Are_Euclidding_Me [e/em/eir]

Are you queer yourself? Because that question, “how do you know you’re right about your sexuality if you’ve never slept with [insert gender here]?”, is a very common queerphobic question that a lot of us get at one point or another, usually from well-meaning cishets who, like yourself, don’t understand that it’s actually a really shit question that is extremely invalidating.

Also, if I’m reading the situation right, you asked a bi man if maybe he is just gay, not bi. That’s a classic biphobic impulse. Bi men tend to be told “you’re probably gay and just pretending you like women”, while bi women are often told “you’re probably straight and just pretending to like women”.

So on two counts here your question was a bad question, and the people telling you this aren’t the people who need to be less sensitive, you are. You need to accept the criticism queer people are giving you and understand that you’re in the wrong here

And exponential improvement means

What do you mean by this? What “exponential improvement” do you see happening with LLMs?

So you’re clearly aware of the real harms this industry causes, such as the absolutely appalling example from your comment.

All I’m asking of you is to be more worried/appalled by real harms than of marketing hype. The real harms are there, they’re things like making people in Kenya view csam, the hypothetical future “superintelligence” is marketing hype.

Hey, you’ve been taken in by marketing hype. Large Language Models (LLMs) will never be “superintelligences” (whatever the fuck people even mean by that term).

Please worry about the very real harms of LLMs (environmental, economic) rather than the hypothetical future harms a “superintelligence” (whatever the fuck that means) might cause.

Please don’t instantly believe google executives who say “superintelligence” is just over the horizon. They are either grifters who are doing marketing or they’re Business Idiots who can’t tell the difference between real work and complete bullshit.