The Dead Internet Theory [video]
youtube.com
youtube.com
Anyway the summary of this link to save anyone clicking through (copied from the youtube summary):
The internet is ever changing, but over the past couple of years things have started to change. The proportion of bots in increasing. A theory dubbed “The Dead Internet Theory” that states that most of the internet is in fact AI bots. These bots have been unleashed to control the populous and rake in profits. But is it true? And if so, how much of it is true?
The only small correction would be AI is not required. What was required to overwhelm everything is Cost of Broadcast dropping to 0. Spammers and scammers can mass bombard billions 24*7 cause there is no big cost per message sent. The cost is borne by the receiver to wade through the ever growing sewage placed in front of them. As a result less than 0.5% of content produced is consumed as mentioned in the report. Its probably similar to HN but HN is not going to measure it or inform you about it.
Just as in real life, there needs to be the possibility to reliably identify yourself easily to others and establish trusted channels of communication. Not just in a P2P manner, but in the sense of a town meeting.
Your attention is worth real money, it is billable working hours. Anonymous actors with ill intent stealing it is a crime without prosecution. That should change.
Think of how many links are not being clicked or read by anyone.
HN can show us the number everyday. If they say for every 100 links submitted only 1 is getting read. And month by month that 100 is growing to 1000 and 10000 why should/will ppl submit links? HN looks alive. But right now who knows what the number is so the submissions and traffic keep flowing and growing.
Scaling it up, when we talk about a dead networks, info exploding for no great reason, is something we need to track and think about more.
People are anonymous, they post anonymously and comment as such. Accordingly, incentive to contribute substantially is minimal at best. Look at your own comment: you didn't really read what I wrote and made some cheap whataboutism-statement.
Useless information rotting away in the void until final deletion is no problem. Incentive to spam others without blowback is.
But yeah, editing those summaries is both huge workload and easily very contentious.
Google is quite good at detecting duplicate text content, apparently duplicate video content isn't so much a thing.
If we accept the premise of the Dead Internet as true, this is a systemic issue which needs to be dealt with on a systems-level (i.e. economic, social and political), and not by the very architects of this flawed system adding their own little patches —which will just result in an AI arms-race— nor by improving the "media literacy" of the World population so they know how to deal with this broken system —a new arms-race.
Avoiding the consequences of the Dead Internet should not rely on saving the current social-media-attention-economy-internet from AI-bot infestation and expecting people to be "educated" against it. It should rely on either fully rethinking, overhauling and/or shutting down this social-media-attention-economy-internet, or alternatively on helping society to abandon-ship and let this bot-infested internet go the way of Myspace and Digg in terms of social relevance. The one thing that AI content cannot do is get in the way of society's ability to touch grass.
In order to be anything other than a big distraction, information on the internet needs to be true. Which means, it needs to be from real humans, about real life, the real world, providing real useful knowledge.
Epistemology intends to explain how that works. In particular, you need the information to take a trustworthy channel from its source to recipients. Anonymity is one principal contradiction there.
There's too much for bad actors and companies to gain, there's no way they're leaving this kind of opportunity on the table, especially as bots get much, much better at appearing human.
It is kind of funny when occasionally a bot messes up and posts the same comment more than once in a thread across more than one account.
I think this is why people increasingly prefer closed off, invite only communities as well. You can vet people coming into a discord community a lot easier and avoid a lot of the nonsense that happens on more open, public, spaces like Reddit, or Twitter.
This has nothing to do with the objective reality that in order to maintain the web a lot of bots are used and contribute to the majority of the traffic
It’s already happening.
Sure, but how long until it becomes objectively true?
I remember going out for drinks during ACM Supercomputing 2011 with a gaggle of talkative folks working for a gov't contractor on (social media) profile management and content generation (i.e. astroturf) for the military and intelligence sector, both for tradecraft and influence purposes. The sector was rolling in dough then, because throwing humans at the problem (even if it scaled) makes maintaining operational security and message discipline hard.
I'm not sure that a majority of ALL "user-generated content" on the internet that is seen by actual humans is synthetic, but I am pretty sure that it is true for topics of strategic or commercial interest to all kinds of parties. (The tech is pretty democratized now.)
Can you elaborate? Were neural language model bots used to stuff the internet with junk before the publishing of Attention Is All You Need?
But it is also important to highlight that because Internet is omniprescent most people would consume "dumb" content in the same way that happened with TV and other media in the past. It is elitist to think that Internet is only for looking at smart stuff and not basic entertainment.
But more seriously, it is important to consider just how much importance any internet discussion should be given. I would say a strong down weight in general is most useful. Consider whether it really helps matters or severely hurts progress to react to the noise. Thinking of troll farms pushing content and adding flamey content to threads. It’d be better if people had some built in skepticism.
Edit-Also, there absolutely will always be someone who disagrees with you/any comment on the internet. And they’re the most likely to respond to that comment. So while there’s a huge amount of actual information on the internet there is far, far more much noise. Actual knowledge and wisdom are going to have to be independently verifiable. Like reading a book, multiple sources, choose sources with demonstrated quality etc.
The problem with this - and it applies to BS in regular media as well - is that it doesn't matter so much what I find important, as what my fellow voters, and real policy makers, find important.
Flat earthers are harmless on their own. If flat earthers start making transport policy, things are going to stop working.
One thing to take into account is that current bots suck major balls. Specially when it comes to utterance purpose, and for longer utterances; they fail the Turing test really hard, provided that the one asking "is this a bot or a human?" is able to interpret utterances based on nearby utterances and non-textual information around it. In other words if the internet was dead already we would've noticed it.
And someone might argue "ackshyually, there are more advanced bots that model language on pragmatic levels, you just don't know about them because it's all hidden!". Well, then you get a conspiracy "theory" with its typical appeal to ignorance, not something worth thinking about.
OMG, real life is beckoning!! I'm going to have to get dressed and go outside again!