Hacker News Reader

I agree that David Gerard is more bombastic and leaps to more conclusions than I would prefer. But I follow this story a little and the connections he and Amy Castor draw here make a lot of things make more sense.

The "staff psychiatrist" at FTX is real. It seemed like an insane indulgence when I read about it, but I didn't follow to the obvious conclusion, that it was to prescribe medications with a subverted doctor. And that implies that any Adderall abuse wasn't just a personal vice; FTX believed in said medications being crucial to its success as an organization. It's not hard to find "rationalist" folks who have the same ideas.

Is there a perfectly straight line between Scott Alexander's writings and some alleged Adderall abuse of a particular EA advocate? I don't know, but there's probably contributing factors there.

That said, your information on the "rationalist" and "effective altruist" cult is out of date. While there might be a lot of people who still cling to its earlier more modest ambitions, the core of the movement changed. It's no longer about choosing charities that are effective in the here and now, but about projecting consequences into a kind of science fiction eschatology. There are many stories online of rationalist/EA groups which descended into worship of big-brained leaders as the patriarchs of the coming galactic civilizations.

Beliefs like this may have made SBF far more dangerous. One interpretation of his actions is that he justified building a massive fortune from fraud in the present, because he planned to save billions of lives from rogue AI in the future.

I appreciate Gerard's framing even if it's a bit overstated. Because it kicks me out of futilely pondering "how could a smart person have done any of this?" and into "ok, this is obvious tweaker fantasy and the real question is how he bamboozled so many into thinking he was the adult in the crypto community".

> While there might be a lot of people who still cling to its earlier more modest ambitions, the core of the movement changed.

I know this is true, and I'm totally willing to admit that my out-of-date impressions of the "movement" are probably coloring my resistance a bit. I've seen some of the "longtermism" stuff and it's crazy to see what was once a legimately "rational" community devolve in that manner. SBF is definitely an example of the worst of the dumb ideas associated with the movement (at least, if you take him at his word).

But dammit, there's still reasonable ways to engage with these ideas, and I hate to see a perfectly good baby thrown out just because it took a shit in the bathwater.

> Is there a perfectly straight line between Scott Alexander's writings and some alleged Adderall abuse of a particular EA advocate? I don't know, but there's probably contributing factors there.

There might be, and there might not be. I'm sure they read the article in question. Just doesn't seem fair to me to take reasonably ... reasonable writing from Scott and then somehow lay SBF's drug abuse at his feet.

I assumed the "54 billion lives" thing was a joke, and I make that assumption because I find Scott to be generally smart and reasonable. If he meant it seriously, well, no mincing - that's pretty dumb.

> Beliefs like this may have made SBF far more dangerous. One interpretation of his actions is that he justified building a massive fortune from fraud in the present, because he planned to save billions of lives from rogue AI in the future.

I have a really hard time buying that SBF truly thought he was some kind of hero here, but that may be my own bias. I'm somewhat reminded of the back-and-forth about Islamism that was in vogue 5-10 years ago, where folks just couldn't seem to believe that groups like ISIS actually believed all that jihad stuff. So I might be making the same kind of error.

Maybe he really does see himself as some kind of AI savior - but it still feels equally reasonable that he just identified a way to convince nominally smart people to give up their money to enrich himself.

EA absolutely means this shit. William MacAskill, who got SBF into EA, is extremely explicitly working for 10^54 future emulated humans, and not for anything so tawdry as existing humans.

I wrote this up earlier in the context of SBF, with cites: https://davidgerard.co.uk/blockchain/2023/02/06/ineffective-...

Picture it with a South Park caption: THIS IS WHAT "EFFECTIVE ALTRUISTS" ACTUALLY BELIEVE

> William MacAskill, who got SBF into EA, is extremely explicitly working for 10^54 future emulated humans

This is where I betray how out of the loop I am on this, I suppose. Last I heard from MacAskill was like a decade ago on Sam Harris's podcast, where he seemed to basically just be an acolyte of Singer, making the whole "would you ruin your suit to save a drowning child" argument.

Guess this is why people don't like utilitarianism, it sure lets you make some silly arguments.

> I assumed the "54 billion lives" thing was a joke

The specific numbers are a joke; but the idea behind it is earnest - the idea that small choices made by people in the movement, might have enormous long-term consequences. e.g. see "most important century".

> One interpretation of his actions is that he justified [...]

That's the interpretation that SBF himself has said.

> I really don't understand the hate boner everyone has for effective altruism - the whole core of the idea was just to choose charities that get the most value for the money donated

I admit to not doing in depth research but suspect the problem is with all that goes beyond that core idea. I really doubt that many people are genuinely complaining that someone decided to measure how effective are modern wood stoves at reducing CO2 emissions.

It's the fact that Sam Bankman-Fried could and did use the ideology as cover. The actual movement taking his money and giving him cover. Some of the weirder longtermist stuff that is at the minimum much harder to justify as charity to the everyman. Earning to give. The demographics of the movement. And a general feeling that at least the part of EA that Sam related to is a 21st century version of an ideology allowing the powerful to get an excuse for whatever they feel like doing from the present day equivalent of the priesthood.

> But just because SBF was a fraud doesn't mean the whole EA movement is bullshit,

I think a lot of people heard about EA because of and in relation to Sam Bankman-Fried. Of course people are going to find the message suspect when the first messenger turns out to be a fraudster using it as cover.

edit: and I say this as someone who is at least a bit concerned that this core idea may get hit in the crossfire.

> And a general feeling that at least the part of EA that Sam related to is a 21st century version of an ideology allowing the powerful to get an excuse for whatever they feel like doing from the present day equivalent of the priesthood.

I think this is the core of what irks me of the movement: it's one thing that their reasoning is often simplistic and arrogantly assumes you can distill the entire world, including its far future, down to some simple formulas - but they are also far too attracted to "repugnant conclusions" than for me to be comfortable.

I find it suspicious how much of their reasoning just happens to justify assholish behaviour and promotes a social darwinist survival-of-the-strongest worldview, all in the name of altruism.

> Earning to give.

This is exactly the kind of "baby with the bathwater" thing that bothers me - when I hear "earning to give" I don't think of SBF, or LessWrong, or even Slate Star Codex - I think of Peter Singer, or Will McAskill. The "philosophical" angle is where I first encountered the concept, not the dumb "rationalist" stuff about how preventing AI apocalypse now counts as charity to all the projected lives across the recovered future of humanity.

And I get it, it's not necessarily fair or accurate to separate those two origins when they've become so intertwined, but I share your concern about a genuinely good, righteous core idea getting tossed out because one public figure who espoused it turned out to be a crook. (Or because an army of weird nerds on LessWrong does a bunch of mystical fantasy "utilitarian math" on their weird little forum).

There's a lot to Singer's basic idea, I concur.

There's not to MacAskill's - he's loudly and explicitly into 10^54 hypothetical emulated future humans over anyone alive and needing help now.

And fundamentally, you can't get away from the EAs being the guys who spent the charity money on not one, but two castles.

Every atrocity in history was justified as being for the greater good.

> This is exactly the kind of "baby with the bathwater" thing that bothers me - when I hear "earning to give" I don't think of SBF, or LessWrong, or even Slate Star Codex - I think of Peter Singer, or Will McAskill. The "philosophical" angle is where I first encountered the concept, not the dumb "rationalist" stuff about how preventing AI apocalypse now counts as charity to all the projected lives across the recovered future of humanity.

I feel a lot of us here in this thread heard about this stuff before, but I get the impression EA became famous because of Bankman-Fried, so yeah.

As for specifically earning to give, yeah it also has defensible roots. But the idea of doing anything that pays a lot of money, even if it's something whose contribution to society is not very good or even negative, can easily go wrong in my opinion. And let's get real here, how many people will actually give the money after they get rich or get a high income and not change their attitude or fall into the hedonistic treadmill? Even for the people who are genuine about earning to give and not just using it as cover for proudly pursing their own material wealth and buying some Caribbean real estate?

Representative quote from the article:

> “by my calculations, [not using adderall] decreased Kelsey’s effectiveness by 20%, thus costing approximately 54 billion lives.” Yeah, rationalists really do say that sort of thing as if it’s normal.

If I remember their reasoning correctly, the billions of lives here are potential future humans that would have been born some milennia from now, had Kelsey just gotten over herself and taken Adderall. But in longermist logic, a future human that never was born is equivalent to the violent death of an actual real-life human. However as of course there are infinitely many more potential future humans than actual living humans, preventing future humans from being born is in practice a much greater offense for longtermists than letting real-life humans die.

That being said, I think the article conflates a bit "rationalists", "effective altruists" and "longtermists". I'm pretty sure the boundaries between the groups are very blurry, but I guess for reasons of fairness one should still make an attempt to distinguish them.

I assure you that the first and the second/third are substantially the same people at different times of day, and the third is the noisiest subgroup of the second.

Yeah. I agree with most of what David Gerard writes, but this attack on Scott Alexander was weak. The "54 billion lives" thing was obviously a joke.

I don't accept that as an excuse - because if it is a "joke", then it's one that only works if they say this stuff all the time seriously.

It's the species of "it's a joke!" that's serious until someone calls them out on it, then they retreat to claiming it was just a joke.

> I really don't understand the hate boner everyone has for effective altruism

I don't think people object to the core idea of EA. I think people object to how the prominent people in it actually behave and a whole lot of the ideas they espouse. A lot of that stuff is really very noxious and harmful, topped with a healthy helping of elitism.

People don't like it because it reads more like cult-lite than a coherent worldview or philosophy, whatever the core is. Goobers like SBF show that the reading isn't totally off base.