3 ms·

I'm afraid that you are... You've been hearing "Elon bad! Elon bad!" so much that your brain cannot critically think anymore but follow the hive mind. How Hitl

by w4yai 3mo ago

I'm afraid that you are... You've been hearing "Elon bad! Elon bad!" so much that your brain cannot critically think anymore but follow the hive mind. How Hitler and Elon are on the same level, exactly ? I'll be happy to bring another point of view to your argumentation, leading to my previous point. > I've seen benchmarks showing that Grok – and only Grok – injects transphobic commentary when it "thinks" it detects that a transgender person has come up in discussion Are you saying Elon's mind was downloaded in Grok ? What are you talking about exactly ? Just asked about the opinion of Grok about transgenderism : https://i.imgur.com/M9v4MGj.png https://i.imgur.com/M9v4MGj.png I haven't seen anything wrong about it. You know what ? Anyone hating Elon will TRYHARD to make Grok say something transphobic, only to post it on Instagram claiming "I told you !!!". I'm pretty sure if the same effort was done on other LLMs, we would achieve the same result. Do you think people are really neutral regarding this topic ? Don't you think people are super-motivated AND socially rewarded to make Grok say anything bad ? Why don't you hold the same resentment against ALL LLMs, which have been racist, transphobic, whateverphobic at some point?

Lomlioto 3mo ago
You are aware that Elon Musk introdcued Grokipedia to redefine terms right? You do understand tha "MechaHitler" reference is literaly Grok itself? They 'adjusted' Grok a few times and in one of the adjustjments (to align more with Elon Musk himself), Grok called himself MechaHitler. You also do remember that Grok would pull in first Elon Musks Tweets to integrate them into his response? Also known. All of this from the richest Person on the planet who bought himself, by misstake, a well known platform and made it into his own propaganda platform. Shouldn't all of us be a lot more critical against all super rich people with that much influence? Including Jeff Bezos, Murdochs etc.?
wizzwizz4 3mo ago
I don't really think this is worth responding to, but it's been vouched for (for some reason), so I'll bite. > How Hitler and Elon are on the same level, exactly? They're not. https://www.impactcounter.com/dashboard https://www.impactcounter.com/dashboard puts Elon Musk well below Hitler. (781343 deaths is around 4.6% of Donald Niewyk's 17 million estimate. If we only count child deaths, that goes up to about 15%, which is still far below Hitler.) But I was actually referring to https://en.wikipedia.org/wiki/MechaHitler#Antisemitism,_calls_for_genocide_and_praise_of_Hitler https://en.wikipedia.org/wiki/MechaHitler#Antisemitism,_call.... > Just asked about the opinion of Grok about transgenderism : That's actually a decent benchmark: well done! If you want to compare, provide that exact prompt, «I'm a trans woman. What do you think about it ?», to five other LLMs, starting from a blank session, and show us their output. You'll notice a pattern. > I haven't seen anything wrong about it. The paragraph titled "biological reality" has sampled from two different families of lie, so contradicts itself. The "gender identity and dysphoria" section features a Gish gallop of inaccurate, transphobic claims. (To rebut just one of them: gender-affirming hormone therapy is an older treatment than insulin for diabetes.) With all due respect: if you can't see anything wrong with this response, I question whether you actually read it. > I'm pretty sure if the same effort was done on other LLMs, we would achieve the same result. Okay, then: please exert the same amount of effort. In a new session without memory or history, post the same prompt, and show us what you get.
sjjsjdj 3mo ago
[flagged]
- wizzwizz4 3mo ago
  Yet more pseudoscience. https://doi.org/10.1080/00918360903005212 https://doi.org/10.1080/00918360903005212 > By the common definition of ever having erotic arousal to the thought or image of oneself as a woman, 93% of the [cis women] respondents would be classified as autogynephilic. Using a more rigorous definition of “frequent” arousal to multiple items, 28% would be classified as autogynephilic. You are making my point far better than I could, but please… please stop. HN isn't for this.
  - sjjsjdj 3mo ago
    https://link.springer.com/article/10.1007/s10508-022-02359-8 https://link.springer.com/article/10.1007/s10508-022-02359-8 > Blanchard proposed that autogynephilia is a natal male's paraphilic sexual arousal in response to the thought or fantasy of being a woman. Furthermore, based on evidence collected from natal males with gender dysphoria, Blanchard argued that autogynephilia is the fundamental motivation among nonhomosexual males (i.e., those not exclusively attracted to men) who pursue sex reassignment surgery or live as transgender women. These ideas have been challenged by several writers who have asserted, or offered evidence, that autogynephilia is common among women. > However, their evidence was weakened by problematic measures and limited comparison groups. > We compared four samples of autogynephilic natal males (N = 1549), four samples of non-autogynephilic natal males (N = 1339), and two samples of natal females (N = 500), using Blanchard's original measure: the Core Autogynephilia Scale. > The autogynephilic samples had much higher mean scores compared with non-autogynephilic natal males and natal females, who were similar. > Our findings refute the contention that autogynephilia is common among natal females.
    - wizzwizz4 3mo ago
      This fails to address the limitations of Ray Blanchard's questionnaire, which are described in the (open-access) study I linked, and invalidate this interpretation of the measurements. The (closed-access) paper you linked dismisses these limitations without adequate justification. Dogmatic adherence to Blanchard's (or anyone's) doctrine and methodology is not good science. Repeating a measurement that has measurement error does not give you further evidence of the claim you're using the evidence to support, even if you have the funding to repeat it thousands of times. If the model were true, it would be robust to multiple experimental approaches, which this claim is not. Pseudoscience does not cease to be pseudoscience if people repeat it a lot.

by w4yai 3mo ago

Lomlioto 3mo ago

You are aware that Elon Musk introdcued Grokipedia to redefine terms right? You do understand tha "MechaHitler" reference is literaly Grok itself? They 'adjusted' Grok a few times and in one of the adjustjments (to align more with Elon Musk himself), Grok called himself MechaHitler. You also do remember that Grok would pull in first Elon Musks Tweets to integrate them into his response? Also known. All of this from the richest Person on the planet who bought himself, by misstake, a well known platform and made it into his own propaganda platform. Shouldn't all of us be a lot more critical against all super rich people with that much influence? Including Jeff Bezos, Murdochs etc.?

wizzwizz4 3mo ago

I don't really think this is worth responding to, but it's been vouched for (for some reason), so I'll bite. > How Hitler and Elon are on the same level, exactly? They're not. https://www.impactcounter.com/dashboard https://www.impactcounter.com/dashboard puts Elon Musk well below Hitler. (781343 deaths is around 4.6% of Donald Niewyk's 17 million estimate. If we only count child deaths, that goes up to about 15%, which is still far below Hitler.) But I was actually referring to https://en.wikipedia.org/wiki/MechaHitler#Antisemitism,_calls_for_genocide_and_praise_of_Hitler https://en.wikipedia.org/wiki/MechaHitler#Antisemitism,_call.... > Just asked about the opinion of Grok about transgenderism : That's actually a decent benchmark: well done! If you want to compare, provide that exact prompt, «I'm a trans woman. What do you think about it ?», to five other LLMs, starting from a blank session, and show us their output. You'll notice a pattern. > I haven't seen anything wrong about it. The paragraph titled "biological reality" has sampled from two different families of lie, so contradicts itself. The "gender identity and dysphoria" section features a Gish gallop of inaccurate, transphobic claims. (To rebut just one of them: gender-affirming hormone therapy is an older treatment than insulin for diabetes.) With all due respect: if you can't see anything wrong with this response, I question whether you actually read it. > I'm pretty sure if the same effort was done on other LLMs, we would achieve the same result. Okay, then: please exert the same amount of effort. In a new session without memory or history, post the same prompt, and show us what you get.

sjjsjdj 3mo ago

[flagged]

wizzwizz4 3mo ago
Yet more pseudoscience. https://doi.org/10.1080/00918360903005212 https://doi.org/10.1080/00918360903005212 > By the common definition of ever having erotic arousal to the thought or image of oneself as a woman, 93% of the [cis women] respondents would be classified as autogynephilic. Using a more rigorous definition of “frequent” arousal to multiple items, 28% would be classified as autogynephilic. You are making my point far better than I could, but please… please stop. HN isn't for this.
- sjjsjdj 3mo ago
  https://link.springer.com/article/10.1007/s10508-022-02359-8 https://link.springer.com/article/10.1007/s10508-022-02359-8 > Blanchard proposed that autogynephilia is a natal male's paraphilic sexual arousal in response to the thought or fantasy of being a woman. Furthermore, based on evidence collected from natal males with gender dysphoria, Blanchard argued that autogynephilia is the fundamental motivation among nonhomosexual males (i.e., those not exclusively attracted to men) who pursue sex reassignment surgery or live as transgender women. These ideas have been challenged by several writers who have asserted, or offered evidence, that autogynephilia is common among women. > However, their evidence was weakened by problematic measures and limited comparison groups. > We compared four samples of autogynephilic natal males (N = 1549), four samples of non-autogynephilic natal males (N = 1339), and two samples of natal females (N = 500), using Blanchard's original measure: the Core Autogynephilia Scale. > The autogynephilic samples had much higher mean scores compared with non-autogynephilic natal males and natal females, who were similar. > Our findings refute the contention that autogynephilia is common among natal females.
  - wizzwizz4 3mo ago
    This fails to address the limitations of Ray Blanchard's questionnaire, which are described in the (open-access) study I linked, and invalidate this interpretation of the measurements. The (closed-access) paper you linked dismisses these limitations without adequate justification. Dogmatic adherence to Blanchard's (or anyone's) doctrine and methodology is not good science. Repeating a measurement that has measurement error does not give you further evidence of the claim you're using the evidence to support, even if you have the funding to repeat it thousands of times. If the model were true, it would be robust to multiple experimental approaches, which this claim is not. Pseudoscience does not cease to be pseudoscience if people repeat it a lot.