All of Emerson Spartz's Comments + Replies

@Ben Pace  Can you please add at the top of the post "Nonlinear disputes at least 85 of the claims in this post and intends to publish a detailed point-by-point response.

They also published this short update giving an example of the kind of evidence they plan to demonstrate."

We keep hearing from people who don't know this. Our comments get buried, so they think your summary at the bottom contains the entirety of our response, though it is just the tip of the iceberg. As a result, they think your post marks the end of the story, and not the opening cha... (read more)

4Adam Zerner6mo
I think it would be helpful to mention some sort of rough estimate of how many of those claims Nonlinear believes to be important and cruxy. 70 of them? 20? 5? Separately, I think it would be helpful to focus on the ones that are important and cruxy. As Kat mentioned, it can take many hours to dispute any one claim. It seems wise to focus on the important ones rather than getting lost in the weeds debating and digging through the evidence for the unimportant ones.
5Ben Pace6mo
I've left an edit at the top.
I think it would be good to word this as "and intends to publish a detailed point-by-point response by September 15th," or whatever the correct date turns out to be.

"Alice quit being vegan while working there. 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."

Seems like other people besides Ruby are confused about this too, maybe also because Ben sometimes says "the Nonlinear cofounders" and Emerson/Kat/Drew

A source of terminological confusion here is that Lightcone team often internally uses the word cofounder to mean ‘person with a very strong investment and generalist skill set, who takes responsibility in a particular way’. Ie have used it to refer to multiple people on the Lightcone team who didn’t literally found the org but are pretty deeply involved. The crux for me with Drew, and I assume with Ruby/Ben, is ‘does he have that kind of relationship with the org?’, rather than ‘did he literally cofound the org’. I do think this terminology is probably confusing for other readers, and seems good to correct, although I would guess not actually misleading in an way that’s particularly relevant for most people's assessment of the situation.

Just FYI Drew is not a cofounder of Nonlinear. That is another inaccurate claim from the article. He did not join full time until April 2022.

Which part of the post claims that? The post seems to say the opposite:  There might be another part that does refer to Drew as a co-founder, but I can't find anything of that sort.

I'd like to kindly remind you that you are making a lot of judgments about my character based on a 10,000 word post written by someone who explicitly told you he was looking for negative information and only intended to share the worst information. 

That is his one paragraph paraphrase of a very complex situation and I think it's fine as far as it goes but it goes nowhere near far enough.  We have a mega post coming ASAP.

Ben has also been quietly fixing errors in the post, which I appreciate, but people are going around right now attacking us for ... (read more)

Ben has also been quietly fixing errors in the post, which I appreciate, but people are going around right now attacking us for things that Ben got wrong, because how would they know he quietly changed the post?

This is why every time newspapers get caught making a mistake they issue a public retraction the next day to let everyone know. I believe Ben should make these retractions more visible.

I used a diff checker to find the differences between the current post and the original post. There seem to be two:

  1. "Alice worked there from November 2021 to June 2022
... (read more)

I definitely think Ben should be flagging anywhere in the post that he has made edits.

7Adam Zerner6mo
Here I am only making one judgement. I agree that the evidence isn't perfect, but even after accounting for that, I still feel reasonably confident in my suspicion. I am basing my judgement off of much more than that paragraph. I don't think that saying "X lacks the skill of being able to lose" is an attack on X's character. Maybe slightly, but not substantially. As discussed elsewhere, I don't think the fact that Nonlinear claims they have evidence of errors means that the conversation needs to be postponed. I think it simply means that we should update our beliefs when the new evidence becomes available. (Yes, humans are biased against doing this well.) Strongly agreed.

This is more false info.  The approximate/expected total compensation was $70k which included far more than room and board and $1k a month.  

Chloe has also been falsely claiming we only had a verbal agreement but we have multiple written records.  

We'll share specifics and evidence in our upcoming post.

Just FYI, the original claim is a wild distortion of the truth.  We'll be providing evidence in our upcoming post.

Rob Bensinger replied with:

I think that there's a big difference between telling everyone "I didn't get the food I wanted, but they did get/offer to cook me vegan food, and I told them it was ok!" and "they refused to get me vegan food and I barely ate for 2 days".


And this:

This also updates me about Kat's take (as summarized by Ben Pace in the OP):

> Kat doesn’t trust Alice to tell the truth, and that Alice has a history of “catastrophic misunderstandings”.

When I read the post, I didn't see any particular reason for Kat to think this, and I worrie

... (read more)

Thanks for sharing your thoughts!

1) I disagree with that interpretation.

2) She was not an employee of Nonlinear at the time, just a friend. Ben's post said she was, but that was a factual inaccuracy, one of many we are working hard to correct in our forthcoming post.

Just to clear up a view things:

  1. It was $70k in approximate/expected total compensation. The $1k a month was just a small part of the total compensation package.
  2. Despite false claims to the contrary, it wasn't just verbally agreed, we have written records.
  3. Despite false claims to the contrary, we were roughly on track to spend that much. This is another thing we will show evidence for ASAP, but there is a lot of accounting/record keeping etc to do to organize all the spending information, etc.

Yes, that is what I intended to communicate here, and I was worried people might think I was trying to suppress the article so I bolded this request to ensure people didn't misunderstand:

For what it's worth, I also interpreted the "if published as is" as "if you do not edit the post to no longer be libelous" and not "if you do not give us a week to prepare a contemporaneous rebuttal".

I think if you wanted to reliably communicate that you were not asking for changes to the text of the post, you would have needed to be explicit about that?

In case it wasn't clear,  we didn't say 'don't publish', we said 'don't publish until we've had a week to gather and share the evidence we have':

3Martin Randall6mo
Please don't post screenshots of comments that include screenshots of comments. It is harder to read and to search and to reply. You can just quote the text, like habryka did above.

I'm trying to support two complementary points:

  • The norm I've been pushing of sharing things with EA organizations ahead of time is only intended for cases where you have a neutral or better relationship with the organization, and not situations like this one where there are allegations of mistreatment, or you don't trust them to behave cooperatively.

  • A threat to sue if changes are not made to the text of the post is not cooperative.

You say "if published as is", not "if published now". Is what you're saying in the comment that, if Ben had waited a week and then published the same post, unedited, you would not want to sue? That is not what is conveyed in the email.

Personally, I think it's correct to update somewhat, but in situations like this where only one side has shared their perspective, I'm much more likely to overupdate ("those monsters!") so I have to guard against that.

Thinking about "situations like this" does sound like it could be helpful. Some come to mind, and caveat that it's hard to remember how I felt at different points in time but: * Case one: if the accused ever gave their perspective, I don't remember it. * Case two: the accused sharing their perspective initially made me more sympathetic to them, but that that was a mistake on my part because it turned out to be full of lies. * Case three: the accused sharing their perspective made me less sympathetic to them. * Case four: I dismissed the accusations offhand and think I was right to do so. So this is weak evidence, but I don't feel like I personally have a history of overupdating in the direction of "those monsters".

I did hear your side for 3 hours and you changed my mind very little and admitted to a bunch of the dynamics ("our intention wasn't just to have employees, but also to have members of our family unit") and you said my summary was pretty good. You mostly laughed at every single accusation I brought up and IMO took nothing morally seriously and the only ex ante mistake you admitted to was "not firing Alice earlier". You didn't seem to understand the gravity of my accusations, or at least had no space for honestly considering that you'd seriously hurt and int... (read more)

Indeed, without context that is a cartoon villain thing to say. Not asking you to believe us, yet just asking you to withhold judgment until you've seen the evidence we have which will make that message seem very different in context.

Was the follow-up promised here ever produced?
How complicated is providing context for that without a week of work on your side? The only plausible exculpatory context I can imagine is something akin to: "If somebody sent me a text like this, I would sever all contact with them, so I'm providing it as an example of what I consider to be unacceptable." I fail to see how hard it is to explain why the claims are false now and then provide detailed receipts within the week. I don't know any of the parties involved here, but the Nonlinear side seems pretty fishy so far.

We were very clear that we felt there were still major issues to address. Here’s another email in the thread a day later:

We also clearly told Ben and Robert in the call many times that there is a lot more to the story, and we have many more examples to share. This is why we suggested writing everything up, to be more precise and not say anything that was factually untrue. Since our former employees’ reputations are on the line as well, it makes sense to try to be very deliberate. 

It's possible there was a miscommunication between you and Ben around ho... (read more)

Honestly, one of the reasons I don't find the Nonlinear narrative credible is the absolute 100% denial of any wrongdoing, a 0% reflection. Clearly, Ben really looked into this and has various accounts from multiple people or really questionable behavior, that seems very credible and to come against all of it with such force and conviction is a tactic of people who want to deny and distort the truth. 

Wait, just so I understand, what I thought happened was that Ben sent you the summary before a call, to which you sent the first email (saying "good summary"). 

Then Ben said that he planned to publish this whole post and shared you on a draft, at which point you sent the email screenshotted in your most recent reply. They are responding to totally different pieces of text.

I absolutely agree that you clearly communicated that you think the full post is full of inaccuracies, but we were talking about whether the specific summary that Ben shared with you... (read more)

There is a reason courtrooms give both sides equal chances to make their case before they ask the jury to decide. 

It is very difficult for people to change their minds later, and most people assume that if you’re on trial, you must be guilty, which is why judges remind juries about “innocent before proven guilty”. 

This is one of the foundations of our legal system, something we learned over thousands of years of trying to get better at justice. You’re just assuming I’m guilty and saying that justifies not giving me a chance to present my evi... (read more)

0Adam Zerner6mo
You seem to be disregarding other considerations at play here. Zooming out, if we forget about the specifics of this situation and instead think about the more general question of whether or not one should honor requests to delay such publications, one consideration is wanting to avoid unjustifiably harming someones reputation (in this case yours, Kat's, and Nonlinear's). But I think habryka lists some other important considerations too in his comment: * Guarding against retaliation * Guarding against lost productivity * Guarding against reality-distortion fields Personally, I don't have strong feelings about where the equilibrium should be here. However, I do feel strongly that the discussion needs to look at the considerations on both sides. Also, I raise my eyebrow a fair bit at those who do have strong feelings about where the equilibrium should be. At least if they haven't thought about it for many hours. It strikes me as a genuinely difficult task to enumerate and weigh the considerations at play.
4Adam Zerner6mo
I don't get that impression. Nothing in the full one stands out to me as important context that would really change anything non-trivially.

I want to note a specific pattern that I've noticed. I am not commenting on this particular matter overall; the events with Nonlinear may or may not be an instance of the pattern. It goes like this:

  1. Fred does something unethical / immoral.
  2. People start talking about how Fred did something bad.
  3. Fred complains that people should not be talking the way they are talking, and Fred specifically invokes the standard of the court system, saying stuff like "there's a reason courts presume innocence / allow the accused to face the accuser / give a right to a defens
... (read more)

Also, if we post another comment thread a week later, who will see it? EAF/LW don’t have sufficient ways to resurface old but important content. 

This doesn't seem like an issue. You could instead write a separate post a week later which has a chance of gaining traction.

People give standing ovations when they feel inspired to because something resonated with them.  They're applauding him for trying to save humanity, and this audience reaction gives me hope.

It's unfortunate that this version is spreading because many people will think it's a low credibility TEDx talk instead of a very credible main stage TED talk.


My background is extremely relevant here and if anybody in the alignment community would like help thinking through strategy, I'd love to be helpful.

^ can confirm! I volunteered for Spartz's Nonlinear org a couple years ago, and he has a long history of getting big numbers on social media.
4Seth Herd1y
We need some expertise on this topic. I actually just wrote a post on exactly that point, yesterday. I also include some strategy ideas, and I'm curious if you agree with them. I think a post from you would probably be useful if you have relevant expertise; few people here probably do. I'd be happy to talk.

I think this is a really promising idea.

If the goal is to unify diverse stakeholders, including non-technical ones, I wonder if it would be better to use a less-wonky target (e.g. "50%" instead of ".002 OOMs")

Going to share a seemingly-unpopular opinion and in a tone that usually gets downvoted on LW but I think needs to be said anyway:

This stat is why I still have hope: 100,000 capabilities researchers vs 300 alignment researchers.

Humanity has not tried to solve alignment yet.

There's no cavalry coming - we are the cavalry.

I am sympathetic to fears of a new alignment researchers being net negative, and I think plausibly the entire field has, so far, been net negative, but guys, there are 100,000 capabilities researchers now! One more is a drop in the bucket.

If ... (read more)

Another example of Overton movement - imagine seeing these results a few years ago:

I wonder if soon the general public will freak out on a large scale (Covid-like). I will be not surprised if it will happen in 2024 and only slightly surprised if it will happen this year. If it will happen, I am also not sure if it will be good or bad.

"I'm an accelerationist for solar power, nuclear power to the extent it hasn't been obsoleted by solar power and we might as well give up but I'm still bitter about it, geothermal, genetic engineering, neuroengineering, FDA delenda est, basically everything except GoF bio and AI"

I also think this approach deserves more consideration.

Also: since BCIs can generate easy-to-understand profits, and are legibly useful to many, we could harness market forces to shorten BCI timelines.

Ambitious BCI projects will likely be more shovel ready than many other alignment approaches - BCIs are plausibly amenable to Manhattan Project-level initiatives where we unleash significant human and financial capital. Maybe use Advanced Market Commitments to kickstart the innovators, etc.

For anybody interested, Tim Urban has a really well written post about Neuralink/BCIs:

Anecdata: many in my non-EA/rat social circles of entrepreneurs and investors are engaging with this for the first time.  

And, to my surprise (given the optimistic nature of entrepreneurs/VCs) they aren't just being reflexive techno-optimists, they're taking the ideas seriously and, since Bankless, "Eliezer" is becoming a first name-only character.

Eliezer said he's an accelerationist in basically everything except AI and gain-of-function bio and that seems to resonate. AI is Not Like The Other Problems.

2Lone Pine1y
Curious, is he accelerationist in atomically precise manufacturing?

Came here to say this. Highly recommend this book for anyone working on deception.

Love this! I used to manage teams of writers/editors and here are some ideas we found useful for increasing readability:

To remove fluff, imagine someone is paying you $1,000 for every word you remove.  Our writers typically could cut 20-50% with minimal loss of information.

Long sentences are hard to read, so try to change your commas into periods. 

Long paragraphs are hard to read, so try to break each paragraph into 2-3 sentences.

Most people just skim, and some of your ideas are much more important than others, so bold/italicize your important points.

Thanks for the feedback! We think a bot could make sense as well - we're exploring this internally.

So glad you're enjoying it! It's mine too - I consume way more LW content because of it.

 I'd expect the most common failure mode for rationalists here is not understanding how patronage networks work. 

Even if you do everything else right, it is very hard to get elected to a position of power if the other guy is distributing the office's resources for votes.

You should be able to map out the voting blocs and what their criteria are, i.e. "Union X and its 500 members will mostly vote for Incumbent Y because they get $X in contracts per year etc"

The mapping of voting blocs seems like a really good idea, very actionable, and a great way to visualize who could be electing you. Putting their requirements, or encouragements out in a visual way, to weigh where the least action can cause the greatest gain. I think that the situation I'm considering has an intensely powerful patronage network that it can relatively easily attach itself to. Other patronage networks will also be necessary.

People are so irrationally intimidated by lawyers that some legal firms make all their money by sending out thousands of scary form letters demanding payment for bullshit transgressions.  My company was threatened with thousands of frivolous lawsuits but only actually sued once. 

Threats are cheap.