So a lab employee recently wrote a tweet about how they spent the night at Lighthaven and, surprisingly(!), not a single person threatened them with violence for working at a lab.
I asked if they were being tongue-in-cheek. They were not. This was noteworthy enough to them that it needed to be remarked upon, because they expected their x followers to doubt it if it weren’t emphasized.
This feels related to the recent Oliver / Scott interaction… it seems pretty clear that something is going HORRIBLY WRONG and I can’t figure out what it is.
edit: Do the downvotes mean this is actually an appropriate concern for lab employees to have? They should feel unwelcome at lighthaven? Maybe personal violence is not as obviously-net-negative as I thought the consensus treated it?
If the labs think that ours is a philosophy of tribal violence then… well, we’re going to alienate people like Scott Alexander, whose Dwarkesh appearance I treated as a very positive event for the timeline and who I think really needs not to be alienated!
At first I thought the accelerationist folk were projecting weird tendencies onto the lesswrong crowd that weren’t there. This post from Scott got me worrying that maybe I was wrong about this: https://x.com/slatestarcodex/status/2043428604818989401 that maybe there was a real problem here.
Theoretically we don’t need to bring around any actual lab employees to our way of thinking, I guess. We can just convince enough of the public to get legislation passed, and then leverage that into an international treaty. But it still feels like a dumb unforced error to negatively polarize anyone who works at a lab against the entire safety memeplex. If we burn the lightcone because some lab employee notices troubling signs during an RSI loop but decides not to speak up because they don’t want to “side with Holly” or whatever… that does not feel like a very dignified way to go.
Do the downvotes mean this is actually an appropriate concern for lab employees to have? They should feel unwelcome at lighthaven? Maybe personal violence is not as obviously-net-negative as I thought the consensus treated it?
After reading your shortform, I have also now downvoted it. I would describe this as “hard-to-follow ramblings about what people said on twitter and how people should update on what other people think without either being very clear or otherwise grounding itself in anything that actually happened outside of people’s failures to communicate.” Like, you didn’t even link to the tweet about someone visiting Lighthaven (which I had to spend ~5 mins searching for to understand the context), nor do you link to the Habryka/Scott conflict, so you’re imposing a fair bit of cost on a lot of people to follow this discussion.
This isn’t terrible but I think it’s pretty confusing to read, and doesn’t make up for that by adding much of value.
I have written this to add information about why you were downvoted.
the former was deliberate, trying to be courteous of privacy requests and avoid directing hostility their way, so if the post is making people search for their tweet my instinct is to hide my post rather than add a citation
fair point re: rambling. I am not sure how to feel about whatever exactly I’m pointing at, or even if the thing-I-am-pointing-at is real. That’s part of why I’m trying to elicit feedback. There’s currently a bunch of discourse on Twitter/X akin to Scott’s post, about the degree of hostility that the lesswrong crowd directs towards the labs (Andrew Critch is probably the most recent example https://x.com/AndrewCritchPhD/status/2048275817256702364). I was more confident that it was just flavor-of-the-month drama before reading the tweet from Scott that I linked above. But I don’t know what to do about that.
Yeah that’s a difficult tradeoff. It is nice to not accidentally brigade someone else’s social media space, but also makes it hard for readers to tell if you’ve correctly assessed or described what’s going on.
Also I’m sorry about my comment, it was needlessly harsh. It is a common annoyance for people to complain about or read a lot into downvotes that isn’t there, and I wanted to push back on that, but I could’ve been kinder in how I did it. I am very sleepy and shouldn’t be writing comments right now.
edit: Do the downvotes mean this is actually an appropriate concern for lab employees to have? They should feel unwelcome at lighthaven? Maybe personal violence is not as obviously-net-negative as I thought the consensus treated it?
I don’t know why you were downvoted but I have no idea how you got from “one person downvoting you” to that.
This feels related to the recent Oliver / Scott interaction… it seems pretty clear that something is going HORRIBLY WRONG and I can’t figure out what it is.
I don’t know what exact words that lab employee said to you (I assume this was in DMs, since I saw the tweet you’re referring to), but I do not think they are well-calibrated about their coworkers and their expectations on that subject.
Ah, I didn’t actually mouse-over it, and assumed going from +1 to −1 required 2 votes. I guess the automatic “self-upvote” is sorta fake, or something? I definitely overreacted though, sorry for that.
What do you think the mood is, inside the labs, regarding the safetyist community? I’m worried that the trend of, for instance, Anthropic researchers (and even the occasional OpenAI researcher!) crossposting results to alignmentforum and lesswrong is a really good example of across-the-aisle cooperation, and might be at risk if the employee’s calibration is accurate.
It’s hard to judge things like this from a distance, I tend to only come to the bay once or twice a year. But Scott’s post about the hostility he perceived towards lab employees has me worried. He’s pretty alert to the whole toxoplasmosis reverse polarization thing, being the tropenamer and all and having experiences with it himself. If he thinks there’s genuine problems in safetyist attitudes towards the labs, it makes me take notice.
But maybe if I lived in the bay and regularly hung out with the actual individuals involved I would know there’s nothing to worry about?
So a lab employee recently wrote a tweet about how they spent the night at Lighthaven and, surprisingly(!), not a single person threatened them with violence for working at a lab.
I asked if they were being tongue-in-cheek. They were not. This was noteworthy enough to them that it needed to be remarked upon, because they expected their x followers to doubt it if it weren’t emphasized.
This feels related to the recent Oliver / Scott interaction… it seems pretty clear that something is going HORRIBLY WRONG and I can’t figure out what it is.
edit: Do the downvotes mean this is actually an appropriate concern for lab employees to have? They should feel unwelcome at lighthaven? Maybe personal violence is not as obviously-net-negative as I thought the consensus treated it?
If the labs think that ours is a philosophy of tribal violence then… well, we’re going to alienate people like Scott Alexander, whose Dwarkesh appearance I treated as a very positive event for the timeline and who I think really needs not to be alienated!
At first I thought the accelerationist folk were projecting weird tendencies onto the lesswrong crowd that weren’t there. This post from Scott got me worrying that maybe I was wrong about this: https://x.com/slatestarcodex/status/2043428604818989401 that maybe there was a real problem here.
Theoretically we don’t need to bring around any actual lab employees to our way of thinking, I guess. We can just convince enough of the public to get legislation passed, and then leverage that into an international treaty. But it still feels like a dumb unforced error to negatively polarize anyone who works at a lab against the entire safety memeplex. If we burn the lightcone because some lab employee notices troubling signs during an RSI loop but decides not to speak up because they don’t want to “side with Holly” or whatever… that does not feel like a very dignified way to go.
After reading your shortform, I have also now downvoted it. I would describe this as “hard-to-follow ramblings about what people said on twitter and how people should update on what other people think without either being very clear or otherwise grounding itself in anything that actually happened outside of people’s failures to communicate.” Like, you didn’t even link to the tweet about someone visiting Lighthaven (which I had to spend ~5 mins searching for to understand the context), nor do you link to the Habryka/Scott conflict, so you’re imposing a fair bit of cost on a lot of people to follow this discussion.
This isn’t terrible but I think it’s pretty confusing to read, and doesn’t make up for that by adding much of value.
I have written this to add information about why you were downvoted.
the former was deliberate, trying to be courteous of privacy requests and avoid directing hostility their way, so if the post is making people search for their tweet my instinct is to hide my post rather than add a citation
fair point re: rambling. I am not sure how to feel about whatever exactly I’m pointing at, or even if the thing-I-am-pointing-at is real. That’s part of why I’m trying to elicit feedback. There’s currently a bunch of discourse on Twitter/X akin to Scott’s post, about the degree of hostility that the lesswrong crowd directs towards the labs (Andrew Critch is probably the most recent example https://x.com/AndrewCritchPhD/status/2048275817256702364). I was more confident that it was just flavor-of-the-month drama before reading the tweet from Scott that I linked above. But I don’t know what to do about that.
Yeah that’s a difficult tradeoff. It is nice to not accidentally brigade someone else’s social media space, but also makes it hard for readers to tell if you’ve correctly assessed or described what’s going on.
Also I’m sorry about my comment, it was needlessly harsh. It is a common annoyance for people to complain about or read a lot into downvotes that isn’t there, and I wanted to push back on that, but I could’ve been kinder in how I did it. I am very sleepy and shouldn’t be writing comments right now.
I don’t know why you were downvoted but I have no idea how you got from “one person downvoting you” to that.
I don’t know what exact words that lab employee said to you (I assume this was in DMs, since I saw the tweet you’re referring to), but I do not think they are well-calibrated about their coworkers and their expectations on that subject.
Ah, I didn’t actually mouse-over it, and assumed going from +1 to −1 required 2 votes. I guess the automatic “self-upvote” is sorta fake, or something? I definitely overreacted though, sorry for that.
What do you think the mood is, inside the labs, regarding the safetyist community? I’m worried that the trend of, for instance, Anthropic researchers (and even the occasional OpenAI researcher!) crossposting results to alignmentforum and lesswrong is a really good example of across-the-aisle cooperation, and might be at risk if the employee’s calibration is accurate.
It’s hard to judge things like this from a distance, I tend to only come to the bay once or twice a year. But Scott’s post about the hostility he perceived towards lab employees has me worried. He’s pretty alert to the whole toxoplasmosis reverse polarization thing, being the tropenamer and all and having experiences with it himself. If he thinks there’s genuine problems in safetyist attitudes towards the labs, it makes me take notice.
But maybe if I lived in the bay and regularly hung out with the actual individuals involved I would know there’s nothing to worry about?