The meta analysis addresses porn in general. That includes fetishized content like violent or "taboo" pornography. It states there's no evidence that it makes sexual aggression more prevalent, and that population studies show that it's at least correlated with a reduction instead.
We can nitpick the wording all day long, but ultimately I think the takeaway is that there's no evidence that it has negative effects, and there's at least some evidence that suggests it has positive effects.
We don't have a specific cordoned off section for meth and cocaine in our brains. Many things trigger those areas of the brain, including some pretty innocuous stuff.
Porn isn't physically addictive like meth and cocaine. It can be psychologically addictive though, but that goes for a lot of things out there.
Stuff like meth and cocaine can actually alter your brain, porn does not.
Anyone can develop an unhealthy relationship with porn, but that goes for just about anything out there.
This is not contradictory.
The meta-study says that pornography contributing to sexual aggression is not proven. Meaning, it doesn't make it worse.
Meanwhile the population study seems to suggest porn usage reduces sexual aggression, or is at least correlated with it.
I personally haven't had to use a chromium browser for anything yet since my swith to Firefox. Only to test a render bug in chromium that Google hasn't bothered to fix in over 9 years for a case that works correctly in every other browser.
There is a clear power disparity between a father and a daughter. It's debatable if the daughter could even realistically consent in that case.
But a depiction of it in porn is in my view not inherently unethical. I can disapprove of it personally, but that doesn't mean we should start banning it based on feelings of inappropriateness.
We depict murder and violence in movies and video games too. Actual murder is of course not exactly ethical, but we have no problem accepting it in a movie, because nobody is actually being murdered. You might not like to watch a movie like Saw or something (I personally don't), but it doesn't make the movie itself unethical. To me, porn is no different. There's a clear separation between fiction and reality.
Where imo a line is crossed, is if said media actually makes a clear effort to promote these acts IRL. But that's not the case here.
https://pubmed.ncbi.nlm.nih.gov/32691692/
Population studies suggested that increased availability of pornography is associated with reduced sexual aggression at the population level
Does he actually have a diagnosis or are you making that up or assuming things?
Porn made with the willing consent of all parties involved, where everyone is compensated appropriately. No harm = no ethical problems as far as I'm concerned. Most big studios these days make sure of this. But there have also been pioneers that push the bar further up (e.g. Lustery, Ersties or Erika Lust).
I can think it's a messed up fantasy, but that doesn't mean it should immediately be banned by a payment processor.
Regardless, there are tons of studies showing that consuming this kind of porn actually helps prevent people from acting on these fantasies. The net result is likely less sexual abuse, not more. Because it's fantasy media, it likely is able to keep the fantasy a fantasy, it gives people an outlet.
Well the thing is, LLMs don't seem to really "solve" complex problems. They remember solutions they've seen before.
The example I saw was asking an LLM to solve "Towers of Hanoi" with 100 disks. This is a common recursive programming problem, takes quite a while for a human to write the answer to. The LLM manages this easily. But when asked to solve the same problem with with say 79 disks, or 41 disks, or some other oddball number, the LLM fails to solve the problem, despite it being simpler(!).
It can do pattern matching and provide solutions, but it's not able to come up with truly new solutions. It does not "think" in that way. LLMs are amazing data storage formats, but they're not truly 'intelligent' in the way most people think.
Completion is not the same as only returning the exact strings in its training set.
LLMs don't really seem to display true inference or abstract thought, even when it seems that way. A recent Apple paper demonstrated this quite clearly.
The inability to distinguish between selection and socialization means there's no evidence for a causal link. At best, it suggests that people who commit sexual aggression generally like porn featuring it more, but even that is apparently a weak correlation apparently.