3 ms·

I'm a bit less apocalyptic, but I wonder about this, the "autophagy" TFA refers to. AI producing and consuming its own output, and the resulting "model collaps

by the_af 3y ago

I'm a bit less apocalyptic, but I wonder about this, the "autophagy" TFA refers to. AI producing and consuming its own output, and the resulting "model collapse", with human beings increasingly out of the loop in thus game of telephone. I've noticed some "summaries" of content here on HN, posted as comments, with the telltale sign of LLMs. Adding no value whatsoever. Could some day be the case that a large number of posts even here are of AI bots interacting with each other? I'm not sure about the "rebels" at the end of the article, how will they find each other in sufficient numbers, how will word of mouth spread in sufficient numbers?

btbuildem 3y ago
Same on twitter (as in the screenshot in the article) -- lots of regurgitative comments, adding words but stating nothing / rephrasing the parent tweet / etc. I think the next step is actually a big leap backwards - human-curated knowledge bases, islands in the sea of AI-generated noise.
- grugagag 3y ago
  LLM will be quick to catch onto the islands. They’d have to be private islands to fend off the corruption.
  - the_af 3y ago
    Do you mean silos that LLMs and bots cannot access? Like the walled gardens of Slack, Discord, Facebook, etc? In that case the cure will be worse than the disease, in my opinion. I don't want a siloed internet. Silos exist and are bad enough now, imagine if in the future the only good, human-produced content is secreted away within them.
  - spookie 3y ago
    Yes, they'll most likely be private islands. Honestly, it won't be too bad now that I think about it.
    - helf 3y ago
      [dead]
- toss1 3y ago
  It will be ironically funny if the original Yahoo! model of fuman-curated search results wins in the end. Also, the LLMs aren't necessarily doomed. Presumably, the humans running them are looking out for the problems of inbreeding or LLMs-eating-their-own-excrement, and will implement some combination of not feeding that crap, and/or freezing the model at a peak point where the results are still primarily based on human creations. Perhaps ChatGPT 4.3 will be the peak? EDIT: typos
surfingdino 3y ago
> I'm not sure about the "rebels" at the end of the article, how will they find each other in sufficient numbers, how will word of mouth spread in sufficient numbers? Friends of friends networks, personal recommendations sent over email, communicators, Discord.
- the_af 3y ago
  Would this scale, when it cannot outcompete garbage content now?
  - surfingdino 3y ago
    I don't know, but I think we need to give the consumers/recipients of information better tools for discovery, ideally at source instead of disseminating it through intermediaries who exist only to provide real estate for ads.
  - dylan604 3y ago
    Yeah, seems like email was already destroyed as anything useful before LLMs kicked in.
wussboy 3y ago
The only true solution to this is seldom mentioned: Real life. "Word of mouth" can still happen (and should happen) by reals words, uttered in the physical world, by real mouths, in the physical world. I suspect many of our current societal ills are caused by too much information spreading too quickly, not forced to go through a multitude of brains or stand the tests of time.
s1artibartfast 3y ago
I dont think rebel numbers is an issue. It didnt take long for Facebook to get massive penetration using other modes of communication. I think this will be a radical change for the better. Once the baseline quality and trust drops low enough, there is as an incentive for curation, walled gardens, and reputation based networks. I hope for the return of webrings, moderated forums, and quality standards for participation. I also think it may finally break the hold of social media obsession. Outrage loses much of its luster when people realize the other end of it is likely a bot. This is a step toward asking "who is this person, and why do I care what they think."
- internetter 3y ago
  This is what I was getting at, but you formulated it so astutely!

by the_af 3y ago

btbuildem 3y ago

Same on twitter (as in the screenshot in the article) -- lots of regurgitative comments, adding words but stating nothing / rephrasing the parent tweet / etc. I think the next step is actually a big leap backwards - human-curated knowledge bases, islands in the sea of AI-generated noise.

grugagag 3y ago
LLM will be quick to catch onto the islands. They’d have to be private islands to fend off the corruption.
- the_af 3y ago
  Do you mean silos that LLMs and bots cannot access? Like the walled gardens of Slack, Discord, Facebook, etc? In that case the cure will be worse than the disease, in my opinion. I don't want a siloed internet. Silos exist and are bad enough now, imagine if in the future the only good, human-produced content is secreted away within them.
- spookie 3y ago
  Yes, they'll most likely be private islands. Honestly, it won't be too bad now that I think about it.
  - helf 3y ago
    [dead]
toss1 3y ago
It will be ironically funny if the original Yahoo! model of fuman-curated search results wins in the end. Also, the LLMs aren't necessarily doomed. Presumably, the humans running them are looking out for the problems of inbreeding or LLMs-eating-their-own-excrement, and will implement some combination of not feeding that crap, and/or freezing the model at a peak point where the results are still primarily based on human creations. Perhaps ChatGPT 4.3 will be the peak? EDIT: typos

surfingdino 3y ago

> I'm not sure about the "rebels" at the end of the article, how will they find each other in sufficient numbers, how will word of mouth spread in sufficient numbers? Friends of friends networks, personal recommendations sent over email, communicators, Discord.

the_af 3y ago
Would this scale, when it cannot outcompete garbage content now?
- surfingdino 3y ago
  I don't know, but I think we need to give the consumers/recipients of information better tools for discovery, ideally at source instead of disseminating it through intermediaries who exist only to provide real estate for ads.
- dylan604 3y ago
  Yeah, seems like email was already destroyed as anything useful before LLMs kicked in.

wussboy 3y ago

The only true solution to this is seldom mentioned: Real life. "Word of mouth" can still happen (and should happen) by reals words, uttered in the physical world, by real mouths, in the physical world. I suspect many of our current societal ills are caused by too much information spreading too quickly, not forced to go through a multitude of brains or stand the tests of time.

s1artibartfast 3y ago

I dont think rebel numbers is an issue. It didnt take long for Facebook to get massive penetration using other modes of communication. I think this will be a radical change for the better. Once the baseline quality and trust drops low enough, there is as an incentive for curation, walled gardens, and reputation based networks. I hope for the return of webrings, moderated forums, and quality standards for participation. I also think it may finally break the hold of social media obsession. Outrage loses much of its luster when people realize the other end of it is likely a bot. This is a step toward asking "who is this person, and why do I care what they think."

internetter 3y ago
This is what I was getting at, but you formulated it so astutely!