A summary of the Nonlinear situation from a third-person perspective, by Ozy Brennan. I find it to be thoughtful, well-written, and well-researched enough to be part of the conversation. It was also posted on LessWrong.
Edit: full text included below (with permission from the author). I had to reformat a lot so assume all formatting errors are done by me.
About a week ago, I became one of a small, elite group—people who have read every single goddamn word of Nonlinear's Evidence: Debunking False and Misleading Claims. I read the appendix. I read the appendixes to the appendix. I have spent more time thinking about whether a complete stranger got a vegan burger than any person should.
If you just want to make up your own mind, I suggest reading The General Situation and referring to Dramatis Personae when you get confused who is who. If you like drama, buckle in, because some shit is going down.
Every blog post is a group effort, this one more than most. I would like to thank Manya, Clara, Elizabeth van Nostrand, Keller, Kira, Wolffy, azrael, and Elena for insightful discussion, answering my “is this normal?” questions, and showing me what they’d discovered in their own research. All opinions—and remaining mistakes—are my own.
Much of this post is speculative. I am sure I got things wrong. In edits, I caught several errors I made in my first draft, and I’m not sure I caught every error I made. And reading 70,000+ words of someone’s writing and drawing conclusions about them is inherently fraught: certainly people believe all kinds of wild things about me from my writing on the Internet. I apologize for any mistakes. There will be an edit log on this post under this paragraph.
How We Got Here
Nonlinear is an effective altruist organization that was founded in 2021 by Kat Woods and Emerson Spartz. It runs a bunch of small projects. Representative examples include a text-to-speech podcast of posts from the Effective Altruism Forum, AI Alignment Forum, and LessWrong; prizes for effective altruists, mostly for writing particular blog posts; and fast grants for charities that lost their funding in the wake of the FTX situation. It also incubated High Impact Recruiting, a recruiter that works with effective altruist organizations.
Rumors have swirled about Nonlinear for quite some time. On September 6, 2023, Ben Pace wrote Sharing Information About Nonlinear, the result of a 100-200 hour investigation into Nonlinear. He found that two of Nonlinear’s employees, Alice and Chloe, were “more dependent on their bosses (combining financial, social, and legally)” than employees at comparable organizations; “both took on physical and legal risks that they strongly regretted, were hurt emotionally, came away financially worse off, gained ~no professional advancement from their time at Nonlinear, and took several months after the experience to recover”; and “had credible reason to be very scared of retaliation.”
Ben Pace wrote a brief followup post. Chloe wrote a comment describing her experiences in her own words. To my knowledge, Alice hasn’t made any public statements.
On December 12, 2023, Kat Woods wrote Nonlinear’s Evidence: Debunking False And Misleading Claims. It’s literally novel-length: more than 70,000 words. This response is what I’ll primarily draw on for my summary of the situation.
Dramatis Personae
- Alice: Alice (a pseudonym) traveled with Nonlinear for a few months, then became their employee for another few months. She was an “incubatee”—that is, she received support from Nonlinear to start her own organization. While employed at Nonlinear, she was asked to commit a felony. Nonlinear raises concerns about Alice’s credibility.
- Ben: Ben Pace is an employee of Lightcone Infrastructure, which runs (among other things) the group blog LessWrong. Ben wrote the original post Sharing Information About Nonlinear.
- Chloe: Chloe (a pseudonym) worked for Nonlinear, mostly at the same time as Alice. She worked simultaneously in operations for Nonlinear (e.g. paying their taxes) and as a personal assistant for Kat and Emerson (e.g. cleaning the house).
- Drew: Drew Spartz is Emerson’s brother and an employee of Nonlinear. He doesn’t have a leadership position.
- Emerson: Emerson Spartz cofounded Nonlinear and is their primary funder. He earned his money through founding iconic Harry Potter fansite Mugglenet and a number of other viral Internet content sites.
- Kat: Kat Woods cofounded Nonlinear. She’s the primary author of Nonlinear’s Evidence: Debunking False And Misleading Claims, so I’ll generally be referring to things said in that document as being said by “Kat.”
Ben’s Credibility
Ben made at least three factual errors in Sharing Evidence About Nonlinear. Ben originally messed up the timeline on when Alice started to work for Nonlinear, but corrected it. Still uncorrected is the claim that Chloe left in July; according to Kat, she left in June. Ben stated that Chloe was risking jail time for driving without a license in Puerto Rico. She wasn't, only a $50 fine.
Ben claimed that Nonlinear had had “ ~4 remote interns, 1 remote employee, and 2 in-person employees,” not counting Drew, Kat, or Emerson. Kat estimates that Nonlinear had had 21 people work for them. My own factchecking suggests that Nonlinear has had at least twelve to fourteen employees, inclusive of interns and not counting Drew, Kat, or Emerson. When I spoke to Ben, he said that he believes that Kat is padding her list with short-term contractors and volunteers. In particular, many of the people on my list were unpaid workers who might be best classified as volunteers. I think how many employees Nonlinear has had is dependent on how you define “employee.” In general, this is enough of a nightmare to figure out that I don’t want to call it for either of them.
Ben claimed that Alice was the only person to go through Nonlinear’s incubator program. I reached out to Kat and she said that five people went through their incubator program: Alice, Alexandra Bos, and three people at High-Impact Recruiting. I think Ben just got this one wrong.
Several claims strike me as misleading. Ben claimed that "they were not able to live apart from the family unit." However, Alice spent a bunch of time away from Nonlinear: she visited her family for a month and lived in the FTX condos for at least a few weeks. Ben seems to have meant that Alice and Chloe weren’t allowed to live in a different AirBnB paid for by Nonlinear, but the phrasing was ambiguous. Further, it seems misleading to me not to mention that Chloe did a disproportionate share of the chores because it was her job. Although cleaning wasn’t included as a duty in the job ad, Kat Woods’s job interview transcripts show it was verbally agreed upon before Chloe was hired.
Ben Pace has never done investigative journalism before and these mistakes strike me as typical for a first-time investigative journalist.
Kat Woods’s Credibility
Kat systematically misrepresents Ben and Chloe. She repeatedly uses quotation marks around paraphrased statements. In many cases, these paraphrases strip essential nuance or make Ben and Chloe sound like they’re thinking in a much more black-and-white way than they actually are. For example, Kat repeatedly claims to have been accused of abuse—often in quotation marks, as if it’s a direct quotation—when neither Ben nor Chloe publicly accused her of abuse.
Let’s look at a fairly representative case of Kat’s misrepresentation, chosen for ease of explanation. Kat writes:
If I (Chloe) asked them to fill out paperwork I would expect it was filled out at least reluctantly and plausibly deceptively or adversarially in some way.
While not in quotation marks, this is a direct quote from Sharing Information About Nonlinear. It is prefaced with “False, questionable, or misleading claim:”, which generally but not always prefaces direct quotes in Nonlinear’s Evidence. Kat Woods proceeds to explain that Chloe thinks this because she kept trying to get other people to fill out paperwork that was her responsibility.
However, the quote is different in its full context:
I did not find the time to write much here. For now I’ll simply pass on my impressions.
I generally got a sense from speaking with many parties that Emerson Spartz and Kat Woods respectively have very adversarial and very lax attitudes toward legalities and bureaucracies, with the former trying to do as little as possible that is asked of him. If I asked them to fill out paperwork I would expect it was filled out at least reluctantly and plausibly deceptively or adversarially in some way. In my current epistemic state, I would be actively concerned about any project in the EA or x-risk ecosystems that relied on Nonlinear doing any accounting or having a reliable legal structure that has had the basics checked.
Personally, if I were giving Nonlinear funds for any project whatsoever, including for regranting, I’d expect it’s quite plausible (>20%) that they didn’t spend the funds on what they told me, and instead will randomly spend it on some other project. If I had previously funded Nonlinear for any projects, I would be keen to ask Nonlinear for receipts to show whether they spent their funds in accordance with what they said they would.
The “I” is clearly not Chloe, it’s Ben. Ben specifically says that his view was based on talking to many people, not just to Chloe. Kat Woods’s reply is completely unresponsive to anything Ben said.
The entire post is like this. I strongly recommend reading Nonlinear’s Evidence with Sharing Information About Nonlinear in another tab so that you can cross-check to make sure Ben actually said what Kat said he said.
Furthermore, Nonlinear’s Evidence could easily have been three-quarters of the length if Kat had stuck to, as it were, Nonlinear’s evidence. Instead, the reader is treated to lengthy digressions about Kat’s metta meditation practice, Kat’s vacation photos, cryptic accusations that Ben has destroyed people’s lives, how Alice and Chloe need to do cognitive behavioral therapy, how Ben has wasted literally millions of dollars by making everyone post about Nonlinear drama instead of doing their jobs, and this charming little poem:
First they came for one EA leader, and I did not speak out --
because I just wanted to focus on making AI go well.
Then they came for another, and I did not speak out --
because surely these are just the aftershocks of FTX, it will blow over.
Then they came for another, and I still did not speak out --
because I was afraid for my reputation if they came after me.
Then they came for me - and I have no reputation to protect anymore.
The General Situation
The good news, given all these credibility problems, is that Kat and Ben actually agree on the general outlines of what happened. Kat’s primary objection to Sharing Information About Nonlinear is that Ben keeps saying things are bad, but actually they’re good.
Throughout this section, I’m going to be concentrating on things everyone agreed happened, and avoiding points which are in dispute. However, sometimes Alice or Chloe make a claim that Kat didn’t mention one way or the other, but which I think contributes to the overall picture. I’ve flagged these by saying that Kat “didn’t dispute” them, but those are more questionable than other statements.
Many claims and counterclaims were made about whether Alice and Chloe justifiably feared retaliation from Kat. I’m not going to go into this, mostly because I thought about it a lot and wound up with deep philosophical confusion about the concept of “retaliation.” I will flag that Kat outed Alice as having been in a mental hospital, which I consider to be completely beyond the pale as a response to criticism.
Kat and Emerson are both experienced entrepreneurs. In particular, Emerson (by his own account) has twenty-four years’ experience building organizations. It’s important to keep in mind that these are not understandable mistakes made by inexperienced people in the process of forming their first organization. Both sides agree that Alice and Chloe had relatively little work experience and therefore very little sense of what appropriate professional norms were.
Enmeshment
Alice and Chloe had very enmeshed relationships with other Nonlinear employees:
- Alice and Chloe both worked for Nonlinear.
- The primary in-person social outlet for both Alice and Chloe was the people they traveled with: Kat, Emerson, and Drew.
- A high percentage of Alice’s and Chloe’s in-person recreational activities were done with some combination of Kat, Emerson, Drew, and people they invited to travel with them.
- Kat writes: “You’re surrounded by a mix of uplifting, ambitious entrepreneurs and a steady influx of top people in the AI safety space. In the morning, you go for a swim with one of your heroes in the field. In the evening, a campfire on a tropical beach. Jungle hiking. Adventure. Trying new foods. Surfing. Sing-a-longs. Roadtrips. Mountain biking. Yachting. Ziplining. Hot tub karaoke parties. All with top people in your field.”
- Alice and Chloe were close friends with Kat, Emerson, and Drew. They (for example) discussed their feelings, insecurities, and life problems together.
- Kat, Emerson, and Drew acted as mentors for Alice and Chloe in both their work and private lives.
- Drew and Alice dated, but to her credit Kat disapproved.
I believe that the employee/friend/housemate/mentee combination predictably created ambiguity about which role Alice and Chloe were inhabiting at any given time, which gave Kat and Emerson an inappropriate amount of control over Alice and Chloe’s lives. For example, as a mentor, Kat gave Alice advice about whom to spend time with (see “Isolation”). But, as a housemate, Kat could easily see how Alice spent all her time and thus that Alice wasn’t following her advice. Alice didn’t have access to normal ways to set boundaries with mentors or employers, such as saying “I have plans” and not specifying that your plans are to hang out with random bartenders.
Another example occurred when Alice told Kat she had money worries. Kat, acting as a mentor and friend, told her that it was anxiety and gave her advice for checking her distorted thoughts. But there’s a conflict of interest if the person providing an outside view on whether you have enough money is also the one paying you. Indeed, Kat was arguably underpaying Alice (see “Finances”).
Finally, in addition to doing operations for Nonlinear, Chloe worked as Kat's and Emerson's personal assistant, responsible for cleaning, grocery shopping, arranging logistics for trips, and similar. Chloe spent about five to ten hours a week doing housekeeping. Kat and Emerson generally didn't contribute. Chloe’s time was considered less valuable than Kat’s, Emerson’s, or Drew’s. (Kat doesn’t clarify if Alice’s time was also considered less valuable.) Kat wrote that “[Emerson]’s not a friend. He’s her boss. She’s his assistant.”
Chloe handled all the logistics for an excursion everyone would enjoy while everyone else chatted; Kat thinks this is all right because Chloe was an assistant and it was her job. Emerson refused to drive Chloe to a medical errand when Chloe couldn’t drive; Kat thinks this is all right because Chloe was an assistant so she shouldn’t cost Emerson time. Ben Pace claims that "if a cofounder spilled food in the kitchen, the employees would clean it up"; Kat didn't dispute the claim, although she didn't explicitly agree that it happened. I believe that the former two are humiliating ways to treat a friend, and the last is a humiliating way to treat anyone.
The ambiguity was generally not resolved in Chloe’s favor. As a personal assistant, Chloe worked while other people had fun and couldn’t ask Emerson to do a favor for her. But, as a friend, Chloe was supposed to see time with Kat, Emerson and their friends as a fun job perk and a satisfying social life, not as billable hours.
To make up for the lower value she placed on Chloe’s (and possibly Alice’s) time, Kat regularly praised them. Alice found the praise insincere and manipulative. Kat says that she’s just a very positive person and seems to believe that praise and positivity can’t be insincere or manipulative.
Kat was aware of the boundary-related issues. She writes:
Alice/Chloe constantly complained about “unclear boundaries” as if we kept them unclear as part of a nefarious plot. If they wanted clear boundaries, they should have applied to Bureacracy Inc, not to be the first team members of a tiny nomadic startup with a tiny budget.
Finances
Nonlinear said that they always paid their employees on time. However, their own screenshots clearly show that Alice's stipend for April wound up in her bank account in mid-June, and her stipend for May wound up in her bank account in July. The first was sent on time but seems to have been held up in verification; I don’t know to what extent that was Nonlinear’s fault. Kat justifies the second by saying that Alice “wanted space.”
Alice and Chloe agreed to and weren’t surprised by Nonlinear’s compensation system. Alice and Chloe were paid a $1,000/month stipend, plus room, board, and expenses. Kat verbally said that room, board, and expenses were expected to amount to about $70,000-$75,000 in value. Nonlinear paid for "expensive music festivals, day trips to St. Barths, beach clubs, ziplining, tours, vacation excursions, and basically any activities she wanted to have paid for." The villas Alice and Chloe stayed at were extremely nice, as we can see from Kat’s photos.
Ben writes:
The employees were very unclear on the boundaries of what would and wouldn’t be paid for by Nonlinear. For instance, Alice and Chloe report that they once spent several days driving around Puerto Rico looking for cheaper medical care for one of them before presenting it to senior staff, as they didn’t know whether medical care would be covered, so they wanted to make sure that it was as cheap as possible to increase the chance of senior staff saying yes.
Kat doesn’t say that the boundaries of what would be paid for were clear, but emphasizes the glamorous nature of Nonlinear’s lifestyle.
Nonlinear didn't agree to cover its employees' medical expenses. Kat framed Emerson paying for Alice’s medical expenses as "Emerson helping a person he cares about out of his own pocket."
In general, $10,000 in expenses is much less valuable than $10,000 cash. Money is freedom. If you have $10,000 cash, you can save it, or donate it, or give it to a friend or family member, or spend it on hobbies you’re too embarrassed to tell anyone about. If someone is willing to buy you tickets to a music festival, you generally don’t get to go “I’d rather not, can you pay for my next four movie tickets instead?”—much less ask them to give you the money so you can save it for future medical expenses. In general, it seems that Alice and Chloe had relatively little control over how money was spent. For example, Alice and Chloe claim that Nonlinear refused to pay for separate AirBnBs; Kat doesn’t dispute this.
An appendix to the appendix provides an accounting of the money Nonlinear spent on Alice and Chloe. It seems to be the accounting Kat and Emerson did when Alice and Chloe were complaining about not getting the pay they were promised. My friend Manya made a spreadsheet.
The amount of money an organization spends on its employees is always higher than the employee’s compensation package, so describing the amount of money Nonlinear spent on Alice and Chloe as their compensation is misleading. As such, I refer to it as “the amount of money Nonlinear spent.” However, Kat views all the money Nonlinear spent on Alice and Chloe as part of Alice and Chloe’s compensation.
Some interesting takeaways:
- It looks like Alice and Chloe were on track to have Nonlinear spend about $70,000-$75,000 on them each year.
- None of the columns add up properly. In particular, the sum of Alice’s individual expenses, the sum of the subtotals, and the total money spent on Alice are three different numbers.
- No, I don’t know why they didn’t just add together the subtotals to get the total.
- A quarter of the money Nonlinear spent on Chloe was in the form of necessities to do her job (flights, car, laptop, phone).
- In general, if you’re flying for business purposes, your company will pay for your flights and not count that as part of your compensation package.
- It is also reasonably common for fly-in-fly-out jobs to provide vehicles (again, without this being considered part of the compensation package).
- Most companies that require computer work provide their employees work laptops without those laptops being part of the compensation package.
- However, unlike most companies, Nonlinear didn’t require Chloe to return her laptop.
- A third of the money Nonlinear spent on Chloe was rent.
- Nonlinear includes, as part of the money they spent on Alice, the rent for rooms that Alice wasn’t staying in (because she was visiting family or staying at the FTX condos).
- While timelines are too unclear for me to do the exact math, Nonlinear likely spent several thousand dollars less on Alice if you exclude the money they spent on rooms she wasn’t staying in.
Kat justifies Chloe’s and Alice’s low cash salary by pointing out that they got to travel a lot. However, the usual practice is that people are paid more, not less, for jobs where they travel a lot. This is because of the numerous downsides of travel. Kat points out several herself in the course of Nonlinear’s Evidence:
- "Not being able to drive on Caribbean islands is terrible - you’re trapped, no independence."
- "It's really hard to eat vegan in rural Puerto Rico."
- “She’s in the middle of nowhere and who else is she going to date [other than her coworker who is also her boss's brother whom she lives with].”
While I’m not any sort of lawyer, it seems to me that this pay structure likely violates tax and minimum-wage laws, as well as immigration laws for people working on an H-1B.
Isolation
Kat believes in the importance of digital nomads remaining socially connected to others. However, Kat and Emerson had a consistent pattern of encouraging Alice and Chloe to spend time with people they considered high value (i.e. effective altruists, especially those working in AI safety) instead of people they considered low-value. To be clear, Kat and Emerson didn't think Alice and Chloe should completely isolate themselves from people who weren’t effective altruists. Kat encouraged Alice and Chloe to call their families regularly. She explicitly supports spending some time with locals. Friends and family who didn't work in AI safety were invited to travel with Nonlinear, although they were lower priority to invite than AI safety people.
However, the vast majority of Kat’s evidence that she didn't isolate Alice and Chloe is evidence that she didn't isolate Alice and Chloe from effective altruists, particularly "top" effective altruists working in AI safety. Alice and Chloe were given lots of access to so-called top effective altruists: there was an average of 7 people living in the house. Nonlinear encouraged networking with FTX people. They traveled with Chloe's boyfriend, whom Kat Woods considered to "have high potential." Inviting people to travel with Nonlinear was framed as "one of the highest ROI things you can do”—that is, as an important means of bettering the world.
Kat and Emerson discouraged Alice from visiting her family because her trip overlapped with "some of the top figures in the field" coming to visit. (The chatlogs are suggestive that Alice timed her visit around a family emergency, but Kat doesn't explicitly mention this.) Kat also discouraged Alice from spending too much time socializing with locals, saying that she would have higher impact if she spent time with higher-value people.
It doesn't seem to have occurred to Nonlinear that Alice and Chloe might have been complaining about being isolated from people who aren't effective altruists.
The Drug Situation
Kat says “this post is long, so if you read just one illustrative story, read this one.” I agree the story is illustrative.
Alice was visiting Mexico for personal reasons. Kat asked Alice to buy her Ritalin and Adderall—Schedule II drugs—which you can buy without a prescription in some parts of Mexico, and to carry these drugs across the border. It turns out that, in this part of Mexico, you can’t buy Adderall and Ritalin without a prescription. Alice apologized for being unable to strategize to get a prescription (presumably an illicit prescription), because she was sick. Kat said that she understood and not to worry about it.
It’s in fact not legal to carry scheduled drugs across the border without a prescription, even from countries where it’s legal to buy them without a prescription. Carrying nonprescription amphetamines across the border is a felony. Kat confessed to asking her employee to commit felony drug smuggling.
I’m sure that Kat didn’t intend to ask Alice to commit a felony: otherwise, she would never have thought of the screenshots as exculpatory. Many techies in big coastal cities are used to living in a world where amphetamines are basically legal. It doesn’t make intuitive sense to them that the U.S. government thinks of Adderall differently than it thinks of SSRIs or antibiotics. But it does. According to the sentencing guidelines, even first-time offenders carrying a small amount of nonprescription amphetamines are sentenced to about a year in prison.
Then Kat says "But then Alice just went and got a prescription anyway." This is not a very clearly written section, but I think we’re supposed to understand that to mean that Alice lied in order to get a prescription for amphetamines to give to someone else? Which is also a quite serious crime?
Kat has a justification for committing conspiracy to distribute controlled substances: Alice wasn’t “a sweet, innocent girl” and she regularly smuggled drugs across borders. It seems to me that this is irrelevant to the main point. Even if your employee gives lots of people oral sex, it’s still sexual harassment to ask that your employee give you oral sex. And even if your employee smuggles drugs, it’s still against usual professional standards and the law to ask that your employee smuggle drugs for you.
Specific Situations
The Vegan Burger Situation
Alice was sick with covid and in complete isolation. She requested a vegan burger from Burger King, but Emerson and Drew refused because they wanted to work in a place with a nicer atmosphere. In spite of being sick herself, Kat went to the local store to get mashed potatoes; this was the only vegan food available at the store that Alice would conceivably want to eat. The chatlogs claim there was other vegan food in the house, such as oatmeal, almonds, tomatoes, or an orange; it's not clear to me why Alice didn't eat that food, and Kat doesn't really address the subject. According to chatlogs, Emerson and Drew ordered Alice a vegan burger the next evening. Alice disputes actually receiving the vegan burger.
Setting aside whether Alice actually received the veggie burger and why she didn't eat the vegan food that was apparently in the house, it seems that the mashed potatoes and a light breakfast were the only food that Alice had between dinner on December 14 and dinner on December 16.
Why am I talking about this? Well, for one thing, it’s a genre convention that all callout posts include something absolutely trivial for flavor.
But also, because Alice apparently characterized this incident as “I had nothing to eat for two days,” Kat refers to Alice as having "a delusional disorder." It’s true that “I had nothing to eat for two days” isn’t exactly the-truth-the-whole-truth-and-nothing-but-the-truth honesty. And Alice probably wasn’t being understanding enough about how hard it is to get vegan food in rural Puerto Rico. But it is totally normal for pissed off people to mildly exaggerate and to not have much empathy for what other people are going through.
Ben’s post correctly said that Alice “barely ate for two days.”
The Driving Situation
Nonlinear inexplicably hired Chloe, a person who didn’t know how to drive, for a position that involved regular driving. Kat, Emerson, and Drew taught Chloe how to drive. Chloe discovered that she couldn’t legally get a license in Puerto Rico, so she drove without a license.
Kat says that no one really cared if Chloe drove or not, that Chloe wouldn't have been fired for refusing to drive, and that Kat was happy to cover the costs of Ubers or public transit. I’m willing to take her at her word that she didn’t think of herself as pressuring Chloe.
However, it seems likely that it was an absolute nightmare to do a job that required regular driving without being able to drive. Public transit in rural areas is usually awful. Kat agrees that Ubers have unreliable coverage in Puerto Rico. There is a dispute about how hard it is to get taxis in Puerto Rico; I'm giving the point to Chloe, the person who actually had to book the taxis.
Kat says early on in Nonlinear’s Evidence that Chloe could have asked people to drive her and no one would have minded. However, when talking about a specific incident in which Chloe asked people to drive her, Kat says, "She is describing Emerson as a friend who’s refusing to help out, rather than a boss. You can see there’s nowhere in her model that her job is to save Emerson’s time and this would do the opposite."
Chloe claims that she had to come up with cool things for people to do in order to get them to drive her places—even for medically necessary trips. About one such trip, Kat says, "it was one of the coolest adventures we went on... If you’re into urban adventures / mysterious buildings, this would have been your dream day." I don't know why she thinks this is a counterargument to Chloe’s claim.
In conclusion: while Kat may not have intended to pressure Chloe to drive without a license, in practice Chloe probably didn’t have a realistic alternative to driving. This is Nonlinear’s fault for hiring a person who didn’t know how to drive for a job with mandatory driving.
Speculation About Drugs
There are contradictory stories about the felony drug smuggling incident. Kat claims that she only asked Alice to bring medication from a pharmacy. Ben says, however, that Alice was asked to bring multiple drugs, “some recreational, some for productivity.”
Furthermore, Kat may have asked her employees to do other drug crimes. She also doesn’t deny Chloe’s claim that she was asked to buy Kat weed:
One of my tasks was to buy weed for Kat, in countries where weed is illegal. When I kept not doing it and saying that it was because I didn’t know how to buy weed, Kat wanted to sit me down and teach me how to do it. I refused and asked if I could just not do it. She kept insisting that I’m saying that because I’m being silly and worry too much and that buying weed is really easy, everybody does it. I wasn’t comfortable with it and insisted on not doing this task. She said we should talk about it when I’m feeling less emotional about it. We never got to that discussion because in the next meeting I had with her I quit my job.
I think that the case that these incidents happened is weaker than the case that the request to smuggle amphetamines happened, because Kat didn’t admit to them. But I think it’s fairly likely that Kat also asked her employees to get her recreational drugs.
Alice’s Credibility
According to Kat, Alice has:
- Committed a crime against Nonlinear employees.
- Done three instances of “[___] involving police”.
- Claimed to have worked for an organization for almost a year when that organization had never heard of her.
- Accused 28 people (outside of Nonlinear) of "oppressing/persecuting/abusing" her.
- Accused six employers of various crimes, including murder.
Also that Alice lied on her resume, which is an extremely normal thing to put on a list next to “falsely accused her employer of committing murder.”
Kat is so good at not retaliating against people that she hasn’t posted Alice’s name, evidence for what Alice has done, or even what [___] is. Kat thinks she should get a lot of credit for how not-retaliate-y she is.
That’s absurd.
If someone is running around the effective altruism movement committing serious crimes, falsely accusing more than thirty people of abuse, and falsely accusing their employer of committing murder, I think you should let people know. I mean, talk about a missing fucking stair.
Kat seems to believe in some bizarre trade where Alice doesn’t talk about Kat paying her employees in scuba trips, and Kat doesn’t talk about Alice falsely accusing her employers of committing murder. But the effective altruism movement is not the Mafia. We don’t have a code of omerta.
Of course, one suspects that the reason Kat isn’t coming forward about Alice is not that Kat took her ethics lessons from Vito Corleone. One suspects that Kat isn’t coming forward because she’s, at best, misrepresenting what actually happened and, at worst, just making shit up.
Kat repeatedly characterizes Alice as "mentally unwell" and "delusional.” Kat might think that she’s being compassionate, because she said that it must be painful to be mentally unwell and that she wants Alice to get therapy. But mentally unwell people can still experience harm. If anything, mentally unwell people are more likely to be manipulated and exploited than mentally healthy people. It’s cruel to assume that someone can’t accurately report their experiences just because they have a severe mental illness—and it’s irrational to reach for an ad hominem attack to bolster your weak case for what you believe.
Kat goes beyond using Alice’s mental health history as an attack to offhandedly out Alice as having been in a mental hospital. As someone who has written publicly about my experiences in a mental hospital, I… honestly do not have words for how much this crosses the line.
Kat says:
Acknowledging the elephant in the room: some people advised us “it’s important for you to speak out about the elephant in the room, like, the common hypothesis that Ben white knighted too hard for Alice” but we are just not going to go there. If you know Alice’s history here you can guess.
[Section redacted. Per the advice of others, we’ll leave the pointer, but not going to go into it.]
This passage seems to me like it’s darkly hinting at some kind of sexual impropriety (?) while pretending to be above dark hinting. Either this is a subject of public interest, in which case you should say what you mean; or it’s private information about Alice’s dating (?) life of little relevance to the situation at hand, in which case you shouldn’t salaciously allude to things that only Ben’s and Alice’s personal friends ought to care about.
The Saga of Alice’s Incubated Organization
Like I said above, Alice was an “incubatee.” In a charity incubation program, the incubating organization provides support to a new charity. This support might include advice and mentoring, seed grants, help with HR or accounting, or a fiscal sponsorship, in which the smaller organization can use the larger organization’s tax-exempt legal status while it works through the process of applying for tax exemptions itself.
Sharing Information About Nonlinear says the following about Alice’s incubated organization:
Alice joined as the sole person in their incubation program. She moved in with them after meeting Nonlinear at EAG and having a ~4 hour conversation there with Emerson, plus a second Zoom call with Kat. Initially while traveling with them she continued her previous job remotely, but was encouraged to quit and work on an incubated org, and after 2 months she quit her job and started working on projects with Nonlinear…
As she quit she gave Nonlinear (on their request) full ownership of the organization that she had otherwise finished incubating…
One of the central reasons Alice says that she stayed on this long was because she was expecting financial independence with the launch of her incubated project that had $100k allocated to it (fundraised from FTX). In her final month there Kat informed her that while she would work quite independently, they would keep the money in the Nonlinear bank account and she would ask for it, meaning she wouldn’t have the financial independence from them that she had been expecting, and learning this was what caused Alice to quit.
Kat disagrees strongly with this view of what was going on. In her view, Alice had only ever worked as a project manager for Nonlinear projects. According to Kat, Alice believing she had an incubated organization is a sign that “she still sees enemies out to get her. The enemies are nebulous and their motives don’t really make any sense upon close inspection, but definitely they’re trying to control her.”
However, Kat was giving at best mixed signals:
As Clara Collier says:
I think this is due to some confusion about how incubation/fiscal sponsorship works. As an example, I run an organization (Asterisk Magazine) which is fiscally sponsored by another organization (Effective Ventures Foundation U.S.). This means that I report to the CEO of EVF, my organization's allocated funding is housed in EVF's bank account, and many of my ops tasks, like accounting and finance, are handled by EVF's staff. However, I have full control over the project I run, and I'm currently in the process of spinning out into a separate organization – at which point the funds I've raised would follow me. In this case, EVF is acting as an incubator to help me get Asterisk off the ground while I build up operational capacity.
Everything Kat shows here is consistent with Alice running her own incubated organization. It's normal for the head of an incubated organization to report to the CEO of their fiscal sponsor, use their fiscal sponsor's operational resources, and be reimbursed by their fiscal sponsor for expenses. I think that the Whatsapp messages support this interpretation – Kat is telling Alice that she can use the money raised from an outside funder for her organization however she wants, but can't use those funds for other projects (just like I can't use funds allocated for another EVF project to support Asterisk).
I’m not saying that Alice was right that Productivity Fund was her organization. I am saying that if you have hired someone to found their own organization, tell them the plan is for them to found their own organization, give them control of the budget for an organization, act in ways consistent with the organization being incubated in the ordinary way, and tell them that their organization might actually be independent within six months (!?)—
It is not delusional or paranoid or unstable for that person to conclude that you’re helping them to found their own organization! It kind of seems like you’re doing that!
In Sharing Information About Nonlinear, Ben mentions Alice expecting a $100,000 grant. Kat only mentions Alice expecting a $240,000 grant from the money Nonlinear had fundraised for the Productivity Fund. I found this confusing so I asked Ben; he explained that FTX had given Nonlinear four $100,000 grants to allocate to their incubated charities. Alice had expected to receive one of those grants as an incubatee. Kat doesn’t address the $100,000 grant or whether it was reasonable for Alice to expect it.
The Saga of Emerson Spartz’s Previous Business Life
Nothing in this section is directly related to Nonlinear. Ben made some claims about Emerson’s character, which Emerson disputed. (In this section, I’ll be referring to the author of Nonlinear’s Evidence as “Emerson,” because he wrote the relevant sections.)
There is a tedious debate about how much you can conclude about whether Emerson is a bad boss from the negative reviews of his company Dose on Glassdoor. I don’t think that you can infer that much from the reviews themselves. Emerson denies that Dose “commissioned” positive reviews. Instead, he says that Dose asked its current employees to review them on Glassdoor (without rewarding them for doing so). Emerson says "many companies" do this.
I talked to several people about whether this is normal. All of them said that they hadn't been asked to do this and several of them called it "creepy,” "slimy," and "weird."
The other thing that came up was Adrian Deck. Briefly, a sixteen-year-old named Adrian Deck created the Twitter account OMGFacts. Emerson bought the account from him. Deck wound up making only $100, even though Emerson had made over $100,000 of profit (and presumably was intending to make more). Emerson didn’t dispute the numbers, so I presume they’re accurate. Deck sued for more money. Emerson knew that Deck’s dream was to become a YouTuber, so he threatened to get two top YouTubers to create videos slamming Deck. In response, Deck settled in what Emerson characterizes as “a deal he was satisfied with that gave him significant financial freedom.”
I haven’t dug deeply into this story—I read a few articles, but I’m mostly working off Emerson’s story. And it kind of makes him look like an asshole? Emerson characterized Deck as “greedy” and “lazy” and points out that he didn’t tweet much from OMGFacts after Emerson’s acquisition. (As I understand it, Deck disputes this characterization of events.) But the OMGFacts brand itself was valuable: that’s why Emerson bought it instead of starting a competing YayFacts Twitter account. Emerson probably paid more to the lawyer who negotiated the contract than he did to Deck—the person who created the brand in the first place.
A California law invalidates any contract with a minor that isn’t overseen by a judge—a crucial protection for child actors, who are often exploited by their parents. Emerson characterizes this law as a “bizarre loophole,” which strikes me as a very cavalier attitude towards children’s labor protections.
If you’re doing business with a sixteen-year-old, you should make an effort to pay them the fair value of their work. It’s true Emerson is only four years older than Deck, but—by his own accounting—at this point he’d had eight years’ experience building organizations. He’s clearly the more experienced party. He bore a responsibility to be fair, not to cheat the person he’s signing a contract with and call them lazy and greedy for objecting.
I don’t know that he did anything illegal, but the whole thing seems slimy and skeezy and leaves a bad taste in my mouth.
This section isn’t very related to Nonlinear. But I think it’s evidence that Emerson is willing to make requests of his employees that go against usual professional norms and to sign unfair contracts with inexperienced people.
Could Nonlinear Employ People?
Ben claims "neither [Alice nor Chloe] were legally employed by the non-profit at any point." I spoke to him and he clarified that he believes Alice and Chloe were paid under the table, without being formally employed by Nonlinear. Kat says:
This is deeply misleading- it takes a normal, legal startup situation and frames it as a sinister illegal act - as if we were deliberately trying to keep them illegally employed. Nonlinear was not yet incorporated, and this was one of the first tasks Kat assigned to Chloe to handle, as our operations manager/assistant. Chloe was literally Nonlinear’s first hire and she’d been hired to set this up. It’s normal to begin operations before formally incorporating. Chloe did not finish the process in the 5 months she was with us.
I spoke to several friends with startup experience and they said that it isn’t legal or ethical to pay your employees under the table. Nonlinear absolutely shouldn't have been hiring employees without being a legal entity of some sort that could legally hire employees.
Furthermore, it’s not clear that Nonlinear couldn’t hire employees.
At the bottom of the page of the Operations Manager job ad it says:
The Nonlinear Fund operates as a project of Spartz Philanthropies, a 501(c)(3) registered non-profit in the US.
Spartz Philanthropies was already incorporated and could have employed Alice and Chloe.
In 2021—well before the job ad was posted in 2022—Spartz Philanthropies's 501(c)(3) status had been automatically revoked for failure to file Form 990 for three years. Drew said that Spartz Philanthropies losing its 501(c)(3) status was an error on the IRS’s part.
While I’m not a lawyer, my understanding of nonprofit law is that Spartz Philanthropies would still be a legal entity that could employ people, just not a tax-exempt one. In the same comment, Drew explains they were fiscally sponsored by Rethink Charity. In short, there is not one but two organizations that could have legally employed Alice and Chloe. I am very confused what the difficulty was.
When I asked him about this, Ben told me that the problem was related to difficulties sorting out visas: in particular, Chloe would be paid once she got an H-1B visa, which she was expected to do all the application for, both employer-side and employee-side. I believe this to be extremely abnormal. The H-1B process is immensely complicated, and it’s never certain that even the most qualified person will be able to get an H-1B. Even setting aside whether it’s legal to do so, you shouldn’t hire someone assuming that they’ll be able to get an H-1B soon and straighten out the legalities—it’s all too likely they won’t be able to. It’s very strange to hire a person to do both the employer and the employee side of the petition for a visa that she can’t be hired without. However, Ben spoke to me shortly before publishing and I didn’t have time to check with Nonlinear, so take this with a grain of salt. It does provide an alternate explanation for why Chloe at least was employed under the table.
Weird sidenote: “Nonlinear operates as a project of Spartz Philanthropies, a 501(c)(3) registered non-profit in the US” is in fact still on Nonlinear’s webpage (archive made two hours before publishing). The text was just turned white. Try highlighting the footer, you can see it for yourself. I have literally no idea why.
Takeaways
People investigating effective altruist organizations for wrongdoing should be aware that investigative journalism is extremely hard and should be more careful than the default not to make factual errors.
I don’t think Kat and Emerson are evil, abusive, or culty. I don’t think they should be completely expelled from the effective altruism movement.
However, I think that they offered inexperienced people deals that those people wouldn’t take if they were more experienced; maintained an enmeshed and boundaryless work environment that would predictably cause harm to their employees; have a lax attitude towards the necessary paperwork to run an organization; asked an employee to commit a felony (probably without knowing it was a felony); outed an employee as having been in a mental hospital, apparently in response to her complaining about Nonlinear’s work environment; and generally show astonishingly bad judgment.
I think the “are Kat and Emerson evil culty abusers?” discourse is a distraction from the main point. For an organization to be considered an effective altruist organization, it’s not enough that the things they do wouldn’t be illegal in a hypothetical libertarian utopia. An effective altruist organization needs to be actively good.
Much of Kat and Emerson’s effective altruist work focuses on mentoring young effective altruists and incubating organizations. I believe the balance of the evidence—much of it from their own defense of their actions—shows that they are hilariously ill-suited for that job. Grants to Nonlinear to do mentoring or incubation are likely to backfire. Grantmakers should also consider the risk of subsidizing Nonlinear’s mentoring and incubation programs when they give grants to Nonlinear for other programs (e.g. its podcast).
I would advise inexperienced people to avoid working for Nonlinear; the job may well have pay or working conditions you wouldn’t agree to if you were more experienced. I would also advise people, especially neurodivergent people, to be careful about disclosing sensitive information to Kat and Emerson.
Further, I believe that we, as the effective altruist movement, should adopt norms such as:
- People should be paid for their work in money and not in ziplining trips.
- A person should not simultaneously be your friend, your housemate, your mentee, an employee of the charity you founded, and the person you pay to clean your house.
- Nonprofits should follow generally accepted accounting principles and file legally required paperwork.
- Employers should not ask their employees to commit felonies unrelated to the job they were hired for.
Kat and Emerson don’t seem to agree with these norms. Until they do, I would advise against allowing them to promote themselves in effective altruist spaces such as Effective Altruism Global, the 80,000 Hours job board, etc.
Putting this here since this is the active thread on the NL situation. Here's where I currently am:
I broadly endorse your judgment on this topic inasmuch as I've observed it, for whatever that is worth. I do want to add the specific note that one of the most serious problems with the process issues is that they led to such an adversarial launching point to the topic that the response was in an unreasonably difficult position from the beginning. Ozy's article (while being on the whole a defensible view) criticizes the approach of the response quite harshly, but speaking with precision and grace in high-pressure, intensely adversarial situations with an audience primed to distrust you is a specific, rare skill with only limited bearing on someone's approach in regular times.
You are likely right that a mistrial is unrealistic at this point, but inasmuch as people take the approach you advocate, it's worth emphasizing the reasons process concerns make putting things together so complicated.
There may be some tricky interactions with US defamation law here. One the one hand, there's the likelihood that fact checking will reduce the risk of error, and there is no defamation without falsity.
On the other hand, getting a source to retract during the pre-publication review process could add heft to a threat (real or feigned) to sue for defamation. In the US, the constitutional minimum for defamation liability is negligence for ordinary figures; "actual malice" (knowledge or reckless disregard of falsity) for public figures. If the would-be author knows prior to publication that the sources have retracted, that could significantly increase the defamation-liability risk under either standard. Thus, the target may have an incentive to pressure the sources into retracting to bolster the target's defamation-suit threat against the would-be publisher.
Every situation is different, and would-be publishers should seek legal advice as appropriate. But my guess is that where the target is a public figure, following certain types of precommitment procedure might well cause a net increase in the risk borne by the publisher.
I think even this under-plays the fact-checking point. It’s very difficult for external observers to evaluate these sorts of situations, particularly when there’s fierce pushback from the accused party. The level of rigour in fact-checking even secondary details is one heuristic people can use to evaluate the journalist’s case. If the accused can point out a litany of errors in the post it makes it much harder to trust the bigger claims. It also makes it immensely harder to persuade those sympathetic to the accused.
Habryka’s suggestion in comments elsewhere that this level of fact-checking is basically too much to ask for seems to miss this.
In that spirit, I feel I should point out that for Alice and/or Chloe the compensation was eventually switched to cash on their request. I don’t think this at all undermines your points - it’s clear that Kat thinks the in-kind compensation structure was an amazing offer that Alice and Chloe should have felt lucky to receive, and that her agreeing to the switch to cash was going above and beyond as an employer. But it seems important to make sure people don’t get an inaccurate impression, or at least don’t think you have one.
I’m a bit confused about the “illegal more like speeding is illegal than like shoplifting is illegal” analogy - speeding seems clearly worse given how many people die in road accidents; large chains just price in a baseline level of theft.
Society defines operating an automobile over a certain rate of speed ("speeding") as illegal. We've decided there need to be some fairly bright-line rules, and so we've made speeding illegal in some circumstances where it is a fairly rational, victimless thing to do. Indeed, we don't want a ton of litigation about whether someone was driving unreasonably fast for the circumstances, so we've defined driving over X mph/kph as illegal (absent special circumstances like a true emergency).
Another offense like this: letting your dog out in public without a leash. It's often a reasonable enough thing to do, but (at least where I live) the effect of the law is that the police officer gets to decide whether you should be ticketed, and the police officer's exercise of discretion is essentially unreviewable. The law's authors perfectly well knew that lots of people would let their dogs off leash, and the cops would generally do nothing or issue a verbal admonition.
In contrast, most people see shoplifting as inherently wrong (lawyers would say: malum in se, rather than malum prohibitium). Although there can be defenses, we do not expect people to regularly break anti-shoplifting laws, or for the police to decide which shoplifters were acting unreasonably.
At least in the US there are some illegal activities that are very common and socially accepted. Learning that someone made a habit of going a small amount over the speed limit, jaywalking, or consuming marijuana at home probably wouldn't make most people think poorly of them. Other things, even ones with arguably low social cost, are seen as pretty unacceptable: shoplifting, insurance fraud, going through red lights when you can see no one is coming.
I don't want to get into whether speeding or shoplifting is actually worse (though you and Will are welcome to keep exploring that!) since what I'm gesturing at is only about how people generally think of the offense; I've edited the comment to change "speeding" to "jaywalking".
Insurance fraud has low social cost? Explain?
I'm not convinced the social cost is low, and I'm not convinced for shoplifting either; hence the 'arguably'. I think insurance fraud, though, is often quite a lot like shoplifting? You're getting something for free from a large company, they have budgeted based on a non-zero amount of it, the costs are spread across all their customers, risk of death to anyone is very low, etc.
"How many people die in road accidents" doesn't tell you much about the badness of speeding without the denominator - which in the US is approximately everybody approximately all the time.
I would still think the ‘micromurder’ of speeding is higher than that of shoplifting? I still think I’m missing something in understanding the analogy
I don't think this is empirically true. US speed limits are typically set lower than the safest driving speeds for the roads, so micromurders from speeding are often negative in areas without pedestrians.