@offby1 @bitprophet @brainwane In other words, I do agree that it makes sense to be more precise in how we talk about the current tech landscape. But it will be tricky, and it will hurt sometimes, especially online.
0 Boosts
0 Likes
@offby1 @bitprophet @brainwane Yeah, I get that. And that's really the rub.... Perhaps it's in response to David Gerard's push to "shame" others in an attempt to curb behavior, which, in my estimation, is the wrong tactic.
(Though, some would argue it's an appropriate toward fascist regimes, but I'm not as well-versed in that position.)
But, bear with me here as I'm not invalidating what you are saying—it can also feel like a provocation when others say "this is the new normal" or "this is what engineering is now" or something like that. Depending on the depth of one's belief, it can "sound" like "eating babies is okay now because we're more productive" (exaggeration to make the point... but sometimes it _feels_ that way in terms of the implied ethics).
Not saying that makes it okay to then act like an asshole, but also just to illustrate how certain viewpoints can have additional weight, depending on where you sit.
@pythonbynight @bitprophet @brainwane I totally agree! I don’t think that “just get in, your input doesn’t matter, we’re all doing it this way now” is a constructive pattern either; it’s dismissive, at best, of a whole set of concerns that are pretty damn reasonable to have.
@offby1 @pythonbynight @bitprophet @brainwane there are intersecting conversations here that cannot be disentangled:
1. is a culture that habitually expresses contempt healthy
2. is tone policing a good idea
3. is shaming a good strategy to change behavior
4. to what extent are "LLM users" a group of people who are doing harm and have an interest in sustaining their ability to do that harm vs. a group of people using a morally neutral tool in diverse ways who have no uniting class interests
@glyph @offby1 @pythonbynight @bitprophet @brainwane the answer to 1–3 is as always: “only if **I** think it's contempt-worthy"
It's the eternal terrorist vs freedom-fighter discussion.
Also you forgot:
5. at what point do we grant actual victimhood to give a free pass on 1–3?
@hynek @glyph @offby1 @pythonbynight @bitprophet @brainwane Also, to what extent have the "bad LLM users" appropriated both the term and the discussions to the point where more reasonable LLM users have been pushed out? In other words, if you want to be acknowledged as a (reasonable) LLM user, is it not fair to expect *you* to police the behaviour of others claiming that name?
@pfmoore @hynek @glyph @offby1 @pythonbynight @bitprophet @brainwane A large part of the problem, I believe, is that the idea that there is *any* good (in the ethical sense) usage of LLMs is very much so not in evidence, and indeed, many of us object to any LLM usage at all on that basis.
Obviously one can disagree with that ethical argument, but trying to separate good and bad LLM usage is on its own unlikely to resolve arguments and community tensions.
@pythonbynight @bitprophet @brainwane
> A user's desire to label/tag ANY LLM output as "slop" could be considered derogatory, but it could also be the shorthand way of saying "stochastic output that has no bearing on veracity which has to be validated in one form or another”.
That’s the trick, right? The conversation I’m thinking of has people *explicitly saying* they want to use the term in a derogatory way and that it is right that they do so.