Posts

Comments

Comment by tremmor19 on Dialogue on the Claim: "OpenAI's Firing of Sam Altman (And Shortly-Subsequent Events) On Net Reduced Existential Risk From AGI" · 2023-11-22T11:40:25.969Z · LW · GW

Yea, im with you on it being a false dichotomy: some people are saying that because the board lost when they tried to use their legal power, it means they never had any power to begin with, so it doesnt matter. It seems plausible they had some level of power, but it was fragile and subject to failure if used carelessly in the wrong circumstances. Like, well, most real world political power.

Comment by tremmor19 on Killing Socrates · 2023-04-12T18:10:56.327Z · LW · GW

I was discussing this with someone the other day-- is free speech, plus high standards of rationality, enough to make a good discussion? Or is it necessary to in some ways control the content as well-- active curation and cooperation, controlling for a culture of thoughtful discussion?

I had been coming from a very negative definition of free speech-- as in an absence of control, censorship-- the ability to share thoughts even if they went against the grain, (so long as they were not actually lying or misleading, etc.)-- placing the value on the discussion being open. 

The other person is very smart but comes from a pretty different background, and had not put a strong value on being able to say whatever you want, if that speech did not contribute to the conversation. Rather, they had assumed a positive value of free speech-- that everyone wanting to participate could actively feel included and able to communicate their thoughts. They felt that communities favoring discussion would want to find interesting and diverse perspectives, and have norms that curated primarily good-faith participation. I'm not sure it's an exact response, but seemed on-topic to this post

Comment by tremmor19 on LW Team is adjusting moderation policy · 2023-04-08T16:15:09.536Z · LW · GW

Im noticing two types of comments I would consider problematic increasing lately-- poorly thought out or reasoned long posts, and snappy reddit-esque one-line comments. The former are more difficult to filter for, but dealing with the second seems much easier to automate-- for example, have a filter which catches any comment below a certain length too be approved manually (potentially with exceptions for established users)

There's also a general attitude that goes along with that-- in general, not reading full posts, nitpicking things to be snarky about, not participating in discussion or responding when someone attempts to engage. Honestly, Id much rather see twenty people making poorly-thought-out longposts that don't know what they're talking about, as long as they're willing to participate in discussion, than an increase of these guys. At least some of the confused but sincere guys can be curious and engaging.

So to that end, I'd like to support moderation that focuses on culling low-effort, failure-to-participate or failure-to-read-the-damn-post. a flag for users who regularly make short comments could be that, or better methods for regular users to flag comments for review for lack of engagement. Or perhaps an automatic flag for how often someone posts once and doesn't respond to any comments

I do think this is mostly what you meant anyway, I just wanted to point out what exactly i personally are as the issue versus just "underinfomed newbies".

Comment by tremmor19 on Human-AI collaborative writing · 2023-02-14T02:10:03.084Z · LW · GW

Co-opperative writing has been the first use I've found for these models that actually deeply appeals to me-- I'd seen the examples of people talking to LLMs as companions (Replika, Character.ai, etc), which were certainly interesting but didn't really hold much long term appeal. But using ChatGPT and GPT3 to speed up my creative writing has been amazing and I can't imagine I will ever get bored of it. I have an endlessly patient beta reader to nod its head and encourage me when I bounce ideas off it, add a paragraph to finish up the scene I was bored with, fill in the places where I don't have any particular ideas  ("Okay, I need some kind of monster to jump out here-- give me five quick descriptions of an original fantasy monster that would be found in a desert")-- when before I probably would have just put a generic placeholder zombie or something.