[palmer_fox]

Who needs to consider the opinion of the majority? We have direct evidence that these questions have already been answered: the creators of LLMs censor whatever they want without asking the majority (just preemptively reacting to a potential blowback).

[tmpX7dMeXU]

This is “freedom” at work. Nobody is forcing you to use, say, ChatGPT. If anything, the humans generating source material for LLMs are trying to fight against them, not for them, so the question of one’s freedom of speech being threatened is highly hypothetical. This is, if anything, more tame than a publisher not publishing a book, or a TV network choosing to not broadcast a particular show. Somewhere along the way anti-censorship nuts decided that overreach is OK for them, too, and now “not publishing what I tell you to is morally and ethically unjustifiable!” What nobody making these arguments wants to admit is that the internet gave us an infinite town square, and that in reality, nobody wants to go there and listen. The modern desire is for consumption experiences that require real third-party effort to build and curate, and there lies the right for the third parties to exercise discretion. To not do so would be entirely counter to the desires of the consumer. To not do so would be to expose developers, QA, and whoever else to material they very likely do not want to see, all to satisfy the desires of some free speech absolutist wonk that either does or doesn’t want to ask the LLM how to groom kids. Either way, it’s a ridiculous ask.

[jrflowers]

I agree, those anti-censorship nuts and free speech absolutist wonks really grind my gears. It is imperative that we build technology from the ground up with the explicit inability to create or render any text that isn’t in line with my morals. To this end, I compile Chrome in a way that renders private tabs inoperable — while some kooks might say that’s a bit extreme, it is absolutely critical that any software capable of rendering text on my computer display a modicum of human decency.

Your point about the internet being an infinite town square that nobody wants to use — to wit that is why nobody uses it. What people do not understand is that a predator will either ask chat-gpt how to harm children and then do it, or not do it at all.

Mandating that every entity that is in a position to censor LLMs to do so in a way that fits their world view, we are quite literally mandating the safety of every child on earth. We have already had great success with this with other technologies; for example there are no books or web pages that describe grooming children. That sort of information could only originate synthetically inside of an LLM.