Minh Nguyen

Karma: 644

I proposed the Nonlinear Emergency Fund and Superlinear as Nonlinear Intern.^[1]

I co-founded Singapore’s Fridays For Future (featured on Al Jazeera and BBC). After arrests + 1 year of campaigning, Singapore adopted all our demands (Net Zero 2050, $80 Carbon Tax and fossil fuel divestment).

I developed a student forum with >300k active users and a study site with >25k users. I founded an education reform campaign with the Singapore Ministry of Education.

^
I proposed both ideas at the same time as the Nonlinear team, so we worked on these together.

Minh Nguyen 25 Nov 2023 4:46 UTC
3 points
0 ∶ 0
in reply to: VictorW’s comment on: VictorW’s Quick takes
I find that asking EAs (or anyone, really) for open-ended feedback tends not to yield novel insight by default. EAs tend to have high openness, and as long something passes the bar of “this has plausible theory of change, has no obviously huge downside and is positive EV enough to be worth exploring” , it isn’t subject to particularly intense scrutiny. Consider also that you have thought about the problem for days/weeks whereas they’ve only thought about it for maybe 10-20 minutes
Haven’t found a 100% perfect solution, but usually I express my 2-3 most pressing doubts and ask them whether there’s a possible solution to those doubts. It scopes the question towards more precise, detailed and actionable feedback.
Alternatively, if the person has attempted a similar project, I would ask them what goals/principles they found most important, and 2-3 things they wish they knew before they’d started.

Minh Nguyen 23 Nov 2023 7:55 UTC
7 points
3 ∶ 2
in reply to: NickLaing’s comment on: Elizabeth’s Quick takes
I will say I also never use the Drowning Child argument. For several reasons:
- I generally don’t think negative emotions like shame and guilt are a good first impression/initial reason to join EA. People tend to distance themselves from sources of guilt. It’s fine to mention the drowning child argument maybe 10-20 minutes in, but I prefer to lead with positive associations.
- I prefer to minimise use of thought experiments/hypotheticals in intros, and prefer to use examples relatable to the other person. IMO, thought experiments make the ethical stakes seem too trivial and distant.
What I often do is to figure out what cause areas the other person might relate to based on what they already care about, describe EA as fundamentally “doing good, better” in the sense of getting people to engage more thoughtfully with values they already hold.

Minh Nguyen 22 Nov 2023 2:14 UTC
3 points
0 ∶ 0
in reply to: James Herbert’s comment on: James Herbert’s Shortform
On CB, my views that are half informed by EA CBs and half personal opinions:
1. Very casual events—If you are holding no events for a long time and don’t have much capacity, just hold low-stakes casual events and follow-up with high-engaged people afterwards. Highly-engaged people tend to show up/follow up several times after learning about EA anyway. 80-90% of the time, I think having some casual events every few weeks is better than no casual events.
2. Bigger events—Try to direct highly-engaged people to bigger and/or more specialised events. The EA community is big and diverse, and letting people know other events exist lets them self-select better. When I first explored beyond EA Singapore, I spent 2 months straight learning about every EA org and resource in existence, individually reviewing all the Swapcard profiles at every EAG. That was absolutely worth the effort, IMO.^[1]
3. 1-on-1s are probably still important − 1-on-1s with someone of very similar interest areas or career trajectories are the most valuable experiences in EA, in my opinion. Only 10% of 1-on-1s are like this, but they more than make up for the 90% that don’t really go anywhere. As much as I try to optimise, this seems to be a numbers game of just finding and meeting a lot of potentially interesting people.^[2]
4. Online resources—For highly-engaged EAs, important information should be online-first. I’m of the opinion that highly-engaged/agentic new EAs tend to read a lot online, and can gain >80% of the same field-specific knowledge reading on their own. This especially holds true in AI Safety, which is like … code and research that’s all publicly available short of frontier models. I think events should be for casual socials, intentional networking and accountability+complex coordination (basically, coworkers).
1. ^
  If you want the ⁸⁰⁄₂₀ for AI Safety, check out aisafety.training, aisafety.world, check EA Forum, Lesswrong and Alignment Forum once a week (~1 hour/week), check 80k job board and EA Opportunities Board once a week (~20 minutes/week), review forum tags for things like prizes, job opportunities and research programs to see what programs were run last year that will be run again this year.
  It is possible to capture all open opportunities this way. The rest is just researching interesting orgs, seeing which ones you vibe with and engaging with them. This is just for AI Safety, for other cause areas I’d expect the same amount of time spent passively checking.
2. ^
  My personal view is people should slightly prioritise “potentially interesting” over “potentially useful”. The few times I’ve met EAs just because they’re high-ranking, the conversation is usually generic and could have been had by Googling and emailing/texting.

Minh Nguyen 21 Nov 2023 13:02 UTC
6 points
2 ∶ 0
in reply to: James Herbert’s comment on: James Herbert’s Shortform
I have heard many people argue against organising relatively simple events
I’m actually very surprised to hear this. What does the “common view” presume then?

Personally, I see 3 tiers of events: 1. Any casual, low-commitment, low stakes events 2. Big EA conferences that I find quite valuable for meeting lots of people intentionally and socially 3. Professionally-focused events (research fellowships, incubators etc)

I think “simple” events like 1 are great for socialising and meeting new people. While 2 and 3 get more done, I don’t think the community would feel as welcoming if the only events occurring were ones where you had to be fully professional.

Sometimes I still want to interact with EAs, but without the expectation of “meeting right” or “networking”. I suspect this applies especially to introverts and beginners. Even just going to a conference with the expectation of booking lots of 1-on-1s vs just chilling feels very different.

Minh Nguyen 21 Nov 2023 9:16 UTC
10 points
3 ∶ 0
in reply to: John G. Halstead’s comment on: CEA is fundraising, and funding constrained
Perhaps the thinking was that FTX is not directly relevant to this specific post?

But I agree, clearly FTX is on the minds of many people reading this, so avoiding the elephant in the room is not working.

Minh Nguyen 18 Nov 2023 2:17 UTC
52 points
9 ∶ 8
on: Sam Altman fired from OpenAI
Found this on Reddit: Anxious_Bandicoot126 comments on Sam Altman is leaving OpenAI (reddit.com)
I feel compelled as someone close to the situation to share additional context about Sam and company.
Engineers raised concerns about rushing tech to market without adequate safety reviews in the race to capitalize on ChatGPT hype. But Sam charged ahead. That’s just who he is. Wouldn’t listen to us.
His focus increasingly seemed to be fame and fortune, not upholding our principles as a responsible nonprofit. He made unilateral business decisions aimed at profits that diverged from our mission.
When he proposed the GPT store and revenue sharing, it crossed a line. This signaled our core values were at risk, so the board made the tough decision to remove him as CEO.
Greg also faced some accountability and stepped down from his role. He enabled much of Sam’s troubling direction.
Now our former CTO, Mira Murati, is stepping in as CEO. There is hope we can return to our engineering-driven mission of developing AI safely to benefit the world, and not shareholders.
Obviously just speculation for now, but seems plausible. The moment the GPT store was released I thought:
“wow that’s really good for business … wow that’s really bad for alignment”

Minh Nguyen 12 Nov 2023 6:44 UTC
11 points
2 ∶ 0
on: “Effective Altruism, Longtermism, and the Problem of Arbitrary Power” by Gwilym David Blunt
Honestly, I like that this essay tries to engage with and understand EA more than other critiques I’ve seen. Usually when I see an article about EA’s association with billionaires/elites, it tends to be a lot less substantive.
I do agree that EA decision making generally biases towards privilege. EA orgs pretty openly bias towards recruiting and hiring from prestigious universities, which are significantly overrepresented in EA demographics. I also recall data from Spencer Greenberg that essentially placed EAs as ideologically centrist, as oppose to more left-leaning social movements popular with similar demographics.
I think her critique here is actually quite sound:
Effective altruism, when it began, had ascetic norms designed to constrain power, but slowly and surely they have been eroded by arguments about ‘effectiveness’, and as they receded the absence of strong institutional constraints became apparent. This leaves effective altruism and longtermism exposed to what we may call ‘the despotism trap’.
It rejects constraints against these visionaries as the pettiness of small minds. They drape themselves in the regalia of liberty to challenge government, but their interest is in being able to exercise uncontrolled power. The excuse is that the ends will justify the means. We can see how these attitudes have influenced effective altruism when it overlaps with ‘philanthrocapitalism’, whereby people like Gates bring the mechanisms of the market into philanthropy on the grounds that is more effective.
I think this is a pretty accurate description. EA and EA-adjacent orgs are generally funded by rich philanthropists, and receive less support and collaboration with governmental bodies. A lot of EA-adjacent or recommended orgs exist in the private sector. I think at the very least, we have to acknowledge that funding sources do affect culture. EA happens to be disproportionately funded by tech billionaires, so EA is more likely to be sympathetic to the views and ideologies of tech billionaires.
I think where outside critics get it wrong, is the interpretation that because EA is funded by social elites, it actively sides/colludes with social elites. I’m not sure how to convey it, but outside coverage seems to imply that EA is essentially a front for social elites to gain more power without actually doing any good, when my take is that EA began in elite college campuses, and so the most accessible sources of support and funding just happened to be social elites. Most notably, coverage of EA often focuses more on ties with social elites than the actual work done by EA orgs in various cause areas. The average article I read contains maybe 1-2 brief sentences about the fairly complex work/research done by EA orgs, and the rest of the article is just speculating about its ties w rich elites.

Minh Nguyen 8 Nov 2023 5:08 UTC
5 points
1 ∶ 0
on: AMA: Ben West, former startup founder and EtGer
When founding something new, how do you balance money and impact?
I’m (attempting to be) an EA entrepreneur, but I find it difficult to balance finding product-market-fit+scaling with actual impact. At different times, I’ve found myself focusing too much on building something “big” that doesn’t have any actual object-level impact, or being too perfectionist about optimising impact versus simply doubling down on something that works and is influential/makes lots of money.
I haven’t figured it out and don’t expect this to be an easy answer, so just curious what your thoughts are on the resource/impact tradeoffs in decision making.

Minh Nguyen 4 Nov 2023 4:25 UTC
3 points
0 ∶ 0
on: Meta Coordination Forum 2023: Project brainstorm
Just gonna weigh in on some of these from my time researching this stuff at Nonlinear.
A common knowledge spreadsheet of directly responsible individuals for important projects.
Strongly agree. It’s logistically easy to do, one person could cover 80% of EA projects within a week. I’ve been using AI Existential Safety Map (aisafety.world) a lot in my list of followups for 1-on-1s.
In the long run, a well-maintained wiki similar to/synced with the EA Opportunities Board (which I also heavily recommend) could make this really comprehensive.
More “public good”-type resources on the state of different talent pipelines and important metrics (e.g., interest in EA).
I read every EA survey I see. They’re often quite interesting and useful. I wouldn’t say they’re neglected since EAs do seem to love surveys, but usually a net positive.
More coherent and transparent communication about the funding situation/bar and priorities.
I am of the opinion that every EA funder should be as transparent and detailed about their funding bar/criteria as possible. Unlike for-profits/VCs, I don’t see a strong reason for secrecy other than infohazards. It helps applicants understand what funders look for which helps both funders and applicants. I believe that applicant misconceptions about “what funders want” can hinder EA a lot in the long run due to mismatched incentives. I see a lot of compelling project directions censored/discarded in the early stages simply because applicants think they should be more generic (because being more generic works well in conventional success pathways).
More risk management capacity for EA broadly as a field and not just individual orgs.
I really liked this post Cash and FX management for EA organizations — EA Forum (effectivealtruism.org) by @JueYan.
Advanced 80K: Career advice targeted at highly committed and talented individuals.
Agree, but I never figured out how to scalably execute this. Usually, if someone has a skillset+motive to do really well in EA, my priority is to 1. inform them of helpful resources to fill in themselves 2. try to link them with someone doing what they’re trying to do.
The problem is that it seems hard to predict in advance who they’d consider a valuable connection. I think none of my most valuable connections in EA so far would’ve been referred to me by someone else.
Tractable idea: A list of helpful links sent to EAGx and EAG attendees post-conference.
A survey to identify why high-value potential members “bounce off” EA.
I actually have bounced off EA for 3 years before (2019-2022). For me, the big reason was that I couldn’t find any follow-up steps to pursue (especially coming Singapore). My experience within EA is very inspiring and exciting interactions followed by not much follow-up (guidance, next steps, pursuing opportunities, encouraging people to start projects etc.).
[just gonna agree with all the AI Safety points, they’ve all come up before in my discussions]
Evaluation and Accountability
Shoutout to @Mo Putera who is working on this.
Media and Outreach
Casual observation that I can’t recall a single EA social media account that I browse simply because it’s fascinating and not because I wanna support EA on social media.
And I’m into weird stuff, too. I just binged hour-long videos on Soviet semiconductors and the history of Chang’an.
Incubators: One respondent stated that incubators are “super hard and over-done,” mentioning that they are too meta and often started by people without entrepreneurial experience.
Agree, this point has been discussed in detail before. What we learned from a year incubating longtermist entrepreneurship — EA Forum (effectivealtruism.org)
I think it’s just hard to do well because there’s so many points of failure, it takes a long time for any results to show and it requires both social skills and technical expertise. That said, I do think a longtermist version of Charity Entrepreneurship seems promising to pilot (actually, I’m gonna bring this up to Kat Woods right now).
Fastgrants and other quick funding mechanisms.
I really like Manifund as a platform!

Minh Nguyen 3 Nov 2023 7:11 UTC
3 points
0 ∶ 0
on: Still no strong evidence that LLMs increase bioterrorism risk
Disclaimer that I am practically a layman on this topic.
My threat model is that creating bioweapons requires a series of steps that are getting easier and easier to do, and LLMs are significantly accelerating one of these steps.
In that sense, open-sourcing LLMs does contribute to increased biorisk, but the action of restricting open-source LLMs to restrict the increase in biorisk seems a disproportionate response by itself?
For example, the internet certainly increased the ease of conducting terrorism, but many people would consider it a disproportionate response to heavily restrict the internet just to restrict terrorism.

Minh Nguyen 27 Oct 2023 13:11 UTC
5 points
0 ∶ 0
on: OpenAI’s new Preparedness team is hiring
There’s a typo!
Outline an experiment plan to (ethically and legally) measure the true feasibility and potential severity of the misuse scenario you described above assuing you have a broad range of resources at your disposal, including an ability to perform human-AI evaluations. *
Hey, at least we know it was written by a human!

Minh Nguyen 24 Oct 2023 7:08 UTC
7 points
1 ∶ 0
in reply to: Linch’s comment on: Linch’s Shortform
Assume by default that if something is missing in EA, nobody else is going to step up.
I’ve been saying this a lot.
I think it’s common for newcomers to assume EA is more structured than it is. For example, they might stumble upon an addressable gap, not find any solutions but simply assume they don’t know enough about EA and that “EA higher-ups know better”.
I think at the very least, new EAs who have strong ideas about what’s missing should make a good effort to ask people who “should know”. If there’s a proposal that makes sense, and there’s no obvious reason people shouldn’t do it, it helps to ask your circles to recommend experts, read the experts’ takes and ask those experts directly if you still have question.
At the very least, you learn about a problem you care about. In some cases, the experts are working on the problem, and open to collaborators. At best, you find an interesting problem to tackle yourself.
In my limited experience, grantmakers love proposals that make them go “wow, someone should have really done that earlier”. These are usually “safe” bets with a high chance of positive impact, and can be done by EA newcomers with relevant knowledge of industry best practices.

Minh Nguyen 17 Oct 2023 7:06 UTC
13 points
6 ∶ 0
in reply to: Calum’s comment on: Politico article on Open Phil, Horizon Fellowship, and EA
the article nevertheless portrays the EA community as if it is pushing frivolous, ill-considered ideas instead of supporting the Real, Serious concerns held by Thoughtful and Reasonable people.
What bothers me is that many criticisms of EA that hinge on “EA is neglecting this angle in a careless and malicious manner” could have been addressed with basic Googling.
I don’t expect the average Joe to actively research EA, but someone who’s creating a longform written or video essay with multiple sources should be held to higher standards.
One example: Effective Altruism and the Cult of Rationality: Shaping the Political Future from FTX to AI — COLUMBIA POLITICAL REVIEW (cpreview.org)
Thus, we must be wary of the power behind a mindset focused solely on the hypothetical future and allow space and empathy for the short term needs of society. A tyranny of the quantifiably rational majority would lead to more quantifying of human suffering than policy change.
This conclusion somehow manages to completely ignore the neartermist cause areas, the frequent discussions about prioritising neartermism vs longtermism and the research on neartermism that does focus on qualitative wellbeing. I genuinely don’t know how someone can read dozens of pages about EA and not come across any reference to neartermism.
Ultimately, I just read so many EA criticism pieces where it feels like the writer hasn’t talked to EAs, or conveniently ignores that most EAs spend most of their work time thinking about how to solve real problems that affect people.
These articles paint a picture of EAs so completely divorced from my actual interactions with EAs. The actual object-level work done by EA orgs is often described in 1-2 sentences, while an entire article is devoted to organisational drama/conflict. Like … how do you talk about EA without mentioning the work EAs do to … solve problems???

Minh Nguyen 13 Oct 2023 19:08 UTC
7 points
2 ∶ 0
in reply to: Denis ’s comment on: How EA is portrayed in “Going Infinite”
People think empathy is about a warm, fuzzy feeling they get when they help someone. But it’s not. Empathy is about getting inside someone’s head, seeing the world from their perspective and understanding what they need.
Not a psychologist, but I’ve heard this distinction described as “cognitive empathy” vs “affective empathy”.
Cognitive and Affective Empathy in Eating Disorders: A Systematic Review and Meta-Analysis—PMC (nih.gov)
Empathy can be separated into two major facets. Cognitive empathy refers to the ability to recognize and understand another’s mental state (part of theory of mind (ToM) or mentalising) while affective empathy is the ability to share the feelings of others, without any direct emotional stimulation to oneself
I’m not aware of surveys being done, but I strongly suspect EAs have average to above-average affective empathy, and unusually high levels of cognitive empathy.
Semi-related: There’s this great video from Dr K about how smart kids develop a reliance on cognitive empathy, and that reminds me of a lot of EAs.

Minh Nguyen 11 Oct 2023 21:09 UTC
15 points
0 ∶ 0
on: Minh Nguyen’s Quick takes
Hello! I’m looking for AI Safety proposals/project ideas to deploy within CivitAI, the biggest open-source AI image model hosting platform.
I own AI Hub, an AI voice model hosting platform w/ ~1 million users. We are discussing integration/merging with CivitAI.com, which is an open source AI model hosting platform and the 7th most visited AI website with 27 million monthly users. Most of their hosting is Stable Diffusion image models, but I’m helping them introduce AI voice cloning model hosting. I’d get a lot of autonomy over distribution for voice and possibly text models.
The founders mentioned they’re also exploring content moderation solutions, and when I asked, they said they were open to working with AI Safety companies. CivitAI is very influential in the open-source community, so I think shaping policy here could significantly shape AI Safety wrt open-source and local models. Some directions I’m exploring:
- Content and open source model moderation frameworks/tools
- Collecting data and running tests that aid alignment research
If anyone has ideas, or knows anyone who might be interested, do LMK! This is exploratory, but I aim to formulate an idea and deploy to users within ~3 months, with 80% confidence. I am literally open to trying any suggestions—for-profit, nonprofit, research, product, governance, alignment research etc.
Even if you have an idea that’s not AI image-related, I have a decent (40%) chance of proposing it to Github, Huggingface etc. Plus, my task is to diversify outside of image models anyway, so LLM-related proposals could still be relevant.
(this is still in exploratory/idea phase, and I wasn’t sure if this should be a full post)

Minh Nguyen’s Quick takes

Minh Nguyen11 Oct 2023 21:09 UTC

5 points

1 comment1 min readEA link

Minh Nguyen 7 Oct 2023 23:04 UTC
3 points
1 ∶ 0
on: What we learned from a year incubating longtermist entrepreneurship
I just discovered this thread, but figured I’d back up some data points! I’ve been in the EA community since 2016-ish and done entrepreneurship and longtermist work for about 2 years. I won’t lean too much into my experience, I’ll generally say things other entrepreneurs and impact-minded people would agree with.
Re: Founder pool
The talent pool size is much smaller (only 100s of potential longtermist entrepreneurs in total vs tens of thousands of entrepreneurs).
It is helpful to think of “longtermism” and “entrepreneurship” as two entirely distinct, and very uncommon skillsets/mental frameworks. Currently, way less than 1% of the world would identify as longtermist, and less than 1% would identify as entrepreneurs. While these traits are slightly correlated, the vast majority of lontermists are not entrepreneurs, and the vast majority of entrepreneurs are not longtermist.
That’s totally fine, but makes things difficult for a longtermist incubator. I mean, if Y Combinator started prioritising longtermist involvement, I’m pretty sure they’d have similar challenges.
Re: Project pool
We found almost no longtermist ideas for traditional startup-minded people to pursue (product-based with quick path to scale, feedback loops), and don’t expect this will be the majority of promising longtermist startups.
Again, overlap is hard. One thing Y Combinator constantly emphasises is that most startups never achieve product-market fit. I.e. it’s really, really hard to build something people want to pay for. So the reliable solution is to try and iterate a lot with fairly little information. A 10% success rate is really good if you try a dozen times a month.
So just building anything successful is hard. Then you have to try and make something impactful, which is another, completely separate hard thing to do. Most successful, lucrative companies are not “effective” or “impactful” in the longtermist sense. I’ve tried this thought exercise myself, sometimes I would think of/test like 9 potential lucrative startup ideas before I find one that’s lucrative and also possibly impactful.
Anyway, point is: This is a difficult thing to do, essentially trying to find a very narrow eligible pool and get them to work on a very small set of possible ideas. I really appreciate that this was tried!

Minh Nguyen 21 Sep 2023 18:06 UTC
6 points
1 ∶ 2
in reply to: richard_ngo’s comment on: richard_ngo’s Shortform
Re: Hacker culture
AI safety becomes the single community that’s the most knowledgeable about cutting-edge ML systems. The smartest up-and-coming ML researchers find themselves constantly coming to AI safety spaces, because that’s the place to go if you want to nerd out about the models. It feels like the early days of hacker culture.
I’d like to constructively push back on this: The research and open-source communities outside AI Safety that I’m embedded in are arguably just as, if not more hands-on, since their attitude towards deployment is usually more … unrestricted. For context, I mess around with generative agents and learning agents.
I broadly agree that the AI Safety community is very smart people working on very challenging and impactful problems. I’m just skeptical that what you’ve described is particularly unique to AI Safety, and think that descriptiom would apply to many ML-related spaces. Then again, I could be extremely inexperienced and unaware of the knowledge gap between top AI Safety researchers and everyone else.
Re: Environmentalism
much more similar to the environmentalist movement. It has broader reach, but alienates a lot of the most competent people in the relevant fields. ML researchers who find themselves in AI safety spaces are told they’re “worse than Hitler”
I was a climate activist organising FridaysForFuture (FFF) protests, and I don’t recall this was ever the prevailing perception/attitude. Mainstream activist movements and scientists put up a united front, and they still mutually support each other today. Even if it was superficial, FFF always emphasised “listen to the science”.
From a survey of FFF activists:
Our data show that activists overwhelmingly derive their goals from scientific knowledge and reject the idea that science could be used imprecisely just as an instrument to attain their goals.^[1]
I’m also fairly certain the environmentalist was a counterfactual net positive, with Will Macaskill himself commenting on the role of climate advocacy in funding solar energy research and accelerating climate commitments in What We Owe The Future. However, I will admit that the anti-nuclear stance was exactly as dumb as you’ve implied, and it embarrasses me how many activists expressed it.
Re: Enemy of my Enemy
Personally, I draw a meaningful distinction between being anti-AI capabilities and pro-AI Safety. Both are strongly and openly concerned about rapid AI progress, but the two groups have very different motivations, proposed solutions and degree of epistemic rigour. Being anti-AI does not mean pro AI Safety, the former is a much larger umbrella movement of people expressing strong opinions on a disruptive, often misunderstood field.
1. ^
  Frontiers | “Listen to the science!”—The role of scientific knowledge for the Fridays for Future movement (frontiersin.org)

Minh Nguyen 8 Sep 2023 17:10 UTC
20 points
8 ∶ 0
in reply to: Vaidehi Agarwalla’s comment on: Sharing Information About Nonlinear
I think given what you know, your level of skepticism is reasonable here.
I mean, obviously, I’m disagreeing based on my subjective experience/knowledge. But these are reasonable concerns for an outside observer to have. My take is that how unreasonable this level of defensiveness is, would vary based on how true the actual claims are. If they’re say, 80% false, vs 80% true.
And honestly, even the most charitable interpretation states that Nonlinear team really dropped the ball on communicating to employees and frequently says a lot of weird, shady stuff. So I’m not gonna pretend like Nonlinear does nothing wrong, just because they’re “my team”.
I mean, for all I know, there’s 2 parties each claiming the other maintains a complex web of deception and lies, and I might be believing the wrong one 🤔
Guess we’ll find out.

Minh Nguyen 8 Sep 2023 9:07 UTC
39 points
5 ∶ 0
in reply to: Vaidehi Agarwalla’s comment on: Sharing Information About Nonlinear
Oh yeah, don’t take this as a direct refutation of Alice/Chloe’s accounts. I definitely agree that the context was different. If the claims are true, then yeah, that sounds really bad.
Re: Core Claims
For the “core claims”. I have a personal opinion that these claims started from unfortunate, honest misunderstanding, and were substantively exaggerated. But those claims are specific and very sensitive. Clearly, at least one party involved is misleading people. So I’ll let Kat/Emerson represent themselves with whatever evidence they showed me.
Re: Patterns of behaviour
I’m addressing parts of the post that cite other concerning behaviour. Maybe it’s unintentional, but this post references a lot of extra details that make the core claims feel much more believable as a pattern of behaviour. If this post was just about “Nonlinear abused this specific employee in this specific context”, that’s one thing. But this post says “Nonlinear abused employees, and they openly brag about how cutthroat/exploitative they are, and they tell employees their problems and time and personal life don’t really matter”. Hell, I’d be convinced.
Also, … I’m disagreeing with this conclusion on a personal level, as the hypothetical person described in this strongly-worded appeal:
I expect that if Nonlinear does more hiring in the EA ecosystem it is more-likely-than-not to chew up and spit out other bright-eyed young EAs who want to do good in the world. I relatedly think that the EA ecosystem doesn’t have reliable defenses against such predators. These are not the first, nor sadly the last, bright-eyed well-intentioned people who I expect to be taken advantage of and hurt in the EA/x-risk/AI safety ecosystem, as a result of falsely trusting high-status people at EA events to be people who will treat them honorably.
Obviously, small sample size, but I’m already 50% of the pool of people this statement applies to.
Re: Adorian Deck
Honestly, it didn’t even occur to me that Deck was worth addressing. I did a quick read of the post+public linked sources, and came to the conclusion that Deck was supposed to co-manage the account (as stated in the contract), became inactive managing it and regretted it afterwards when it became successful. It’s incredibly common, almost expected, for young content creators who go viral to become inexplicably inactive and neglect obligations, as I’ve already experienced multiple times in my generative AI venture. It happens like … 70-90% of the time you sign such deals.
Yes, maybe Emerson’s account is entirely fabricated, but I find it easier to believe that this is a very common dispute, and not “behaviour that’s like 7 standard deviations away from usual norms in this area”. I mean, mathematically, believing someone is behaving many standard deviations outside the norm is a bit harder than believing a teenager lost interest in maintaining their viral account, which happens most of the time.^[1]
I can elaborate, and even cite personal examples. But point is: it’s really common in industries that work closely with content creators/affiliates, which is why it didn’t register as a red flag for me.
1. ^
  In any case, Emerson directly refutes he sent stalkers. This claim sounds really hard to prove/disprove given the information presented, so I … didn’t want to go down that weird rabbit hole.

Minh Nguyen

Tractable idea: A list of helpful links sent to EAGx and EAG attendees post-conference.

Evaluation and Accountability

Media and Outreach

Hello! I’m looking for AI Safety proposals/​project ideas to deploy within CivitAI, the biggest open-source AI image model hosting platform.

Minh Nguyen’s Quick takes

Re: Founder pool

Re: Project pool

Re: Hacker culture

Re: Environmentalism

Re: Enemy of my Enemy

Re: Core Claims

Re: Patterns of behaviour

Re: Adorian Deck

Hello! I’m looking for AI Safety proposals/project ideas to deploy within CivitAI, the biggest open-source AI image model hosting platform.