• 0 Posts
  • 277 Comments
Joined 2 years ago
cake
Cake day: June 11th, 2023

help-circle

  • I’m not necessarily saying they’re conflicting goals, merely that they’re not the same goal.

    The incentive for the generator becomes “generate propaganda that doesn’t have the language chatacteristics of typical LLMs”, so the incentive is split between those goals. As a simplified example, if the additional incentive were “include the word bamboo in every response”, I think we would both agree that it would do a worse job at its original goal, since the constraint means that outputs that would have been optimal previously are now considered poor responses.

    Meanwhile, the detector network has a far simpler task - given some input string, give back a value representing the confidence it was output by a system rather than a person.

    I think it’s also worth considering that LLMs don’t “think” in the same way people do - where people construct an abstract thought, then find the best combinations of words to express that thought, an LLM generates words that are likely to follow the preceding ones (including prompts). This does leave some space for detecting these different approaches better than at random, even though it’s impossible to do so reliably.

    But I guess really the important thing is that people running these bots don’t really care if it’s possible to find that the content is likely generated, just so long as it’s not so obvious that the content gets removed. This means they’re not really incentivised to spend money training models to avoid detection.








  • But the standards for an organisation like Amnesty International saying a state is committing genocide are much higher than a random person on the internet.

    To make a claim like that, they have to have specific evidence satisfying the actual definitions in international law, which is what this whole report is about. It’s all well and good for you to go “well it’s obvious to me”, but that doesn’t meet the standards of evidence for a reputable NGO like them to make a statement like that.

    They agree with your stance, so I’m not sure I understand why your response to them - explicitly - saying “this is genocide” is to chew them out for it.





  • Making decisions and recognizing a state are fundamentally different things though, right?

    Recognition is a very specific thing where a nation formally acknowledges their existence as a state, which also affects their ability to e.g. make diplomatic agreements.

    But doing so is totally separate from how you act toward that nation in practice.

    Russia, for example, recognises Ukraine as a country (currently), but actively does not respect their right to self-determination or their internationally recognised borders. But it would be wrong to claim that they don’t recognise Ukraine, despite that.





  • Sure, but you’ll notice I explicitly referenced that when I said hamas is incentivised to minimise risk of them being hit by Israeli strikes, especially by keeping them separately in members’ homes and in tunnels. The same goes for starvation, in that, if at all possible, hamas actively wants to keep them fed enough to survive, where that isn’t at the cost of starving to death themselves.

    You’re right that some/many (maybe even most, I have no data wrt specific numbers) of the hostages have, and will continue to die due to Israel’s actions, but that’s significantly different from your initial claim of “every single hostage is dead”