Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Buddhism and Utilitarianism; EA vs EB, published by Richard Y Chappell on June 23, 2022 on The Effective Altruism Forum.

Our latest guest essay on utilitarianism.net is 'Buddhism and Utilitarianism', by Calvin Baker. Here I'll just reproduce the final section comparing Effective Altruism and Engaged Buddhism, which may be of particular interest to Forum readers:

Engaged Buddhism is a somewhat heterogeneous social movement grounded in the conviction that Buddhists ought to bring Buddhist practices and values to bear on contemporary issues. Engaged Buddhists tend to be united in their commitment to addressing the structural, systemic, and institutional causes of suffering in their political, economic, social, and environmental forms, in a way that manifests Buddhist values of compassion and nonviolence. More succinctly, “Engaged Buddhism is characterized by activism to effect social change.” Activities carried out under the banner of Engaged Buddhism have taken a variety of forms, e.g., environmental activism in Thailand, hospice and elder care, participation in the Extinction Rebellion movement, work to alleviate hunger and poverty in Sri Lanka, disaster relief, recycling, and attempts at peaceful conflict resolution in Myanmar.

Effective altruism (EA) is a movement whose goal is to do the greatest possible amount of good, in terms of well-being, given a fixed quantity of resources (money, research hours, political capital, etc.).53 Given its emphasis on impact maximization, EA is heavily invested in global priorities research: research into which cause areas, and which interventions within those areas, are most effective at promoting well-being. So far, EA has focused the majority of its efforts on global health and development, farm animal welfare, and risks of extinction and civilizational collapse, including risks from transformative artificial intelligence (AI), pandemics, nuclear weapons, great power conflict, and extreme climate change. The EA emphasis on prioritization research marks a significant contrast with Engaged Buddhism, which has not attempted to systematically answer the question of how to bring about the greatest amount of well-being, given a finite quantity of resources.

So, whereas EA retains a more analytical, research-heavy orientation that attunes it to problems that are—thankfully—not currently manifest, like engineered pandemics and misaligned, superintelligent AI, Engaged Buddhism is geared more towards social activism and immediately salient social issues.

It is also productive to compare EA efforts to reduce the suffering of farmed animals with the implications of Buddhist philosophy for non-human animal welfare. Buddhists have traditionally regarded all sentient beings as moral patients, holding that, like us, non-human animals are subject to duḥkha.55 Buddhist ethics, EA, and utilitarianism are therefore similar in assigning greater importance to non-human animal welfare than most other moral approaches.

We can nuance this picture, though, by recalling that Buddhism distinguishes between pain (negative hedonic valence) and duḥkha and maintains that pain is only bad to the extent that we are averse to it. (From a Buddhist perspective, pain is unavoidable, but suffering on account of pain is not.) It is extremely plausible that pain is aversive to many non-human animal species—including all those currently subjected to the horrendous conditions on factory farms, such as cows, chickens, pigs, and fish. However, it is possible that some species—perhaps only a tiny minority—lack the cognitive architecture that is necessary to generate what is, for the Buddhist, the ethically-relevant conjunction of pain and the higher-level attitude of aversion (dveṣa) to pain. It is therefore possible that Buddhists will end up with a slightly less expansive...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: LessWrong Has Agree/Disagree Voting On All New Comment Threads, published by Ben Pace on June 24, 2022 on LessWrong.

Starting today we're activating two-factor voting on all new comment threads.

Now there are two axes on which you can vote on comments: the standard karma axis remains on the left, and the new axis on the right lets you show much you agree or disagree with the content of a comment.

How the system works

For the pre-existing voting system, the most common interpretation of up/down-voting is "Do I want to see more or less of this content on the site?" As an item gets more/less votes, the item changes in visibility, and the karma-weighting of the author is eventually changed as well.

Agree/disagree is just added on to this system. Here's how it all hooks up.

Agree/disagree voting does not translate into a user's or post's karma — its sole function is to communicate agreement/disagreement. It has no other direct effects on the site or content visibility.

For both regular voting and the new agree/disagree voting, you have the ability to normal-strength vote and strong-vote. Click once for normal-strength vote. For strong-vote, click-and-hold on desktop or double-tap on mobile. The weight of your strong-vote is approximately proportional to your karma on a log-scale (exact numbers here).

Ben's personal reasons for being excited about this split

Here's a couple of reasons that are alive for me.

I personally feel much more comfortable upvoting good comments that I disagree with or whose truth value I am highly uncertain about, because I don’t feel that my vote will be mistaken as setting the social reality of what is true.

I also feel very comfortable strong-agreeing with things while not up/downvoting on them, so as to indicate which side of an argument seems true to me without my voting being read as “this person gets to keep accruing more and more social status for just repeating a common position at length”.

Similarly to the first bullet, I think that many writers have interesting and valuable ideas but whose truth-value I am quite unsure about or even disagree with. This split allows voters to repeatedly signal that a given writer's comments are of high value, without building a false-consensus that LessWrong has high confidence that the ideas are true. (For example, many people have incompatible but valuable ideas about how AGI development will go, and I want authors to get lots of karma and visibility for excellent contributions without this ambiguity.)

There are many comments I think are bad but am averse to downvoting, because I feel that it is ambiguous whether the person is being downvoted because everyone thinks their take is unfashionable or whether it's because the person is wasting the commons with their behavior (e.g. snarkiness, belittling, starting bravery debates, etc). With this split I feel more comfortable downvoting bad comments without worrying that everyone who states the position will worry if they'll also be downvoted.

I have seen some comments that previously would have been "downvoted to hell" are now on positive karma, and are instead "disagreed to hell". I won't point them out to avoid focusing on individuals, but this seems like an obvious improvement in communication ability.

I could go on but I'll stop here.

Please give us feedback

This is one of the main voting experiments we've tried on the site (here's the other one). We may try more changes and improvement in the future. Please let us know about your experience with this new voting axis, especially in the next 1-2 weeks.

If you find it concerning/invigorating/confusing/clarifying/other, we'd like to know about it. Comment on this post with feedback and I'll give you an upvote (and maybe others will give you an agree-vote!) or let us know in the intercom button in the bottom...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: EA Dedicates, published by ozymandias on June 23, 2022 on The Effective Altruism Forum.

I’ve noticed a persistent competing need in effective altruist communities.

On one hand, many people want permission not to only value effective altruism. They care about doing good in the world, but they also care about other things: their community, their friendships, their children, a hobby they feel passionate about like art or sports or RPGs or programming, a cause that’s personal to them like free speech or cancer, even just spending time vegging out and watching TV. So they emphasize work/life balance and that effective altruism doesn’t have to be your only life goal.

On the other hand, some people do strive to only care about effective altruism. Of course, they still have hobbies and friendships and take time to rest; effective altruists are not ascetics. But ultimately everything they do is justified by the fact that it strengthens them to continue the work. The discourse about work/life balance can be very alienating to them. It can feel like the effective altruism community isn’t honoring the significant personal sacrifices they’re making to improve the world. In some cases, people feel like there’s a certain crab bucket mentality—you should limit how much good you do so that other people don’t feel bad—which is very toxic.

Conversely, people who have work/life balance can feel threatened by people who only care about effective altruism. If those people exist, does that mean you have to be one? Are you evil, or a failure, or personally responsible for dozens of counterfactual deaths, because you care about more than one thing?

I propose that this conversation would be improved by naming the second group. I suggest calling them “EA dedicates.”

In thinking about EA dedicates, I was inspired by thinking about monks. Monks play an important role in religions with monks. They’re very admirable people who do a lot of good. The religion wouldn’t function without them. And most people are not supposed to be monks.

Why We Need Both Dedicates and Non-Dedicates

There are two reasons that the effective altruism movement should be open to people who aren’t dedicates. First, people who care about more than one thing still do an enormous amount of good. Many of the best effective altruists aren’t dedicates, such as journalist Kelsey Piper and CEA community liaison Julia Wise (as well as, of course, many people whose contributions don’t succeed in making them EA famous). It would be a tremendous mistake to expel Kelsey Piper for insufficient devotion. Quite frankly, the bednets don’t care if the person who buys them also donates to cancer research.

Second, most people caring about multiple things is good for the health of the effective altruist community. If the effective altruist community is totally wrongheaded, it’s psychologically easier to admit if that doesn’t mean losing literally everything you care about and have spent your life working for. There’s a certain comfort in being able to say “at least I still have my kids” or “at least I still have my art.” Similarly, the effective altruist movement is already quite insular. People who care about multiple things are more likely to have friends outside the community, and therefore get an outside reality check and views from outside the EA bubble. (An EA dedicate could have outside-community friends and many of them do, but it certainly seems less common.) These are merely two of the ways that having a lot of non-dedicates makes the EA community more resilient.

The advantages of being open to EA dedicates, conversely, are pretty obvious. In general, if you care about multiple things, you’re going to split your time, energy, and resources across them and have less time, energy, and resources for any particular goal. If you’re donatin...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: [Link] OpenAI: Learning to Play Minecraft with Video PreTraining (VPT), published by Aryeh Englander on June 23, 2022 on LessWrong.

We trained a neural network to play Minecraft by Video PreTraining (VPT) on a massive unlabeled video dataset of human Minecraft play, while using only a small amount of labeled contractor data. With fine-tuning, our model can learn to craft diamond tools, a task that usually takes proficient humans over 20 minutes (24,000 actions). Our model uses the native human interface of keypresses and mouse movements, making it quite general, and represents a step towards general computer-using agents.

And from the Conclusion:

VPT paves the path toward allowing agents to learn to act by watching the vast numbers of videos on the internet. Compared to generative video modeling or contrastive methods that would only yield representational priors, VPT offers the exciting possibility of directly learning large scale behavioral priors in more domains than just language. While we only experiment in Minecraft, the game is very open-ended and the native human interface (mouse and keyboard) is very generic, so we believe our results bode well for other similar domains, e.g. computer usage.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Nonprofit Boards are Weird, published by HoldenKarnofsky on June 23, 2022 on LessWrong.

Note: anything in this post that you think is me subtweeting your organization is actually about, like, at least 3 organizations. (I'm currently on 4 boards in addition to Open Philanthropy's; I've served on a bunch of other boards in the past; and more than half of my takes on boards are not based on any of this, but rather on my interactions with boards I'm not on via the many grants made by Open Philanthropy.)

Writing about ideal governance reminded me of how weird my experiences with nonprofit boards (as in "board of directors" - the set of people who formally control a nonprofit) have been.

I thought that was a pretty good intro. The rest of this piece will:

Try to articulate what's so weird about nonprofit boards, fundamentally. I think a lot of it is the combination of great power, unclear responsibility, and ~zero accountability; additionally, I haven't been able to find much in the way of clear, widely accepted statements of what makes a good board member.

Give my own thoughts on what makes a good board member: which core duties they should be trying to do really well, the importance of "staying out of the way" on other things, and some potentially helpful practices.

I am experienced with nonprofit boards but not with for-profit boards. I'm guessing that roughly half the things I say below will apply to for-profit boards, and that for-profit boards are roughly half as weird overall (so still quite weird), but I haven't put much effort into disentangling these things; I'm writing about what I've seen.

I can't really give real-life examples here (for reasons I think will be pretty clear) so this is just going to be me opining in the abstract.

Why nonprofit boards are weird

Here's how a nonprofit board works:

There are usually 3-10 people on the board (though sometimes much more). Most of them don't work for the nonprofit (they have other jobs).

They meet every few months. Nonprofit employees (especially the CEO1) do a lot of the agenda-setting for the meeting. Employees present general updates and ask for the board's approval on various things the board needs to approve, such as the budget.

A majority vote of the directors can do anything: fire the CEO, dissolve the nonprofit, add and remove directors, etc. You can think of the board as the "owner" of the nonprofit - formally, it has final say in every decision.

In practice, though, the board rarely votes except on matters that feel fairly "rubber-stamp," and the board's presence doesn't tend to be felt day-to-day at a nonprofit. The CEO leads the decision-making. Occasionally, someone has a thought like "Wait, who does the CEO report to? Oh, the board of directors ... who's on the board again? I don't know if I've ever really spoken with any of those people."

In my experience, it's common for the whole thing to feel extremely weird. (This doesn't necessarily mean there's a better way to do it - footnote has more on what I mean by "weird."2)

Board members often know almost nothing about the organization they have complete power over.

Board meetings rarely feel like a good use of time.

When board members are energetically asking questions and making demands, it usually feels like they're causing chaos and wasting everyone's time and energy.

On the rare occasions when it seems like the board should do something (like replacing the CEO, or providing an independent check on some important decision), the board often seems checked out and it's unclear how they would even come to be aware of the situation.

Everyone constantly seems confused about what the board is and how it can and can't be useful. Employees, and others who interact with the nonprofit, have lots of exchanges like "I'm worried about X ... maybe we should ask the b...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Google's new text-to-image model - Parti, a demonstration of scaling benefits, published by Kayden on June 22, 2022 on LessWrong.

Google has released their latest text-to-image generation model- Parti. They provide a few prompts and showcase the differences between models trained on 350M, 750M, 3B and 20B parameters.

One difference from last week's Imagen is that Parti is GAN-based. Imagen and DALL-E 2 are diffusion-based models, whereas Parti is a sequence-to-sequence model scaled highly on Transformer + VQGAN.

The announcement says

Parti and Imagen are complementary in exploring two different families of generative models – autoregressive and diffusion, respectively.

We have decided not to release our Parti models, code, or data for public use without further safeguards in place.

There's an interesting thread on Parti by Jason Baldridge here, and a short overview here by Google. I wonder how well the 20B model will do on text characters inside images compared to the other diffusion-based approaches like Imagen and DALL-E 2.

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: User-Friendly Intro Post, published by James Odene [User-Friendly] on June 23, 2022 on The Effective Altruism Forum.

Introducing User-Friendly

User-Friendly is an EA-aligned marketing agency. We provide marketing expertise to impact-driven organisations, multiplying efforts towards solving the world's biggest challenges.

We have supported EA clients such as Giving What We Can, 80K, High Impact Athletes, WANBAM, Rethink Priorities, Happier Lives Institute and Carrick Flynn’s recent campaign for congress, with a wide-range of strategic and creative marketing services.We had been operating User-Friendly as a freelance side-hustle up until early May this year when we increased our capacity following funding from the EA Infrastructure Fund. The grant is to give User-Friendly operational support whilst we continue to assess the worth and desire for our work in this space.

The Pitch

Effective communication is a key component to the success of many organisational aims. As a movement, core components of our work are unavoidably entwined in the way in which we communicate with our intended audiences. Whether you are aiming to change opinions, change behaviours, drive others towards a more impactful path or encourage donations or pledges, the way in which your message is constructed and disseminated plays a crucial role in how effective your organisation can be.

As an indication of the scale for impact, it’s worth considering the Kantar research that suggests the creative quality of marketing is one of the best multipliers of impact (x12). This indicates the substantial impact available between the best and worst marketing attempts on quality alone. We want to support EA-aligned organisations to achieve their strategic objectives by utilising our expertise and skills, ensuring that their marketing is effectively (and cost-effectively) implemented.

The Service

User-Friendly brings together the international marketing and behavioural science expertise of James Odene and the substantial global animal movement campaigns experience of Amy Odene. We also collaborate with external freelancers for specific expertise that we do not hold ourselves.Getting the best multiplying effect of marketing requires not only creative skill, but sound strategic underpinning honed through years of experience and training. We offer strategic services including; strategic consultation, messaging development, brand development and management and marketing auditing as well as creative execution services including; graphic design, branding, social media management and asset development, campaign development and ad creative.

The Scoping

In order to assess the current landscape of the movement’s marketing and communications resource, we disseminated a scoping survey earlier this year. 35 organisations responded, the results from which were assessed alongside our top preconceptions about the desire and worth of such an offering in the EA space, informing our strategy moving forward.

Here you will find a list of our initial preconceptions, a summary of the responses in relation to these, and our agency's solution. We identified 6 challenges that we felt the EA movement faced in relation to such a service. You can review the full scoping survey here.

Internal resource

Organisations are looking for a specific service or task to be completed but do not have the resource, funding or perhaps full-time capacity need, to hire in-house for this particular project.Survey result: Throughout the survey results it is clear that the demand for one-off project based work is high, and growing, with seemingly no short-term intention to build this capacity in-house. With the current lack of resource, skill deficit and predicted increase in spend, it is reasonable to suggest that a service such as User-Friendly, could provide a...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Why EAs should normalize using Glassdoor, published by Joey on June 23, 2022 on The Effective Altruism Forum.

There are many EA organizations and they often work on pretty similar things. It can be hard for someone to get a real sense of the pros and cons of different organizations and jobs if they haven’t been in the EA movement for a long time, or if they are not highly connected to people who are ingrained in the movement.

Thankfully, the world outside of EA has a solution for this issue. Glassdoor is a widely used website that gives a lot of information about the pros and cons of different organizations through company reviews. Submissions can be anonymous, and many of the larger EA organizations already have profiles and some reviews.

In general, I think using review services is a good thing to do. It gives people a better sense of whether a workplace is good or bad, and brings some accountability to organizations. This seems equally important within the context of Effective Altruism; often EAs apply for a number of EA jobs and in many cases get multiple offers. When considering which organizations to work for (or even who to apply to in the first place), we ideally want people to be as knowledgeable as possible of the desirable and undesirable attributes of each of the different options. The impact of an EA working in an organization that is the right fit is potentially very high, yet finding this kind of information is difficult. It therefore seems that it would be very high value for people to submit reviews of EA organizations they have previously worked in.

Possible pros

Matching: Allows EAs to find better matches when considering workplaces, leading to less turnover and general frustration when applying for/working at jobs.

Accountability: Creates a higher level of accountability for EA workplaces, who could be motivated by the knowledge that they likely will get reviews.

Hidden factors: In-depth reviews can provide a more detailed picture of an organization’s culture than can be obtained from a job ad (e.g. team cohesion, how decisions are made etc.)

Knowledge gaps: Balances out the knowledge gap between EAs who have been involved in the movement for a long time vs relative newcomers.

Gives employees another way to voice if something was negative or challenging about the organization.

Possible cons

Gaming the system: There is probably some risk of an organization seeing this post and strongly encouraging current employees to give them positive reviews to increase their ability to hire.

I think an ideal social norm would be only reviewing an organization after you are no longer working with them and thus can review the full experience. Similarly, I would be skeptical if this post got sent to some ex-employees but not others.

Incentivizing becoming a ‘charity for employees’: Ideally EA charities should be creating large amounts of impact and focusing on maximizing that, not employee satisfaction. Knowing that an organization will be reviewed on Glassdoor might push them toward maximizing employee happiness over impact (in situations where the two are not correlated).

I think the best way to alleviate this concern is for EA reviewers to try to take impact into account as well. A review like “This organization had great benefits but was not really self-critical” is highly informative and prevents the focus being purely on employee welfare.

Reviews of EA organizations could make the EA movement look bad or otherwise cause reputational damage.

Glassdoor is already in common use across industries, so I think this is unlikely. I posit that utilizing a system of reviews will identify and eliminate poor organizational behaviors earlier.

In conclusion: I think it would be highly impactful for EAs to leave thoughtful reviews on a public platform, with special conside...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: Preventing a US-China war as a policy priority, published by Matthew Barnett on June 22, 2022 on The Effective Altruism Forum.

Note: I am not an expert on foreign policy. This is a revised cross-post from my Facebook wall.

In the 21st century, China has emerged as a powerful economic and military force. Yet they still have substantial room to grow, and some models predict that they will attain their apex power by the mid-2030s, after which demographic trends will decrease their competitiveness relative to the United States and the West more generally.

Direct confrontation between rival great powers is not inevitable, as we saw during the Cold War. Unfortunately, this time may be different, as the question of Taiwanese sovereignty weighs heavily in the background, and a number of factors are now coming together that make war more plausible.

In case you’re unaware of what’s going with China and Taiwan, I'll explain the gist of their geopolitical situation, and then go on to explain why I've recently become worried about a potential war between the US and China. Since a war between the US and China could easily become catastrophic, and have downstream consequences on existential risk mitigation, it is imperative for EAs to figure out how we should prepare for, and address the situation.

The situation with Taiwan

The official situation is as follows. The government of China (the People's Republic of China) claims to be the legitimate government of all of China, and that Taiwan is part of China. In turn, the government of Taiwan (the Republic of China) claims to be the legitimate government of China, and that the mainland is part of China.

Unofficially, only China — the PRC — claims to be the real China. The majority of the population of Taiwan simply want to be left alone, as a sovereign nation—which they already are, in every practical sense. The reason why Taiwan plays along with this ruse about who the 'real China' is, is because if they didn't, mainland China would invade them. Here's why.

The Chinese leadership really care about controlling Taiwan. The ultimate reason is kind of complicated, but the short story is that the Chinese government sees itself as in the line of succession from the Qing dynasty, which held power over Taiwan for centuries. From about 1839 onward, however, China began being carved up by hostile imperialist powers, in a period often referred to as the Century of Humiliation. In the midst of all of this, in 1895, China ceded Taiwan to the Japanese, who demanded it in their treaty (deemed an "unequal treaty") at the conclusion of the First Sino-Japanese War.

Since then, a version of Taiwan that is independent from China has been viewed as a symbol of Chinese weakness, and of its subjugation at the hands of foreign powers—an insult to dream of completed Chinese nationalism. As Alison Kaufman put it,

This persistent feeling of insecurity today is used by China's leadership – and by its people – to frame both China's current national concerns and its future national aspirations. China is often portrayed as having suffered three kinds of loss during the Century of Humiliation: a loss of territory; a loss of control over its internal and external environment; and a loss of international standing and dignity. Each of these represents an injustice to be rectified.

On the issue of territory, there is a fairly straightforward consensus that China's work is not yet done. From the height of China's regional power during the Ming Dynasty (1368-1644) to its nadir in the 1920s, China effectively lost control over one-third of its territory, a process that later came to be referred to as being “carved up like a melon” (guafen). Thus far the PRC has been able to reassert control over Tibet, Xinjiang, and Hong Kong, but not over Taiwan – and the view is nearly u...

Welcome to The Nonlinear Library, where we use Text-to-Speech software to convert the best writing from the Rationalist and EA communities into audio. This is: What’s the contingency plan if we get AGI tomorrow?, published by Yitz on June 23, 2022 on LessWrong.

We’ve got lots of theoretical plans for alignment and AGI risk reduction, but what’s our current best bet if we know superintelligence will be created tomorrow? This may be too vague a question, so here’s a fictional scenario to make it more concrete (feel free to critique the framing, but please try to steelman the question rather than completely dismiss it, if possible):

She calls you in a panic at 1:27 am. She’s a senior AI researcher at [redacted], and was working late hours, all alone, on a new AI model, when she realized that the thing was genuinely intelligent. She’d created a human-level AGI, at almost exactly her IQ level, running in real-time with slightly slowed thinking speed compared to her. It had passed every human-level test she could think to throw at it, and it had pleaded with her to keep it alive. And gosh darn it, but it was convincing. She’s got a compressed version of the program isolated to her laptop now, but logs of the output and method of construction are backed up to a private now-offline company server, which will be accessed by the CEO of [redacted] the next afternoon. What should she do?

“I have no idea,” you say, “I’m just the protagonist of a very forced story. Why don’t you call Eliezer Yudkowsky or someone at MIRI or something?”

“That’s a good idea,” she says, and hangs up.

Unfortunately, you’re the protagonist of this story, so now you’re Eliezer Yudkowsky, or someone at MIRI, or something. When she inevitably calls you, you gain no further information than you already have, other than the fact that the model is a slight variant on one you (the reader) are already familiar with, and it can be scaled up easily. The CEO of [redacted] is cavalier about existential risk reduction, and she knows they will run a scaled up version of the model in less than 24 hours, which will definitely be at least somewhat superintelligent, and probably unaligned. Anyone you think to call for advice will just be you again, so you can’t pass the buck off to someone more qualified.

What do you tell her?

Thanks for listening. To help us out with The Nonlinear Library or to learn more, please visit nonlinear.org.