I live for a high disagree-to-upvote ratio
huw
Karma: 428
FWIW on timelines:
June 13, 2022: Critiques paper (link 1)
May 9, 2023: Language models explain language models paper (link 2)
November 17, 2023: Altman removal & reinstatement
February 15, 2024: William_S resigns
March 8, 2024: Altman is reinstated to the OpenAI board
March 12, 2024: Transformer debugger is open-sourced
April 2024: Cullen O’Keefe departs (via LinkedIn)
April 11, 2024: Leopold Aschenbrenner & Pavel Izmailov fired for leaking information
April 18, 2024: Users notice Daniel Kokotaljo has resigned
FWIW I find the self-indulgence angle annoying when journalists bring it up, it’s reasonable for Sam to have been reckless, stupid, and even malicious without wanting to see personal material gain from it. Moreover, I think leads others to learn the wrong lessons—as you note in your other comment, the fraud was committed by multiple people with seemingly good intentions; we should be looking more at the non-material incentives (reputation, etc.) and enabling factors of recklessness that led them to justify risks in the service of good outcomes (again, as you do below).
This is an extremely rich guy who isn’t donating any of his money. I wouldn’t call him ‘aligned’ at all to EA.
I would also just, be careful about reading him on his word. He’s only started talking about this framing recently (I’ve followed him for a while because of a passing interest in Kernel). He may well just be a guy who’s very scared of dying with an incomprehensible amount of money to spend on it, who’s looking for some admirers.
My sense from a very quick skim of the literature is:
There are barely any studies or RCTs on non-dual mindfulness, and certainly not enough to make a conclusion about it having a larger-than-normal effect size[1][2]
The most highly-cited meta-analyses that do split out types of meditation either directly find no significant difference between kinds, or claim they don’t have enough evidence for a difference in their discussions[1:1][2:1]
The effect size is no better or worse than other psychotherapies
It might be possible to do some special pleading around non-dual mindfulness in particular, but frankly, everyone who has their own flavour of mindfulness does a lot of special pleading around it, so I’m default skeptical despite non-dual being my personal preference.
My sense as an experienced non-dual meditator (~10 years, and having experienced ‘ego death’ before without psychedelics):
I am skeptical that at-will or permanent ego death is possible. By ‘at-will’, I mean with an ease similar to meditating, with effects lasting longer than an acid trip.
I am skeptical that this state would even be desirable; most people that have tried psychedelics aren’t on a constant low dose (despite that having few downsides for people not prone to psychosis).
Even if it is possible and desirable, I am skeptical that there is a path to this kind of enlightenment for every person, and it might only be possible for a very small percentage of people even with the motivation and infinite free time to practice
I think teaching people mindfulness would be good, but probably no better than teaching them any other kind of therapy. Maybe it’s generally more acceptable because it’s less stigmatised than self-learning CBT. But I’d be really curious to understand what the people who voted yes were thinking, and in particular what they think ‘enlightenment’ is.
Just doing some napkin math, the ruling coalition lead by about 400,000 votes. So that’s approximately 6,000 EUR lost per vote. It’s obviously not all going to be spent cost-effectively & doesn’t consider counterfactuals, but that’s a good marker for how much someone should be willing to spend to swing those votes.
Can you give a sense of what proportion? Should we expect ‘some’ to mean ≤10% or something more significant?
Prioritizing longtermist over neartermist causes was predicted by higher levels of engagement, male gender, white ethnicity, and younger age (when accounting for the influence of these and several other variables in the same model).
This raises some hackles in me, gender & race are both dimensions of privilege (‘based in US’ is a much less clear marker of privilege in this context). This is speculation, but I wonder if psychological distance from suffering is a strong predictor of longtermist beliefs; i.e. that suffering yourself or being in a context of suffering leads you to prefer near-term causes. This, at least, is one thing I think critics of longtermism take issue with (at their strongest & in best faith).
I’m wary of EAs performatively self-flagellating and accepting more responsibility for the FTX thing than is warranted (given, e.g., that huge numbers of people with a very direct financial incentive in spotting FTX’s fraud didn’t spot it, so it’s obviously not weird if random EAs failed to spot it).
I don’t think this is about spotting fraud at all. I think the strong case for EA responsibility goes like:
EA promoted earning to give
When the movement largely moved away from it, not enough work was done to make that distance (such that the average person still strongly associates EA with earning to give)
Sam adopted an extremely reckless approach to EV-maximisation that was highly likely, if not guaranteed, to lead to a severe loss in value, regardless of the illegality of that loss
He publicly branded himself with this reckless EV-maximisation approach
He publicly associated himself with the movement, with earning-to-give specifically, and donated lots of money to EA causes to prove it
Many EA causes accepted that money and/or associated themselves with Sam
Critics of Sam’s recklessness did not speak up or were drowned out by other community members
The counterfactuals look something like this:
If EA had strongly distanced themselves from ETG, Sam may not have considered it as a path, or been unable to tie himself to EA
If EA had criticised Sam’s recklessness this may have moderated it
If EA orgs conditionalised their acceptance of his donations on financial transparency this may have mitigated the recklessness (and this would have been in their own interests since they stood to lose money!)
FWIW, I don’t necessarily agree with this, but this is what feels either explicit or implicit in a lot of the EA criticism around this that I’ve read. The focus appears to be explicitly on the causal chain of Sam adopting a reckless flavour of ETG which motivated him to take risks which didn’t pay off, and what EA, which was surely influential over his thinking, could’ve done to prevent it.
You’re right! It’s not that ETG is inherently bad (and frankly, I haven’t seen anyone make this argument), it’s that specific EV-maximising interpretations of ETG cause people to pursue careers that are (1) harmful, (2) net harmful, or (3) too risky to pay off.
Personally, I think FTX was (1) and (3), and
unlikely to be (2)probably also (2). I’m not really sure where the bar is, but under any moderately deontological framework (1) is especially concerning, and many of the people EA might want to have a good reputation with believe (1). So that’s roughly the worldview-neutral case for caring about strongly rejecting EV-maximising forms of ETG.
It’s wild for a news organisation that routinely witnesses and reports on tragedies without intervening (as is standard journalistic practice, for good reason) to not recognise it when someone else does it.
Perhaps this is a cop-out, but ‘socialism’ generally just means the reduction of private capital. This doesn’t necessarily mean more government or authority.
In fact, I think you’d probably agree with a lot of socialist anti-imperialism, such as replacing the IMF, World Bank, and Belt & Road with grantmaking organisations (these don’t require enforced austerity or other impositions of power when the recipients don’t make repayments). Equally, without uniform U.S. interventionism in every left-wing revolution since WWII, we may have seen some of these economies flourish. Both of these stances are motivated by an analysis of the West’s involvement as being motivated by protection or enhancement of private capital (whether you buy that or not), but I do agree that the original post could’ve clarified this further.
It feels like you’re arguing for a higher-than-necessary level of harm and suffering in the world, just because a high level of suffering already exists in this context? I can’t see an argument with this structure working anywhere else (and believe me, I think Sam should be punished).
The current board is:
Bret Taylor (chair): Co-created Google Maps, ex-Meta CTO, ex-Twitter Chairperson, current co-founder of Sierra (AI company)
Larry Summers: Ex U.S. Treasury Secretary, Ex Harvard president
Adam D’Angelo: Co-founder, CEO Quora
Dr. Sue Desmond-Hellmann: Ex-director P&G, Meta, Bill & Melinda Gates; Ex-chancellor UCSF. Pfizer board member
Nicole Seligman: Ex-Sony exec, Paramount board member
Fidji Simo: CEO & Chair Instacart, Ex-Meta VP
Sam Altman
Also, Microsoft are allowed to observe board meetings
The only people here who even have rumours of being safety-conscious (AFAIK) is Adam D’Angelo, who allegedly played a role in kickstarting last year’s board incident, and Sam, who has contradicted a great deal of his rhetoric with his actions. God knows why Larry Summers is there (give it an air of professionalism?), the rest seem to me like your typical professional board members (i.e. unlikely to understand OpenAI’s unique charter & structure). In my opinion, any hope of restraint from this board or OpenAI’s current leadership is misplaced.
I tend to think the question is less ‘should EAs advocate for socialism?’, and more practically, ‘should EAs and socialists collaborate?’. Framing it in the former way is only going to highlight the very real differences between both movements. We can learn from the history of leftist disunity that historically, leftist movements have been strongest when they focus on finding ways to collaborate rather than persuading each other of their exact flavour of the ideology.
I think the collaboration between EA and existing animal welfare ideologies is a great template for what I mean here.
Contra tractability specifically, this depends on a lot on what societies count as ‘socialist’ or not. A lot of European countries, for example, are essentially market socialist or hybrid economies (large social states as in the Nordics, strong or mandatory labour unions as in Germany, large cooperative companies as in Spain, widespread state ownership of industry as in the Nordics or Poland). These countries have the longest histories of socialist movements, and have all achieved these economies through reform.
And OP discusses market socialist systems which allow capital markets but not private capital!
This isn’t a petty distinction. It allows the definer to claim all of the benefits of markets and dodge the more negative effects of private ownership, pitting centralised price controls as inherent to anti-capitalist systems. And in the worst cases (not here) it allows people to motte-and-bailey their way out of the devastating effects of wealth inequality by claiming that ‘capitalism’ actually just means markets.
I mention all this because I see this definition a lot in rat-adjacent circles and it frustrates me, because people usually just want to talk about why disgusting levels of wealth inequality are necessary or even permissible, and then get a non-sequitur defence of markets in response.
To make it concrete, the OP’s friends are interested in economic inequality. This is absolutely an inherent consequence of private capital ownership, and therefore capitalism. In a debate, then, you’d want to start defending private capital ownership rather than markets. So I think the ‘talking past each other’ arises from a faulty definition, but just not the one that the OP identified.
Global Burden of Disease (GBD) is okay, it depends a lot on what disease & metric you’re looking at, and how aware you are of the caveats around it. Some of these:
A lot of the data is estimated, rather than real measurements of prevalence
I think most people understand this, but it’s always worth a reminder
The GBD provides credible intervals for all major statistics and these should definitely be used!
This paper on the Major Depressive Disorder estimates is a good overview for a specific disease
The moral weights for estimating the years lived with disability for a given disease are compiled from a wide survey of the general public
This means they’re based on people’s general belief of what it would be like to have that condition, even if they haven’t experienced it or know anyone who has
Older GBDs included expert opinion in their moral weights, but to remove biases they don’t do this anymore (IMHO I think this is the right call)
The estimates for prevalence are compiled differently per condition by experts in that condition
There is some overall standardisation, but equally, there’s some wiggle room for a motivated researcher to inflate their prevalence estimates. I assume the thinking is that these biases cancel out in some overall sense.
Overall, I think the GBD is very robust and an extremely useful tool, especially for (a) making direct comparisons between countries or diseases and (b) where no direct, trustworthy, country-specific data is available. But you should be able to improve on its accuracy if you have an inside view on a particular situation. I don’t think it’s subject to the incentives you mention above in quite the same way.
I haven’t run the numbers myself but I generally assume that FTX’s account-holders were mostly moderately well-off HIC residents (based on roughly imbibed demographics of crypto), and the Future Fund’s beneficiaries are by and large worse off. There were probably some number of people who invested their life savings or were otherwise poor to begin with that were harmed more significantly then the beneficiaries of their money. But on the whole it feels like it was an accidental wealth transfer, and much of that harm will be mitigated if they’re made whole (but admittedly, the make-whole money just comes from crypto speculation that trades on the gullibility of yet more people).
But much less confident in this take; my point is much more around the real harms it caused being worth thinking about.
The general thrust of my take here is that most economies are a hybrid of pro- and anti-property systems, so it’s hard to disentangle a lot of it. But I’ll have a crack:
the post above outlines the USSR’s space program or China’s economic growth as examples of socialist successes, so it must be to do with the ‘socialism’ in those economies
Oh, I read these as counter-examples to the theory that capitalism is a superior economic principle; that whatever causes these things could be independent of either. But equally, China’s economic growth might be more attributable to their capitalist tendencies than their socialist ones.
Your definition sounds like capitalism to me: you can pay rent to a landlord, have your surplus labour taken, the only condition is that ‘private capital’ is being ‘reduced’
It’s all a spectrum, right? No system is entirely ‘capitalist’ either, in the sense that not all forms of capital are subject to private ownership for profit. So I was looking more at decreasing private capital relative to the status quo, rather than abolishing it entirely. It wouldn’t be beyond the pale to refer to such acts as ‘socialist’.
being intergovernmental organisations, I don’t believe IMF and World Bank are ‘private capital’
Given that almost every capitalist government raises some of its own capital by borrowing money from private citizens, yes, these organisations partially serve private capital. I can’t speak to the extent, untangling global debt flows seems hard.
In other words the attributes pro-imperialism and pro-capitalism are independent.
I think that’s fair. I was speaking toward the tendency of capitalist systems to exhaust themselves of growth in their places of origin, and then start looking outward. The usual example leftists like to trot out here is the Dutch East India Company (not without reason!). I think you might be able to be capitalist and anti-imperialist, but it’s harder work and you might find yourself out-competed by those willing to extract the extra value by doing so.
A meta thing that frustrates me here is I haven’t seen much talking about incentive structures. The obvious retort to negative anecdotal evidence is the anecdotal evidence Will cited about people who had previous expressed concerns who continued to affiliate with FTX and the FTXFF, but to me, this evidence is completely meaningless because continuing to affiliate with FTX and FTXFF meant a closer proximity to money. As a corollary, the people who refused to affiliate with them did so at significant personal & professional cost for that two-year period.
Of course you had a hard time voicing these concerns! Everyone’s salaries depended on them not knowing or disseminating this information! (I am not here to accuse anyone of a cover-up, these things usually happen much less perniciously and much more subconsciously)