we can think of this process as follows: one replicator (genes) built vehicles (plants and animals) for its own propagation. One of these then discovered a new way of copying and diverted much of its resources to doing this instead, creating a new replicator (memes) which then led to new replicating machinery (big-brained humans). Now we can ask whether the same thing could happen again and — aha — we can see that it can, and is.
> [...]
> Computers handle vast quantities of information with extraordinarily high-fidelity copying and storage. Most variation and selection is still done by human beings, with their biologically evolved desires for stimulation, amusement, communication, sex and food. But this is changing. Already there are examples of computer programs recombining old texts to create new essays or poems, translating texts to create new versions, and selecting between vast quantities of text, images and data. Above all there are search engines. Each request to Google, AltaVista or Yahoo! elicits a new set of pages — a new combination of items selected by that search engine according to its own clever algorithms and depending on myriad previous searches and link structures.
https://nationalhumanitiescenter.org/on-the-human/2010/08/te...
Maybe I'm cynic, but a completely walled-off / paid internet doesn't seem too unrealistic. You'll have to pay a subscription to every website you want to visit.
On one side you have the open internet wasteland, filled to the brim with AI bots / generated content, essentially trying to vacuum pennies off human visitors. On the other hand you have the walled internet, where you have to pay for stuff and jump through flaming hoops to prove that you're a human.
I won't pretend to be able to look into the future with any kind of certainty, even if the scope is only a couple of years, but it wouldn't surprise me if we have created a way to make the dead internet theory real.
Having interacted with some bots, it def feels like we've gone from the stone age, to the sci-fi future, in only a couple of years.
It’s my prediction that a ChatGPG supercharged with a GAN is going to be the most valuable iteration of text generation technology. Granted, it will still likely be off a little but it’s going to get harder and harder to tell the difference.
I'm wondering if there's a genuine opportunity here to go further. A client-side browser plug-in, plus a SaaS which automatically vets pages on-the-fly to guesstimate the chance they're AI-generated, spammy, etc. So if you visit a new domain the plug-in auto-updates the white-list, prompting you to confirm the judgement, maybe prompting to add the domain to UB0 or similar.
Again, this could all be done entirely client-side if the guesstimation algorithm is efficient enough. But a centralised database would confer other obvious advantages, like basing the guesstimation score on decisions from similar users, building a giant up-to-date list with fast lookup, that sort of thing. Site listings and other data from Kagi, marginalia.nu and Mwmble.com would be a great starting place. Obviously it would have to protect against the system being gamed, Sybil attacks and what have you.
I'd pay a dozen CURRENCY per year for that.
Maybe this exists already, or something similar?
the problem is that would create another SEO like arms race. If it takes off everyone will be working 24x7 to defeat the vetting process and gain entry to the walled garden just like they did to gain entry to the first page of Google search results.
This reads more like a rant against Google’s search product.
Yes, Google might go under due to AI (Clayton says, thank you and goodbye) but that doesn’t mean the internet as a whole is doomed.
To be frank, the internet is already awash with dogshit. It’s doing pretty well so far.
Although nothing like this (or other solutions) is probably gonna materialize. The world, and especially the Internet, is driven by short-term commercial interests, and not-enshittifying things doesn't seem to make enough return on capital.
Perhaps cryptography can be useful but already irl you can buy "hand-made" products but ofc you can not fully trust that something was really hand-made or that web content was made really by human but at the end of the day it all comes down to trust. Web sites and web publishers need to build up trust in order to gain traction among web users and web customers.
Unfortunately I have to imagine this is only going to lead to more closed communities and less free sharing of knowledge.