Ah perhaps I misunderstood you then—it sounds like this quote was specifically your own takeaway from reading Ben’s original article, rather than a characterization of the article itself. It’s possible that I’m seeing your position a bit better now—previously I thought you largely agreed with Ben’s article, but on another reread of your comment it seems that you generally hold significantly more moderate view on Nonlinear. (Although your other comment implies that you do believe “Ben’s account holds up”, so I remain confused.)
Vlad Firoiu
Karma: 151
Just read your comment again and there were a few things that I felt strong disagreement toward. One was you saying that
The stories from Chloe and Alice painted a picture of Non-Linear. A close, ambitious, high-stress, often renegotiated environment.”
This feels like a pretty big euphemism for Ben’s piece, which paints Nonlinear as cruel and abusive.
Thinking about this more, my guess is that by “uncollaborative” you were specifically referring to Nonlinear’s threat to file for libel against Ben. I agree you could call it that, but I don’t see it as disproportionate given the adversarial nature of Ben’s investigation and the massive cost it has had on Nonlinear. I’d be happy to hear your thoughts on this point.
I’m not really sure what we’re arguing at this point. My initial reply was about how collaborative Nonlinear had been, which I don’t think you’ve addressed and isn’t particularly related to whether Ben said true things. I’d also add that in my view Ben posting without getting Nonlinear’s side of the story was itself pretty uncollaborative, and so the “retaliation” against him (in the form of criticizing him for the way he wrote his post) to me seems entirely justified.
Ok, so it sounds like a crux for you is that Ben in fact had high confidence in what he was relaying from Alice being true. In a dispute like this I don’t think you can do very good due diligence when avoiding the people who are most likely to have counter-evidence; even if it is well-intentioned, it’s a sort of conscious confirmation bias. Ben sort of admits to using poor epistemics in his disclaimer (at the top of his original post) about how to update from reading his post, but doesn’t seem to update much on this himself (?), which seems like an error to me particularly when the stakes are this high. Perhaps it’s unnecessary, but I will also point out that deliberately using poor epistemics feels pretty contrary to the spirit of rationality, which for good reason has fought for truth and against poor epistemics.
(I further argue against the premise of the disclaimer and Ben posting without hearing both sides here).
Hm, a lot I disagree with here, but a crux is that I think you’re not really replying to TracingWoodgrain’s original point, which was that Ben knew there might be significant evidence contradicting much of his post but decided not to wait for it and published anyways (which TW considers to be a bad norm). Instead you seem to be changing frame to “did Ben publish anything which he knew for sure wasn’t true”, which is quite different, particularly in this case where evidence is deliberately not being looked at.
Just to clarify, I was specifically referring to untrue things that the employees said, not Ben (and likewise retaliation against the employees, not against Ben).
If the line you’re taking is that “Ben technically only relayed information given to him by Alice, while admitting that she might be unreliable”, I don’t think that’s very tenable. Publishing like that is implicitly an endorsement, and unlike you I suspect most people ignored the disclaimer, because it would be strange for someone to publish such damaging things that they actually weren’t sure were true. This comment I made on Ben’s original post also touches on this.
This squares very starkly in contrast with Nonlinear’s perception of things. It seems to me that all the work in your comment is being done by the “we did not believe them on” bit, which is very subjective and frankly would be ridiculous in something like fair trial—it would be like saying “the defense is not allowed to bring witnesses or make a case, because despite them claiming that they’ll make a strong case, we (the prosection) just don’t believe them”. You can argue about whether Nonlinear’s eventual response was satisfactory (though their evidence seems compelling to me), but I’m not seeing your case on this point in particular.
There’s a lot in here but I was immediately confused by “Nonlinear seem to move pretty quickly from collaborate mode to conflict mode. Quicker than I do, at least”. My understanding is that they were hearing about their ex-employees saying damaging untrue things for over a year but chose not to retaliate partly because they didn’t want to hurt their ex-employees’ reputations, until Ben forced their hand with his deliberately one-sided “Sharing Information” post. That sounds fairly (some might say overly) collaborative to me.
Edit: Here by “retaliate” I mean defending themselves in the way they did with this post, which does have the side effect of harming Alice and Chloe’s reputations. Even then, they purposefuly decided not to de-anonymize their employees, and have a section on how they don’t consider them to have had ill intentions.
A lot of people have been angry about these texts made by Kat towards Alice:
“Given your past behavior, your career in EA would be over in a few DMs, but we aren’t going to do that because we care about you”
“We’re saying nice things about you publicly and expect you will do the same moving forward”This sounds like a threat and it’s not how I would have worded it had I been in Kat’s shoes. However, I think it looks much more reasonable if you view it through the hypothesis that a) the bad things Alice is saying about Nonlinear are untrue and b) the bad things Kat has been holding off on saying about Alice are true. Basically, I think Kat’s position is that “If you [Alice] keep spreading lies about us, we will have to defend ourselves by countering with the truth, and unfortunately if these truths got out it would make you look bad (e.g. by painting you as dishonest). That’s why we’ve been trying to avoid going down this route, because we actually care about you and don’t want to hurt your reputation (so you can find jobs), so let’s both just say nice things about each other from now on and put this behind us.”. My sense is that Kat, out of fear that her reputation was being badly and unfairly damaged, emphasized the part where bad things happen to Alice in an attempt to get her to stop spreading misinformation. Again, while this isn’t how I’d have worded those messages, given this context I think it’s much more understandable than it might first seem.
Disclaimer: I’m friends with Kat and know some of her side of the story.
The second-to-last line of the proof is also very confusing. I think GPT-4 is using the same variable names to mean different variables, which leads it to say silly things like “the uniqueness of $y$ given $y$”.
On further reflection I agree that diminishing return are pretty important. One consequence of them is that there is effectively a cap P on the total positive utility in a given time T. That turns into a risk cap of P/C per time period T.
Ok, so I’m guessing your position is that a) you, having read Nonlinear’s reply, continue to believe that most of what Ben relayed from Alice was true, and b) if a few things turn out to be untrue it’s not a big deal because it doesn’t change the overall story, and in any case Ben admitted that Alice might be unreliable.
I’m not entirely sure how you weigh (a) and (b) but it makes more sense to me if your crux is (a), that most of Alice’s claims are true. For that, I’m not sure where to start; as far as I’ve seen they all seem to be false. I guess we could start with the claims about not being paid, e.g. from Ben’s high level overview:
Nonlinear has several rows in their overview table which contradict this account:
Alice “wasn’t getting paid” only due to her own rather strange mistakes, such as not logging her expenses or not checking her own bank account to see that the money was actually there.
Alice eventually got to choose her own salary.
Alice claimed to be making significant income from her side business.
Alice had much less involvement and ownership of “the organization” than she claimed, and was repeatedly informed of this (this section of the appendix is relevant).
Ben also admits that “[Alice] also had a substantial number of emergency health issues covered [by Nonlinear]”.
We could also talk about Alice’s accusations of not being fed vegan food or being forced to travel with illegal drugs. I’m not sure if this is what you meant by “grievous error” though—please let me know if I’m barking up the wrong tree.