4 ms·

What if, future AIs gain capabilities to sort, categorize, discriminate, deduplicate and rank new post-2022 datasets? When we ask for a text to be human produce

What if, future AIs gain capabilities to sort, categorize, discriminate, deduplicate and rank new post-2022 datasets? When we ask for a text to be human produced, what we are really asking for is novelty and groundedness to facts. Most of the opinions on reddit is regurgitated hive mind thoughts.

ChatGTP 3y ago
Not the valuable stuff, the the world evolves and there’s an almost infinite supply of new stuff to talk about.
- dylan604 3y ago
  and yet here we are like a broken record talking about chatbots
  - retrocryptid 3y ago
    A Comment section post, like language, is a virus.
  - ChatGTP 3y ago
    What should we talk about then ?
juujian 3y ago
Once you have a reliable model to detect AI-content, you also have a new tool to train AI-models against. So ironically, any tool to detect AI content will just serve to make AI content more difficult to detect.
- PartiallyTyped 3y ago
  Which you can use to train a better model to detect AI content. I wonder whether some models fit this - well, you could certainly call it generative, and obviously adversarial - framework of networks?
- passion__desire 3y ago
  The new AI models will seek novelty[0], pieces of facts not recorded before. It will directly hone in on newness not how it was delivered. Writing styles could be another vector in which novelty could be detected but that's a different matter. [0] https://people.idsia.ch/~juergen/sice2009.pdf https://people.idsia.ch/~juergen/sice2009.pdf
- bee_rider 3y ago
  But the ability to detect, say, content generated by pre-2033 chatbots will open up a whole decade of previously contaminated content.
- collegeburner 3y ago
  right, you're describing a gan-like system/positional arms race. which is how adversarial tech always develops. doesn't mean we won't get better AI out of it.
pessimizer 3y ago
I don't think that the ability to fuck something implies that there's an ability to unfuck something. I don't think there's any necessity to think that there will be an continually increasing ability to distinguish faked from real in past data. The analogy would be how we can't distinguish fantasy, parable, and propaganda in histories written in the past, even though we have a lot more technology than they had 2500 years ago. We rule out most things that seem fantastic, like dragons or people rising from the dead (maybe the equivalent of finding extreme statistical improbabilities), while also entertaining the possibility that the past could have been radically different than we see it now, and some of the things we ruled out may have been real. We make suppositions based on the political alignments of the historians, and whether they could have had access to the information that they claim to record. We use geological and anthropological records to check credibility, and use confirmations found as sources for priors when judging other, maybe unrelated, things that said by that same scholar. We've seen what current AI does when confronted by a situation like that, it fills in the gaps with plausible fantasy and hallucinates. edit: not that there isn't low hanging fruit in detecting AI creations, but that's just a sign of early AI. AIs can't even draw hands.
- retrocryptid 3y ago
  Aye. Enshitification and Data Pollution appear to be thermodynamic processes. Easy to go one way, hard to go the other.
  - passion__desire 3y ago
    How about a new metaphor, complexification. The maximal complex state is useless and the initial simple state is also useless. Interesting patterns arise only in the middle like while mixing coffee and cream [0]. [0] https://www.youtube.com/watch?v=NgAtvbRqckQ https://www.youtube.com/watch?v=NgAtvbRqckQ
    - retrocryptid 3y ago
      +1 on Sean Carroll. I think he overestimates the strength of human cognition, but that's not so vile sin as underestimating it. And he's a very good speaker I terms of getting his point across. But yeah. There are definitely entropy analogies in software development. A couple of places I've worked at I've described as "brownian." Engineering would move forward a bit then marketing would change the requirements and we would make progress towards the new requirements, but before long we would get new requirements. If you mapped out progress starting at the origin and then put different objectives along the peremiter of a unit circle, progress would look like the random walk of an atom in a warm gas. Those are the places where you ship a product when progress accidentally meets expectation.

ChatGTP 3y ago

Not the valuable stuff, the the world evolves and there’s an almost infinite supply of new stuff to talk about.

dylan604 3y ago
and yet here we are like a broken record talking about chatbots
- retrocryptid 3y ago
  A Comment section post, like language, is a virus.
- ChatGTP 3y ago
  What should we talk about then ?

juujian 3y ago

Once you have a reliable model to detect AI-content, you also have a new tool to train AI-models against. So ironically, any tool to detect AI content will just serve to make AI content more difficult to detect.

PartiallyTyped 3y ago
Which you can use to train a better model to detect AI content. I wonder whether some models fit this - well, you could certainly call it generative, and obviously adversarial - framework of networks?
passion__desire 3y ago
The new AI models will seek novelty[0], pieces of facts not recorded before. It will directly hone in on newness not how it was delivered. Writing styles could be another vector in which novelty could be detected but that's a different matter. [0] https://people.idsia.ch/~juergen/sice2009.pdf https://people.idsia.ch/~juergen/sice2009.pdf
bee_rider 3y ago
But the ability to detect, say, content generated by pre-2033 chatbots will open up a whole decade of previously contaminated content.
collegeburner 3y ago
right, you're describing a gan-like system/positional arms race. which is how adversarial tech always develops. doesn't mean we won't get better AI out of it.

pessimizer 3y ago

I don't think that the ability to fuck something implies that there's an ability to unfuck something. I don't think there's any necessity to think that there will be an continually increasing ability to distinguish faked from real in past data. The analogy would be how we can't distinguish fantasy, parable, and propaganda in histories written in the past, even though we have a lot more technology than they had 2500 years ago. We rule out most things that seem fantastic, like dragons or people rising from the dead (maybe the equivalent of finding extreme statistical improbabilities), while also entertaining the possibility that the past could have been radically different than we see it now, and some of the things we ruled out may have been real. We make suppositions based on the political alignments of the historians, and whether they could have had access to the information that they claim to record. We use geological and anthropological records to check credibility, and use confirmations found as sources for priors when judging other, maybe unrelated, things that said by that same scholar. We've seen what current AI does when confronted by a situation like that, it fills in the gaps with plausible fantasy and hallucinates. edit: not that there isn't low hanging fruit in detecting AI creations, but that's just a sign of early AI. AIs can't even draw hands.

retrocryptid 3y ago
Aye. Enshitification and Data Pollution appear to be thermodynamic processes. Easy to go one way, hard to go the other.
- passion__desire 3y ago
  How about a new metaphor, complexification. The maximal complex state is useless and the initial simple state is also useless. Interesting patterns arise only in the middle like while mixing coffee and cream [0]. [0] https://www.youtube.com/watch?v=NgAtvbRqckQ https://www.youtube.com/watch?v=NgAtvbRqckQ
  - retrocryptid 3y ago
    +1 on Sean Carroll. I think he overestimates the strength of human cognition, but that's not so vile sin as underestimating it. And he's a very good speaker I terms of getting his point across. But yeah. There are definitely entropy analogies in software development. A couple of places I've worked at I've described as "brownian." Engineering would move forward a bit then marketing would change the requirements and we would make progress towards the new requirements, but before long we would get new requirements. If you mapped out progress starting at the origin and then put different objectives along the peremiter of a unit circle, progress would look like the random walk of an atom in a warm gas. Those are the places where you ship a product when progress accidentally meets expectation.