Weird Internet Communities

[Othko97]

I greatly enjoyed that Molly White piece, her writing on the cryptocurrency world in general is well worth reading.  It's almost surprising to see such clear and readable code for doing fraud---you'd expect something more obfuscated than an "infinite money" flag.

On the topic of Sam Bankman-Fried, the linked tweet in https://quomodocumque.wordpress.com/2023/10/08/underestimating-shakespeare-and-real-numbers/  to me exemplifies the nonsense within the over-quantification that comes in "rationalist" spaces.  It's utterly alien to me that the fact that it's unlikely that Shakespeare is literally the greatest writer that has and will ever be is taken as a reason to dismiss his works.  As the article itself suggests, I don't think we can assign a real number to "literary greatness," and even if we could there's still plenty of reason to read works not by the Empirical Greatest Writer of All Time.

[dubsartur]

I can't see that quote because its a screenshot behind a scriptwall on twitter and not archived on nitter.

One of the many reasons I wonder what happened to education in the USA is that "garbage in, garbage out" is classical computer science (like Charles Babbage is supposed to have made a joke about it and technicians were including it in their 'introduction to electronic computing' for journalists in 1957).  The rationalists often have little or no formal education in probability, statistics, or computer science and the crypto grifters often know even less about the technical side of blockchain but Sam Bankman-Fried is supposed to have an undergraduate degree in physics.  But making up numbers to express opinions in pseudo-statistics is fashionable among the rationalists, and SBF was trying to appeal to them or blend in with them. 

Which gets us back to how these communities are interlinked, conduct money and influence, and often have a front which looks mostly harmless and a covert wing which is dubious or criminal.  Where most rationalists and the effective altruists stand on a "useful idiots to swindlers" spectrum is a judgment call but there are groups at the core of these spaces with clear, sinister goals.

[Jubal]

I can't see that quote because its a screenshot behind a scriptwall on twitter and not archived on nitter.

Voila, here it is

[dubsartur]

People in these weird Internet communities tend to simultaneously dismiss the value of education and worship pop science books, so suggesting that going to university makes you a better writer would be controversial there. But SBF was just a clever superficial dude high on proscription stimulants running a bunch of fraud and embezzlement so I'm not sure how much thought went into that tweet.

In general, I see tweets as the equivalents of what a drunk friend says in private at 11 pm.

Edit: Scott Alexander essay from 2017 describing how many US persons in finance, competitive exam-taking, etc. want adderall; others in the rationalist subculture experimented on themselves to self-optimize with various substances https://slatestarcodex.com/2017/12/28/adderall-risks-much-more-than-you-wanted-to-know/ (Alexander did not discuss the common side effect of impulsivity, just uncommon side effects such as psychosis and Parkinson's disease, and he said that in his professional opinion as a psychiatrist adderall will help most people focus)  FTX's pet psychiatrist had some weasel words about how his patients among the staff used these drugs at similar rates to other finance workers.

David Gerard misrepresents this in service of a higher truth as "he (= Scott Alexander) told the rationalist subculture it  made you into a financial genius" ("focus better" is not the same as "genius").  IME he does this a lot; he knows these people's Internet posts and the Internet gossip about their private lives, but you can't trust him on the details.

I hope the fact that a subculture that grew out of Southern California encourages experimenting with mind-altering substances and unconventional sexual relationships does not surprise readers.  Or that this movement has gurus and Leaders who get accused of taking financial and sexual advantage of their position.

Edit: one of the American tech billionaires just successfully seeded a manifest onto Hacker News only for the posters there to notice "hey, he cites the Futurist Manifesto as a model and a lot of their followers became Fascists!  And if 'Our enemy is the ivory tower, the know-it-all credentialed expert worldview, indulging in abstract theories, luxury beliefs, social engineering, disconnected from the real world, delusional, unelected, and unaccountable – playing God with everyone else’s lives, with total insulation from the consequences.' how is that different from venture capitalists deciding who to shower with money and connections and publicity because they worked for the right company in 1997?"  So even the wannabe Silicon Valley startup people are getting the message.

[dubsartur]

One useful thing from Gerard: SBF spilled his guts in a text thread with journalist Kelsey Piper (currently at Vox). Piper promptly published part of the exchange.

Piper was one of Scott Alexander's examples of someone who desperately needed Adderall but was kept from it because people with ADHD have trouble completing long bureaucratic procedures, and he had her on his blogroll under Rationality.  So she seems to have been part of the SoCal rationalist subculture in 2017.  And SBF has complained that he is not getting the heavy dose of adderall which he was accustomed to.  So SBF texted indiscreet things to a fellow traveller, who was also a journalist and shared his interest in proscription stimulants, not a random journalist.

This is a small densely-connected space and the connections are not always obvious to a casual observer even if they are documented in endless detail online.  I already discussed how little interest the media had in exploring what Dominic Cummings' connections to the rationalists implied about his ideas and associations, because those were blog posts not tweets.

Edit: I also note that the online-and-tabloid, Moscow-based newspaper The eXile (1997-2008) had both journalist Matt Taibbi (who I don't know but seems to have slipped into that Glenn Greenwald space 'calls himself a lefty but really likes Putin, Assad, and conspiracy theories about lefty Anglos') and pottymouthed commentator John Dolan alias Gary 'War Nerd' Brecher (today,YouTuber LazerPig has a similar voice).  Dolan supposedly taught as an adjunct English professor in BC circa 2006-2008 and was interviewed by understimulated Internet racist Steve Sailer in 2003.  Around that time Christopher Wylie the Cambridge Analytica whistleblower was leaving school in the same city.

[Jubal]

I genuinely wonder if a prosopography of internet thinkers and movers and shakers of the 2000s-2010s would actually be a useful piece of research work for someone to build (I'm not volunteering, I should hasten to add, though it's the sort of project where I'd happily give thoughts on the data modelling).

[dubsartur]

I genuinely wonder if a prosopography of internet thinkers and movers and shakers of the 2000s-2010s would actually be a useful piece of research work for someone to build (I'm not volunteering, I should hasten to add, though it's the sort of project where I'd happily give thoughts on the data modelling).

I think some people tried something like that with their projects on Rational Wiki, but they didn't have any thoughts on visualization or organization, and they didn't have a Rankean goal of just figuring out who was connected to whom. Jon Evans, "Extropia’s Children" (2022) tried to be a prose account in the style of Nevala-Lee's Astounding!

Edit: the Pinkerite blog has an essay with diagrams on what it sees as Steven Pinker's racist and far-right connections (they make a good case that Pinker is intrigued by race 'science' and quietly boosts some people who believe in it)

Piper's Vox article certainly did not spell out that she had social connections to Sam Bankman-Fried, although she describes her section of the newsroom as "effective altruism-inspired".

I'm just trying to understand how these weird Internet people were so much more successful at offline networking and getting resources for big projects than my weird Internet people.  Ok "do crimes" and "lie a lot" can help for getting access to resources!

Edit: a few years ago I would have pegged all these spaces as somewhere between the Los Angeles SF Society in the 1960s and the kind of people who hang out in comment threads, as big on talk and unconventional opinions but mostly harmless. There are worse things in the world than shy people with opinions I think are bad. But it seems like they got their hands on some levers of power.

Edit: anyone familiar with the rationalists sees Piper's page at Vox and her Twitter account, sees "Tyler Cowen Bryan Alexander AI risk existential risk YIMBY effective altruism," let alone the endorsement by Scott Alexander, and hears "here be a rationalist." And that is obviously relevant to a story where an EA- and rationalist-linked fraudster sends her texts confessing to misdeeds. But it took another clever angry lonely person in Australia to spell out the connection.

Edit: David Gerard's blog post has a beautiful sad moment where after criticizing rationalists for disagreeing with (all?) psychologists about IQ, and criticizing the rationalists as lacking offline achievements, he has to criticize psychiatrist Scott Alexander for talking about Adderall in the wrong way.  To my knowledge Gerard has no verifiable expertise in psychiatry, he is just a sysadmin and journalist.  Gerard thinks offline achievements and expertise are valuable until someone he hates has them (and because he obsessively follows the rationalists' internet posts but seems to rarely attend their parties, he blames a lot on Internet posts and less on the fact that drugs are popular in the USA and Adderall and cocaine are especially popular with bankers in the USA and "its medicinal!" is a really popular excuse for getting zonked)

I have seen this kind of pseudo-skepticism, where people cite 'the authorities' when they mean Wikipedia or what a friend told them experts think, a lot since smartphones came out.  Gerard rightfully criticizes the rationalists for reading pop science books and thinking that makes them experts, but that does not stop him from pronouncing about the G/IQ construct with the confidence of Eliezer Yudkowski.

[dubsartur]

Patrick McKenzie has a good long rant about how on one hand cryptocurrency scams grow out of the small worlds of Internet subcultures and colourful characters, but on the other hand they spread through Anglo institutions like cholera through an army camp.  Neither journalists, not police forces, nor financial regulators, did much to stop them; two Canadian pension funds invested sums in the $100 million range in crypto companies which any graduate of an accounting program could see were frauds.  Different fraudsters tried to buy small Pacific islands with UN membership, and got both major US parties to start proposing legislation written by the fraudsters by convincing each party that the fraudsters would give them lots of money and let them crush the other party.  Even now, the CBC of all places is printing stories full of quotes from crypto advocates which don't mention the stream of bankruptcies and convictions for defrauding people of billions of dollars (this story was published one day (!) after one of those convictions of a fraudster who cost a Canadian pension fund big money). 

So on one hand there are stories about people like "had Kelsey Piper had informal chats with SBF before he sent her those indiscreet texts?" or "how close was Dominc Cummings to these spaces?" but on the other hand there are stories of how some confidence men with the right lingo can run wild through established institutions in the rich Anglo countries. Most of us as individuals just have to know to keep away from anyone associated with rationalism, LessWrong, longtermism, longtermist Effective Altruism, blockchain, cryptocurrency, NFTs, etc., but our institutions struggle to do that much.

https://www.bitsaboutmoney.com/archive/a-review-of-number-go-up-on-crypto-shenanigans/

Edit: Molly White has a handy list of some of the aspects of crypto fraud which are open for investigation.  There are 80 billion Tether tokens in circulation, each theoretically worth 1 USD, but since a theft in 2016 (!) they have not had enough actual money or money-like assets to redeem those.  Some was stolen, some was embezzled, some is in bank accounts frozen by governments that don't like money laundering, and some may have never existed at all- their current wording gives them room to back Tether tokens with other cryptocurrency or equity in other crypto companies rather than boring old US dollars that their customers can pay their boring old taxes and mortgages with. And this is the version of events that they have more or less admitted in court, its possible that the rot goes further back because crypto companies are run by incompetent people and crooks and crypto software makes it easy to, for example, lose large amounts of cryptocurrency on an old hard drive.

https://newsletter.mollywhite.net/p/the-stones-left-unturned

If you assume that most of those 80 billion Tethers are not actually backed with anything you can sell for USD, then you have two to four Madoffs worth of bezzle (money which the mark thinks they have but the embezzler has already stolen)
Tether alone could be a USD 80 billion bezzle (the gap between when the embezzler has stolen the money, and the victim realizes it is missing).  And when it collapses there will be fireworks.

[dubsartur]

The board of OpenAI just removed its CEO, a random billionaire.  The chairman of the board the quit in protest.  The board consisted of six people: Greg Brockman (chairman and president), Ilya Sutskever (chief scientist), Sam Altman (CEO), Adam D’Angelo, Tasha McCauley, and Helen Toner.  Of them, McCauley and Toner are involved in rationalism or effective altruism, and Sutskever throws around the term 'AI safety' which can be a LessWrong term of art for their specific dreams of how machine intelligence could go bad. 

Microsoft bought 49% of OpenAI for USD 13 billion.  So Effective Altruists and LessWrong rationalists made up about half of the board of a USD 26 billion company in a sector which could have very big consequences (OpenAI's charter presumes that "artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work" is coming in the next few decades).

NYT - Wired - Castor and Gerard (with G's trademark personal attacks, gossip, and statements that don't match my observations of these movements online)

[Jubal]

Yeah, I think a lot of how much this particular tiff matters is quite dependent on how much you believe maximalist claims about AI's capabilities. From what I'm reading it sounds like there seems to be a bit of a struggle between the people who want to run OpenAI as a Rationalist/EA nonprofit (possibly aligned to Sutskever) and those who want to run it more like a standard business (possibly aligned to Altman).

At the moment I'm definitely seeing LLMs and generative models becoming increasingly in-use by people in general and colleagues in neighbouring fields, but most of what I'm hearing about in terms of job losses is in things like content writing, and I think that may be an outcome of the SEO-isation of content anyway, something I wonder if there'll be a backlash too as the current generation of internet systems end up creaking under the strain of autogenerated garbage.

I see another AI lead has resigned over the exploitative nature of using copyrighted materials in training sets.

[dubsartur]

Yeah, I think a lot of how much this particular tiff matters is quite dependent on how much you believe maximalist claims about AI's capabilities. From what I'm reading it sounds like there seems to be a bit of a struggle between the people who want to run OpenAI as a Rationalist/EA nonprofit (possibly aligned to Sutskever) and those who want to run it more like a standard business (possibly aligned to Altman).

At the moment I'm definitely seeing LLMs and generative models becoming increasingly in-use by people in general and colleagues in neighbouring fields, but most of what I'm hearing about in terms of job losses is in things like content writing, and I think that may be an outcome of the SEO-isation of content anyway, something I wonder if there'll be a backlash too as the current generation of internet systems end up creaking under the strain of autogenerated garbage.

I see another AI lead has resigned over the exploitative nature of using copyrighted materials in training sets.

I am interested because it shows that Rationalists and Effective Altruists now control significant money and power (eg. Clarkesworld magazine had to close submissions due to a deluge of LLM spam which overwhelmed their slush readers, and its not clear that essay assignments have a future).

I don't think the Internet needs my speculations about politics between people I heard of today or the future of spicy autocomplete.

The problem with copyright is that copyright is a racket too.  I have never seen a justification for copyright longer than 20 or 30 years after the creator' death, its just an excuse for corporations to collect rent.  But its concerning that these programs often spit out their original training data and that some are designed to take all the money and recognition out of things that some people really love to do.

[Jubal]

The problem with copyright is that copyright is a racket too.  I have never seen a justification for copyright longer than 20 or 30 years after the creator' death, its just an excuse for corporations to collect rent.  But its concerning that these programs often spit out their original training data and that some are designed to take all the money and recognition out of things that some people really love to do.

Agreed on the general awfulness of copyrighted, though I think a lot of the biggest negatives of AI are avoided if it can't train on reasonably copyrighted (author still alive or very recently deceased) content, because an AI where most of its data set has to be ca fifty years old is not exactly going to be producing the kinds of outputs where it can purport plausibly enough for nonspecialists to be an up to date artist, journalist, or scientist. I think I'd be finding AI a lot more fun if what it could mostly produce was rehashed variants of 19th century book illustrations or even 1940s era comics, the kid of stuff that is or should be reasonably in the public domain anyway and where there isn't an issue of people basically providing unpaid labour in vast quantity to the AI developers.

[dubsartur]

This essay is billed as a criticism of Effective Altruism but is really a criticism of Utilitarianism. https://www.theintrinsicperspective.com/p/why-i-am-not-an-effective-altruist

I still say that if you look at things like We Charity, "raising awareness," or scams run by churches and gurus, you have to ask "couldn't we be doing this better? shouldn't we ask to verify the results? we say we want to reduce suffering but do our actions help to achieve that?"  And in actual emergencies, nobody starts raising deep philosophical questions about the ethics of triage, they just start allotting their resources with the desperate efficiency of Homer's widow spinning someone else's wool to feed her children.

Likewise, every day we make decisions which will create suffering somewhere far enough away that we don't have to think about it.  Every time our local council votes on bus lanes vs. bike lanes vs. automobile lanes they are making a choice like that, because commutes take time and people die of car crashes and air pollution and not being able to get to a hospital quick enough.  Isn't it better to explicitly consider the tradeoffs than to let some rich guy decide on a whim?

Edit: the essayist wants you to know that he was a bold cryptocurrency trader (ie. gambler or fraudster).  Ia ia Cthulhu ftagn.

Edit: another essay in an internet magazine aimed at Americans with university education and above-average incomes https://www.currentaffairs.org/2023/05/why-effective-altruism-and-longtermism-are-toxic-ideologies

[dubsartur]

'Cecil Adam' of the Straight Dope column (1973-2018) returned in 2023 with a column on longtermism https://boards.straightdope.com/t/straight-dope-1-13-2023-is-longtermism-the-worlds-most-dangerous-belief-system/978173  Apparently some people believe 'Cecil' is a team of writers or a series of writers.

Edit: short life of Napoleon Hill, the early-20th-century swindler and self-help author (specialty: how to get rich quick).  Many aspects of his career overlap with dubious people in the spaces in these threads. https://gizmodo.com/the-untold-story-of-napoleon-hill-the-greatest-self-he-1789385645

The Royal Fraternity of the Master Metaphysicians was founded by James B. Schafer and is largely forgotten today. Born around 1896, Schafer came from Michigan to New York sometime around 1930 and by the mid-30s had amassed a following through his speeches on the spiritual potential hidden in the material world. He explained to crowds of hundreds at Carnegie Hall each Sunday morning that the human mind had the ability to change everything around it. If you could simply imagine it, those thoughts could become real. By some estimates Schafer counted nearly 10,000 people amongst his followers by the end of the decade. ...
Schafer’s intentions with the cult were unclear. He seemed to believe every word he breathed, but he also saw that his status afforded him access to a great deal of money and women. There’s a strange psycho-sexual component to the Master Metaphysicians that’s always hinted at in news articles of the day, but never said outright.

[dubsartur]

An animal-rights activist has alleged that she had an affair with Peter Singer, the utilitarian philosopher with many crappy opinions, and that he had a series of physical relationships with female co-authors.  This is obviously a deeply personal conflict (the plaintiff sued about events in 2002-2004 in 2022 and represents herself) but it adds to the suspicion that many influential men in this space are in it for the chicks, and that some of the public controversies are shaped by private interpersonal drama.

Singer was one of the early advocates for Effective Altruism, although he seems to have aged out of the spotlight by 2022 (as a famous tenured professor at a rich private university he does not need the money or attention).

Teacher-student relationships (not just mentor-mentee relationships) seem very common between tenured faculty and students at wealthy US universities.

Docket 22CV01792 at https://portal.sbcourts.org/CASBCIVILPORTAL/  Many content warnings (cosmetic surgery, messy breakup with someone who likes to share crappy opinions)