Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

How do I know this is not Sora?


Chain of reputation. If you can trace back the claim to a person or persons who have reputation to stake on this, then it's unlikely to be completely fabricated.

There are tech-related ways to tell for now but eventually it's going to come back to this.


> Chain of reputation. If you can trace back the claim to a person or persons who have reputation to stake on this, then it's unlikely to be completely fabricated.

AKA "Provenance" but digital, for those who want to look at existing methodologies. https://en.wikipedia.org/wiki/Provenance


Additionally, in a very concrete technical sense, "whatever1" must rely on the "chain of reputation" of the https certificate system to have confidence that what they saw is not sora.


One angle could be to consider it from a game theory perspective.

Is this the sort of organisation that would be negatively impacted by publishing unverified stories?

If so, what is the likelihood that the content is just Sora?


This is an important analysis to perform but it's far from a sure thing. Motives can be murky and hard to assess. Maybe there is one particular scientist that has a baby on the way and fears he is about to be laid off unless he can get a sensational article published asap. A little helping hand from ai could be just the thing, and it's based on a true story just touched up a little bit and besides it's not like the readers will suffer any real harm from this tiny little transgression. Just one little shortcut one little time off course after that it's right back to honest science.


Motives can be murky and hard to assess.

FOH with that FUD, if you care about corruption go after the obvious examples instead of making up new ones.


Game theory or Bayesian inference?


I'm an expert in neither, but I would say both?

I thought of game theory initially because I framed the situation as a repeated game, where every article published is a new round.

https://en.wikipedia.org/wiki/Repeated_game

But then I went and muddied it with the word "likelihood" :)


It is good to be skeptical, but there is a large amount of detail in the paper itself that would have taken quite a bit of effort to fabricate (for no good reason): https://www.sciencedirect.com/science/article/pii/S235198942...

There is enough detail there to book a trip to Germany and set up infrared cameras if we are so inclined, repeatability is a large part of science.


Because the person filming it brought the camera with approved government ID. Got their camera serial number recorded in the government database. The camera then embeds its serial number into the video using hidden watermarks. Just Joking .... for now.


The bat and rat were identified as non visa holders by ICE.


Sora is too smooth




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: