popoi

In general there are pretty obvious signs that an F1 car is being driven well or poorly. The problem here is that we have a tool that's weighted toward responding positively enough that it will confirm obviously wrong answers, and it doesn't seem like people are doing the work necessary to guard against that. Appealing to the idea that someone could use the tools doesn't seem to address that, and it seems like exactly the same kind of generalization you got on me about, just in the opposite direction.

Many studies show that gender-affirming care up to and including surgical and hormonal interventions is effective at treating gender dysphoria in many people, and it is endorsed by a large number of medical organizations for that reason. Evidence for the alleged downsides like regret has failed to materialize, and the previous review I posted as well as others show that regret is very low, both in an absolute sense (.3-3.8% in the first and <1% in the second), and relative to other surgeries and life decisions. Most of the evidence that what if has been appealing to from Sweden makes the claim that the existing studies are inconclusive, but that seems to be mostly because they're just deeming most of the studies "low quality" due to things like lacking a randomized controlled trial, without really explaining how we're supposed to do such a trial ethically or practically, and ignoring that "only a minority of outcomes for health care interventions are supported by high-quality evidence". Insisting that we can't make any conclusion from "low-quality" evidence is holding gender-affirming care to an unreasonably high standard, I think most likely for political reasons.

An LLM shouldn't be regarded as a source. At best, it can point to things that are actual sources and provide a summary. I would have been happy to address those actual sources if anyone would have actually posted any, but the last few posts in that thread were only appealing to the reported existence of those sources, and nobody seemed interested in addressing any of the contradictory sources I posted. Unsurprisingly, it's a study I was already aware of which has a bunch of problems and also doesn't actually measure the thing we were talking about, which was suicidality. This is why it's important to post sources for evaluation and discussion. I don't want the last word, I want people to respond substantively to the arguments I make and the questions I ask. On top of all the other problems, all this LLM use is making everyone really boring. I provided plenty of support for my argument against the specific session where possible, but people weren't actually addressing those arguments, and at the end weren't even posting the information necessary to evaluate specific sessions. I can't do anything with "I asked GPT and it agreed with me and said it could provide sources" other than pointing out why that doesn't mean much due to the design of LLMs. Having a more substantive discussion is a two way street.

Following up on this from a now closed thread from mike: The reason to assume that things might be made up is that LLMs have an extremely well demonstrated tendency to make things up sometimes. That's why every one of them that I'm aware of includes some kind of "This might be wrong, double check it" disclaimer, and why people whose use it for purposes that demand accuracy like lawyers keep getting punished when it produces fake citations that they don't catch. Leaving aside that presenting sources for evaluation is a basic part of proper debate, it's even more important when you can't reliably assume that the source actually exists in the first place.

I'm sorry, you didn't actually check the evidence it said it had? I'm still not watching an hour long video without some more specific detail from you about how you think it supports your case. I posted a response to the Cass review, and your only response so far has been to blow it off and pivot to an entirely unrelated blog article. You still haven't offered anything to demonstrate that those countries are particularly receptive in a way that makes their data more legitimate, and you haven't had any response to the other countries that are reaffirming their stance in favor of the effectiveness of affirming treatment. Quote them. You have so many of these bugbears that you persistently refer to, but you almost never actually point to what specific thing you're talking about, so as usual I have no idea what you're referring to. This doesn't seem to actually explain any of the evidence though. It just lets you dismiss anything you don't want to address. Because it just doesn't seem to be true. David Reimer's case seems like a stark (and disastrous) failure of that idea when it comes to gender identity. He still had a strong internal sense of gender identity that was separate from what he was assigned. And the very existence of trans people outside of environments where transition is an acceptable option, or before there were any available treatments seems like a strong indicator as well. Where do you imagine someone like James Barry learned that behavior in late 1700's England? Or Albert Cashier in late 1800's America? Are we to believe those were "receptive" environments as well? You should not come in to treatment with the intent of changing someone's gender identity. It's perfectly fine to come to the conclusion that something else is causing their dysphoria if the evidence suggests it. This seems like just a straight up lie. Cite any law that says that. No, the research doesn't say that. Do you understand what they mean when they say the evidence is "inconclusive"? It doesn't though? I've presented plenty of sources from other countries (as well as professional medical organizations) saying "Actually no, the evidence is conclusive, the treatment is effective, and we should be doing it".

I think you're conflating the correctness of the answer with the meaning of getting the answer. If your calculator is broken and sometimes gives you incorrect answers, you may still get correct answers sometimes, but it compromises the meaning of the calculator giving you a particular answer because we can no longer expect as a matter of course that the answers are correct. At the very least you have to do additional work to confirm the answers, and people by and large don't seem to be doing that, because they're just using it as a source of external validation. The problem when you do something like that is that if you're just looking for confirmation, reassuring you about the ways you're right can be interpreted as it just confirming you you're right. I do want to avoid it altogether, but my argument has pretty consistently been that it's too easy to get it to agree with an argument. I'm sure I could get it to agree with me, but the nice part about earning confidence by doing your own research is that you don't need anything telling you "Wow what a great point" every other sentence.

I believe you got it to say something you think agrees with you. The problem is that's pretty much meaningless given the demonstrated propensity for LLMs to agree with the user. And doubly so since you're not only continuing to not quote anything from the actual research in support of your argument, you're not even posting the chat so we can see what "the opinion i gave in this thread" is and what it actually said in response. Do you think you would get the same answer if you pretended to hold a different opinion? Have you tried? I'm willing to admit it's possible for LLMs to be useful sometimes but it doesn't seem like you (or most other people I see using them that enthusiastically) are using them in a way that accounts for their obvious deficiencies.

No? The authors of both the paper and the blog are included in both, and they don't have any names in common. Whose guns do you think are blazing there and for what purpose? The article seems entirely correct that the claim that pimozide is a legitimate general treatment for gender dysphoria on the basis of a single case report with very complex particulars is bunk. And the existence of multiple (since removed) videos making the claim seems to indicate it is or was a common enough talking point to be worth debunking. Are you saying we should accept their claim based on that single data point?

I agree it's probably not going to change my mind, but I think that has much more to do with the poor support for your position and your generally poor understanding of the subject than it does any kind of intransigence on my side. You shouldn't be reading the research and providing quotes to support your argument because you think it will change my mind, you should do it because it's the basic standard of arguing a point. Because you never posted anything other than a vague hand wave at an entire book that didn't appear to have much if anything to do with the subject. I am not going to read an entire book to try to figure out what you're talking about, you need to do the basic work of communicating what that book says that you think supports you and how. It doesn't look like it did, no. I don't see anything in the most recent chats that mentions record keeping. What part are you interpreting as having to do with record keeping? And not to overload you with reading that you're not going to do anyway, but here is a detailed response as to why the Cass Review is junk. It really doesn't seem like it has. It seems more like you're doing what is a pretty standard move for you where you take research that you think agrees with you on some point, in this case that the evidence for the benefits of hormones in minors is inconlusive, and trying to spin it as if it supports every other position you hold like "transition has no benefit", when it simply doesn't. About that...

No we're pretty firmly in the last category here. Many of your statements just aren't supported by the studies you're waving at. The way to disprove me is to quote from those studies with specific statements that show you're right, which unfortunately will require you to actually read them. Whining about it isn't going to work. Is it when you ask for a specific subset of research and then don't engage with any other research that disagrees? Because that seems more like cherry picking to me. They can though. There are many studies and countless anecdotes that show that. This is a strawman, nobody thinks that. It is entirely possible to be treated and function in society as a gender other than the one you were assigned when you were born though, which is what trans people actually want. And all available information shows they are tiny, especially relative to other common surgeries. The standard is not that nobody will ever regret their decision, because that would disqualify pretty much any medical treatment you could think of. The standard is that the benefits outweigh the risks, and that patients are given enough information about both to make an informed decision. None of what I read supports that conclusion. You need to quote something specific if you think it does.

This doesn't appear to be the case. Reading through the reviews (have you actually done this?) it looks like the main reason they consider it inconclusive is due to the rated quality of studies, not because of uncertainty or conflict about metrics. A huge chunk of medicine is based on people's subjective reports of their symptoms. "Patient reports gender dysphoria/depression/leg pain" is a perfectly scientific observation to make, and it's entirely possible to draw conclusions if we have enough of such observations to identify common patterns. We're talking about medicine, not law, and there are in fact several medical standards that define gender dysphoria and its treatments. At best it doesn't agree with you either, but the Swedish research is focused specifically on hormonal treatment for minors, so even then it's only addressing part of the question. Nothing I've seen indicates even any low quality studies that support your claim. In fact it seems like it's pretty much the exact opposite of what you claimed. Anecdotally, I've seen many people describe dysphoria in similar terms to "know[ing] they aren't the real deal", but it's with respect to their assigned gender. Once they're allowed to live as the gender they identify with, they report feeling not just better, but right. Again, no it doesn't. "Inconclusive" doesn't mean that they know it doesn't help, it means they can't draw a conclusion. Is there some hidden body of conclusive evidence for the idea that conversion therapy was ever helpful?

It doesn't actually show that though. You specifically asked it to only give you a subset of research, and now you're trying to pretend it's representative of the total state of research, but you're not actually addressing what I'm showing you that there is research showing positive effects to transitioning. If you're dedicated to listening to a computer tell you want to think you could at least go back and take it up on the offer to discuss different perspectives. Or even better you could actually try to engage with those different perspectives that I am helpfully showing you! What metrics do you want to use that show higher than a single digit percent of regret? Because that seems pretty straightforward, especially given that it's something you were pretty vehement about before. I asked you for a source for the claim that "swedish suicide rates are on par with the global rate among transgender". Leaving aside that this chat is not a source, do you understand why "transgender people have a higher rate of suicidal ideation and attempts than cisgender people" does not address that claim? You can at least demonstrate that they're significantly correlated, as the study I posted shows: I still haven't seen any basis to the claim that Sweden is particularly accepting, or that the rate of suicidality is consistent across other countries. If this were true then it seems like we wouldn't expect to see improvement after transition, but we do. Conversion therapy wasn't ever "help". I mean you could try to find and present some evidence or address some of mine, that might help? Or if you actually think that the evidence is completely inconclusive you could stop making claims that contradict that notion. It was though. The study I posted above is from 2014, and cites studies on the subject from the early 2000's. But also, one of the pioneering research facilities was burned down by the Nazis in 1933. You may not have been aware of the issue 10 years ago because fewer people stood to gain politically from making you mad about it, but that doesn't mean it didn't exist.

There don't appear to be any changing definitions in play here. If it seems like things are changing, my suspicion is that you're bouncing off of your own misunderstanding. But also, you've been discussing it this whole time despite your reservations about it being useless, so it seems like you don't even believe your own claim. And Germany, Austria, Switzerland, France, Australia and New Zealand, Canada, and I'm sure I could go on have reaffirmed their support. You are dishonest if you are trying to present things as if there is one evidence-based wave that is pushing back. It isn't though. Nothing about the definitions involved seems to have changed, and I've seen multiple studies that show regret to be single digit or lower percentages, and nothing that showed it to be higher than that. If you have some you want to actually post a link to this time, feel free. Or is there some particular definition you had in mind that you think changed that casts doubt on those studies? Should I take this to mean you didn't (or couldn't) read the many studies I've posted that show transitioning does produce improved outcomes and mental health? From the second post in this thread for example: This notion also runs contrary to you pointing to the Swedish stuff, because once again they specifically claim that there isn't enough evidence to assess effectiveness. From "Care of children and adolescents with gender dysphoria: Summary of national guidelines": To be fair I think they're wrong and they're artificially disqualifying most of the evidence that shows effectiveness, but either way it seems to leave you lacking when it comes to evidence of ineffectiveness. Source? But also, this Sweden right here? Oh for sure it sounds like a paradise!

I'm sorry, are you saying I should have accessed your chat session and asked it for references? Why didn't you already do that and post them? It is though. Many do, yes, which is where the transition care comes in. If they don't, then there is certainly room for helping them figure out if that's something they want, so long as it isn't just an excuse to string them along indefinitely. For at least the third time, no. Correct, because as every reputable organization I'm aware of maintains, conversion therapy is ineffective and harmful. What is your basis for concluding that Sweden is uniquely receptive such that the only valid data can come from there? What is your basis for talking about "poor outcomes" given that they themselves say there's not enough information to determine effectiveness? I dunno, it seems an awful lot like the Swedish recommendations are just a common talking point that you're parroting. The first thing it gave you to read from Sweden was "Care of children and adolescents with gender dysphoria: Summary of national guidelines", and from Finland was "Finnish national guidelines and evidence reviews", and both of the documents I linked include extensive citation of supporting research, so your complaint is off base in both directions.

Oh ok because you told the computer that it was because you didn't trust the data. I'm not sure why data from a less receptive country would be less trustworthy if the other side is supposed to be the unscientific one though. It doesn't seem like it does. This is just you appealing to the same stuff you did before that says "We don't have enough evidence" as if it supports statements like "the people that transition now will know the truth 10 years down the road, and i'll GUARENTEE you they will either regret transitioning or rope themselves" when it just doesn't. Even if we take the conclusion that there isn't sufficient evidence as a given (which is disputed), that also means that there isn't sufficient evidence for the idea that it's ineffective or dangerous. But if we're interested in non-US examples, since the last time this thread was active, recommendations from Germany, Austria, and Switzerland and France among others have rejected the idea that there isn't sufficient evidence and supported the effectiveness of transition care. So the implication that there's some international wave pushing back in a single direction doesn't really stack up.