928 karmaJoined Mar 2019


Cool, thanks for checking it out! I'll update the post slightly to make it clearer that I'm talking about beliefs rather than the truth.

Hi Tyler, thanks for your thoughts on this! Note that this post is not about the best philosophical objections, it's about what EAs actually believe. I have spoken to many EAs who say they are utilitarian but don't believe in objective moral truth (or think that objective moral truth is very unlikely) and what I'm responding to in this post is what those people say about what they believe and why. I also have spoken to Jeff Sebo about this as well! 


In point 1 and 2 in this post, namely, "1. I think (in one sense) it’s empirically false to say that "only utility is valuable" and "2. It can be psychologically harmful to deny your intrinsic values" I'm making a claim about human psychology, not about philosophy.

So it sounds like you're mainly respoding to this point in my essay: "3. I think (in one sense) it’s incoherent to only value utility if you don't believe in moral realism"

That's totally fair, but please note that I actually solicited feedback on that point from tons of people, including some philosophers, and I wrote a whole essay just about that claim in particular which is linked above (of course, I wouldn't have expected you to have read it, but I'm just pointing that out).  Here is that essay:


I will update the post slightly to make it clearer that I have a whole essay discussing objections to that point.

Note that at the bottom of that other essay I discuss every single reasonable-ish objection I've heard to that argument, including some from philosophers. Perhaps you have other objections not mentioned there, but I do delve into various objections and have sought a lot of feedback/criticism on that already!

I would be very happy to discuss this topic more with you, and hear your more detailed objections/points you think I am getting wrong - let me know if you'd like to do that! 


Finally, I will note though that most of the objections you mention in your comment above are NOT the many EAs I've spoken to use to defend their beliefs, so even if they are strong arguments I don't think they are doing work in why most EAs (who deny objective moral truth but say they are utilitarians) believe what they do.

Hi Rebecca. To clarify: that’s not what I’m saying. What I’m saying is that in the version Ben showed me hours before publication none of the disparaging Glassdoor comments he used in the post (that he claimed were all about Emerson) were actually about Emerson. He has acknowledged this point. Based on me pointing this out, he hastily fix these mistakes before releasing the public version, hence you won’t find this error in the version of his post above. I use this as an example of just one of a number of what I see as important errors (based on the evidence I have access to) in his draft that was shared with me right before publishing, which made me fear his research was done in a biased, sloppy and/or error prone way, with (from my point of view) not enough care being taken to avoid making false harmful claims.

Thanks Tiresias for your thoughtful comments. I agree with much of what you say but I seemingly have a few important differences of opinion:

"I disagree with holding all misconduct reports to incredibly high standards, such that in a report with as many allegations as this, people feel the report is basically wrong if it includes a few misinterpretations. In an ideal world, yes, all summaries of patterns of misconduct would not contain any errors. But in reality, I've found that almost all allegations of behaviors that turn out to be -- for all intents and purposes -- true, contain some level of mistakes, misattributions, specific allegations that are overstated."

I agree. I don't think I was holding the report to an incredibly high standard though. When I read it I was immediately chagrined by the amount and severity of false information (i.e., false as far as I can tell based on the evidence I have access to). I was also distressed that Ben was not seeking out evidence he could have easily gotten from nonlinear.

"People who allege misconduct are under intense scrutiny. And absolutely, scrutiny is warranted. But as someone who has reported misconduct and spoken to other people that report misconduct, the expectation of perfection is, to put it mildly, chilling. It means people do not come forward, it means people who do come forward are further traumatized, it means allegations that are 80% truthful are dismissed outright."

Good point. I would differentiate between the standard for people privately reporting bad behavior (where I think the bar should be way lower) and large scale investigations that are made public (where I think the bar should be much higher for the claims made - e.g., that the investigator should be very careful not to credulously include damaging false information).

"Does a third or more of what Ben wrote comport with your general understanding? If so, these allegations are still concerning to me."

I think this framing doesn't quite work because the post contains some very minor concerns and some very major ones, and I think it's much more important whether the major concerns are accurate than that the minor concerns are accurate, so counting up the number of inaccuracies doesn't, I think, reflect what's important. But based on the evidence I've seen, some of the damning claims in his original post seemed to me to be false or missing critical context that make them very misleading.

"And on the Kat screenshots/food question, I do not think they delegitimize what Ben wrote here. At worst, Ben somewhat overstated the food situation. But, my overall impression from those screenshots was what Alice said was basically true. Kat's framing of what the screenshots say make me doubt Kat's account more, not less."

I think people should decide for themselves what they think is true about this after reviewing the evidence. Here is a side-by-side comparison of what Ben says and what Kat says:

Ben: "Alice claims she was sick with covid in a foreign country, with only the three Nonlinear cofounders around, but nobody in the house was willing to go out and get her vegan food, so she barely ate for 2 days."

Kat: "1. There was vegan food in the house (oatmeal, quinoa, mixed nuts, prunes, peanuts, tomatoes, cereal, oranges) which we offered to cook for her. 2. We did pick up vegan food for her."

And here are the screenshots Kat provided to back up her account: https://forum.effectivealtruism.org/posts/5pksH3SbQzaniX96b/a-quick-update-from-nonlinear

"I'll also say as someone who has experienced harassment, that people really underestimate how much bias they have towards their friends accused of misconduct."

Absolutely agreed, this is a significant issue to watch out for.

Just to clarify, nonlinear has now picked one claim and provided screen shots relevant to it, I’m not sure if you saw that.

I also want to clarify that I gave Ben a bunch of very specific examples of information in his post that I have evidence are false (responding to the version he sent me hours before publication). He hastily attempted to adjust his post to remove or tweak some of his claims right before publishing based on my discussing these errors with him. It’s a lot easier (and vastly less time consuming) to provide those examples in a private one-on-one with Ben than to provide them publicly (where, for instance, issues of confidentially become much more complicated, and where documentation and wording need to be handled with extreme care, quite different than the norms of conversation).

The easiest to explain example is that Ben claimed a bunch of very bad sounding quotes from Glassdoor were about Emerson that clearly weren’t (he hadn’t been at the company for years when those complaints were written). Ben acknowledged somewhere in the comments that those were indeed not about Emerson and so that was indeed false information in the original version of the post.

My understanding, trying to interpret Ben’s comments on this point (if I’m mistaken, please correct me of course), is that Ben thinks it’s not a big deal that he almost included these false claims about Emerson (and would have had I not pointed it out right before publication) because he doesn’t view these as cruxy for his own personal hypotheses.

On the other hand, I view it as a very big deal to make severely negative, public, false claims about another person, and to me this one example is indicative of the process used to generate the post - a process that, from my point of view based on the evidence I’ve seen, led the post to contain a bunch of false claims.

Of course Ben didn’t purposely say anything he knew to be false, but I think Ben and I have different opinions on how bad it is to make public, false, potentially damaging claims about people, and the standard of care/evidence required before making those claims.

Nonlinear says they will provide lots more specific examples in the coming days of what they see as misinformation in the post - of course it will be up to you to judge whether you find their evidence convincing.

From my point of view, it’s best to reserve judgment until the evidence is released, assuming they do it within a reasonable time frame (e.g., a week or two - if they failed to release the evidence promptly that would be another matter).

I see no reason to jump to conclusions or take sides before we’ve seen all the evidence since it sounds like we’ll have access to it very soon.

Yes, here two examples, sorry I can’t provide more detail:

-there were claims in the post made about Emerson that were not actually about Emerson at all (they were about his former company years after he left). I pointed this out to Ben hours before publication and he rushed to correct it (in my view it’s a pretty serious mistake to make false accusations about a person, I see this as pretty significant)!

-there was also a very disparaging claim made in the piece (I unfortunately can’t share the details for privacy reasons; but I assume nonlinear will later) that was quite strongly contradicted by a text message exchange I have


Hi all, I wanted to chime in because I have had conversations relevant to this post with just about all involved parties at various points. I've spoken to "Alice" (both while she worked at nonlinear and afterward), Kat (throughout the period when the events in the post were alleged to have happened and afterward), Emerson, Drew, and (recently) the author Ben, as well as, to a much lesser extent, "Chloe" (when she worked at nonlinear). I am (to my knowledge) on friendly terms with everyone mentioned (by name or pseudonym) in this post. I wish well for everyone involved. I also want the truth to be known, whatever the truth is.

I was sent a nearly final draft of this post yesterday (Wednesday), once by Ben and once by another person mentioned in the post.

I want to say that I find this post extremely strange for the following reasons:

(1) The nearly final draft of this post that I was given yesterday had factual inaccuracies that (in my opinion and based on my understanding of the facts) are very serious despite ~150 hours being spent on this investigation. This makes it harder for me to take at face value the parts of the post that I have no knowledge of.  Why am I, an outsider on this whole thing, finding serious errors in the final hours before publication? That's not to say everything in the post is inaccurate, just that I was disturbed to see serious inaccuracies, and I have no idea why nobody caught these (I really don't feel like I should be the one to correct mistakes, given my lack of involvement, but it feels important to me to comment here since I know there were inaccuracies in the piece, so here we are).

(2) Nonlinear reached out to me and told me they have proof that a bunch of claims in the post are completely false. They also said that in the past day or so (upon becoming aware of the contents of the post), they asked Ben to delay his publication of this post by one week so that they could gather their evidence and show it to Ben before he publishes it (to avoid having him publish false information). However, he refused to do so.

This really confuses me. Clearly, Ben spent a huge amount of time on this post (which has presumably involved weeks or months of research), so why not wait one additional week for Nonlinear to provide what they say is proof that his post contains substantial misinformation? Of course, if the evidence provided by nonlinear is weak, he should treat it as such, but if it is strong, it should also be treated as such. I struggle to wrap my head around the decision not to look at that evidence. I am also confused why Ben, despite spending a huge amount of time on this research, apparently didn't seek out this evidence from Nonlinear long ago.

To clarify: I think it’s very important in situations like this not to let the group being criticized have a way to delay publication indefinitely. If I were in Ben’s shoes, I believe what I would have done is say something like, “You have exactly one week to provide proof of any false claims in this post (and I’ll remove any claim you can prove is false) then I’m publishing the post no matter what at that time.” This is very similar to the policy we use for our Transparent Replications project (where we replicate psychology results of publications in top journals), and we have found it to work well. We give the original authors a specific window of time during which they can point out any errors we may have made (which is at least a week). This helps make sure our replications are accurate, fair, and correct, and yet the teams being replicated have no say over whether the replications are released (they always are released regardless of whether we get a response).

It seems to me that basic norms of good epistemics require that, on important topics, you look at all the evidence that can be easily acquired.

I also think that if you publish misinformation, you can't just undo it by updating the post later or issuing a correction. Sadly, that's not the way human minds/social information works. In other words, misinformation can't be jammed back into the bottle once it is released. I have seen numerous cases where misinformation is released only later to be retracted, in which the misinformation got way more attention than the retraction, and most people came away only with the misinformation. This seems to me to provide a strong additional reason why a small delay in the publication date appears well worth it (to me, as an outsider) to help avoid putting out a post with potentially substantial misinformation. I hope that the lesswrong/EA communities will look at all the evidence once it is released, which presumably will be in the next week or so, in order to come to a fair and accurate conclusion (based on all the evidence, whatever that accurate final conclusion turns out to be) and do better than these other cases I’ve witnessed where misinformation won the day.

Of course, I don't know Ben's reason for jumping to publish immediately, so I can't evaluate his reasons directly.

Disclaimer: I am friends with multiple people connected to this post.  As a reminder, I wish well for everyone involved, and I wish for the truth to be known, whatever that truth happens to be. I have acted (informally) as an advisor to nonlinear (without pay) - all that means, though, is that every so often, team members there will reach out to me to ask for my advice on things.

Note: I've updated this comment a few times to try to make my position clearer, to add some additional context, and to fix grammatical mistakes.

The way you define values in your comment:

"From the AI "engineering" perspective, values/valued states are "rewards" that the agent adds themselves in order to train (in RL style) their reasoning/planning network (i.e., generative model) to produce behaviours that are adaptive but also that they like and find interesting (aesthetics). This RL-style training happens during conscious reflection."

is just something different than what I'm talking about in my post when I use the phrase "intrinsic values." 

From what I can tell, you seem to be arguing:


[paraphrasing] "In this one line of work, we define values this way", and then jumping from there to "therefore, you are misunderstanding values," when actually I think you're just using the phrase to mean something different than I'm using it to mean. 

Preference utilitarianism and valuism don't have much in common.

Preference utilitarianism: maximize the interests/preferences of all beings impartially.

First, preferences and intrinsic values are not the same thing. For instance, you may have a preference to eat Cheetos over eating nachos, but that doesn't mean you intrinsically value eating Cheetos or that eating Cheetos necessarily gets you more of what you intrinsically value than eating nachos will. Human choice is driven by a lot of factors other than just intrinsic values (though intrinsic values play a role).

Second, preference utilitarianism is not about your own preferences, it's about the preferences of all beings impartially.

Glad you find it interesting! We tested maybe about 150 statements. Just to clarify, it's not that the depression side doesn't correlate with the anxiety side - because anxiety and depression are so correlated, any statement correlating with one is likely to correlate with the other. But when you statistically separate them (i.e., you look at what correlates with anxiety when you've controlled for depression, or the reverse), this clearer picture emerges). 


While it would be great for someone to replicate these findings (to increase confidence in them), and I hope someone does that, the sample size (n=500) is fine in my view for this kind of result. There is diminishing benefit to larger sample sizes (the right sample size depends on the analyses being performed and the level of noise). So 10,000 people isn't as much better than 500 people as it may sound.

For instance, at n=500 a measured correlation of r=0.50 has a 95th percentile confidence interval of r=0.43 to r=0.56. At n=10,000 it's r=0.49 to r=0.51. For many purposes the latter isn't much more useful than the former. See this confidence interval calculation tool for correlations for more details: http://vassarstats.net/rho.html

Load more