November 2023 BenGoldhaber.com Newsletter
AI drama, AI investments, and AI meditation. Plus, twitter drama.
This ‘November-ish’ edition of the BenGoldhaber.com newsletter is brought to you totally on time by our sponsors, the common cold and holiday travel.
#links
How to Hopefully Ethically Make Money off of AGI: I enjoyed this LessWrong dialogue for its concrete recommendations and advice on the topic of investing for AGI.
I'm open to buying individual stocks, but it's very difficult for me to recommend specific stocks to other people under extreme uncertainty! How can an average investor with <15 hours of research as Oliver stipulated know they're doing anything with positive expected value over buying a broad index? Anything that's a mega-winner will in fact get captured by a broad index fund.
I do think it's fine for younger more risk seeking people with time on their hands to give this some thought and pick individual stocks. If nothing else you will learn a lot by doing this, both about markets and about your psychology, and it can incentivize you to learn more about the companies in question once you have skin in the game. That said, I would never let your index fund exposure go to 0 in favor of individual stocks, just to avoid the risk of ruin that you captured none of the eventual winners.
I generally endorse an efficient market hypothesis for the stock market - I don’t expect to be able to beat the combined brainpower of millions of people who are all incentivized to find an edge before I do. But, events that are outside of the training distribution of the economy - your Covids, your Cryptos - do seem to be cases where individuals with a strong inside view can outperform broad index funds.
My general approach in such situations is that you want to look for investments/gambles that would not be too expensive (in EV terms) if your thesis was chosen at random, or even would be good ideas anyway, but that would then benefit greatly if your thesis turns out to be true.
If you’re into kremlin-ology of AI Labs it sure was an exciting month; the best summary of the OpenAI drama was Matt Levine’s MoneyStuff writeup. I’m not going to dive too deep into it, because there has been so much coverage of it from all corners already (see Zvi). Also I’m still confused about what happened and what it means. But a few general notes on it:
I have a lot of respect for the board members and I hate the initial coverage that tried to summarize it as incompetence - still, I don’t understand why no public explanation from the board was ever disclosed, and given the magnitude of the action, one was needed.
I feel like, in corporate politics and politics politics, coalitions always lose to the single, strong, live actor.
AI Safety took a medium size L, as the SV VC class seems to have been radicalized against it.
If it’s true that they torched an $80 billion company cause they thought they were too close to building God, then that’s orders of magnitude the most punk rock thing I’ve ever heard of. - Joe Weisenthal
Shallow Review of Live Agendas in Alignment: A broad overview of the actively pursued research agendas in AI Alignment. Doing well researched ‘roadmaps’ like these is a public service. Very comprehensive.1
Defensive Acceleration, d/acc, is Vitalik’s nuanced call for the right kind of technological progress - aka mindful of risks from advances in AI and BioTech, but still excited about shooting rockets into space and living forever.
The Chinese proverb 天高皇帝远 ("tian gao huang di yuan"), "the sky is high, the emperor is far away", encapsulates a basic fact about the limits of centralization in politics. Even in a nominally large and despotic empire - in fact, especially if the despotic empire is large, there are practical limits to the leadership's reach and attention, the leadership's need to delegate to local agents to enforce its will dilutes its ability to enforce its intentions, and so there are always places where a certain degree of practical freedom reigns.
With AI, no longer. In the twentieth century, modern transportation technology made limitations of distance a much weaker constraint on centralized power than before; the great totalitarian empires of the 1940s were in part a result. In the twenty first, scalable information gathering and automation may mean that attention will no longer be a constraint either.
d/acc stands in contrast to the e/acc ‘movement’ that has been picking up steam on tech twitter, which celebrates the no holds barred, acceleration of technology. I put movement in air quotes because it’s probably wrong to think of it as an ideology, and more accurate to call it an aesthetic stance. Move very fast and break everything.
In this sense it is pretty punk, and the essay Punk as Ideology is a good read for how to think about a community and vibe with a set of underspecified beliefs.
It is a style, it became an ideology, and it remains a community. All in all, a good way to spend your teens.
I must admit to having a bit of contempt for the VC class who are adding e/acc to their twitter handles. It’s one thing if you’re an anonymous account on twitter posting about how cool Nick Land is; it’s another if you’re a centi-millionaire living in the Palo Alto suburbs to sign up for a brand that includes “nothing human makes it out of the near-future”.2 As I said last year, I remain opposed to destroying our social institutions.
One more bit of prognostication about cultural trends in tech - the meditator / techie singularity is nigh. I agree with this tweet from Sarah Constantin/Nick Cammarata
There's an AI/meditation cluster of people who are well funded and are *highly* interested in neuroenhancement tech, including transcranial ultrasound, to become "awakened" (whatever that means)
I recently met the creator of Jhourney, a biofeedback device for practicing blissful meditation, and I came away impressed with their rigorous and reasonable approach; biofeedback might be able to train people to enter specific mental states significantly faster and easier, and the meditation community is a great testbed.
If non invasive neurotech can decrease the time it takes to enter an advanced meditation state, I think you’re on the path to the dream of the hippies of the 60’s to spark a ‘consciousness’ revolution. That vision seems silly decades later - jaded children of the 2000s that we are - but finding new ways to relate to one another through the magic of LSD must have seemed like a worthwhile political cause.
Finetuned neurotech, with fewer chemical downsides, might enable similar experimentation in new social political realities. On that note I suspect we underestimate the degree to which our current social reality is shaped by relatively recent biotech interventions - some estimates:
1% to 3% of American men are on testosterone replacement.
3% to 4% of individuals in the US have a prescription for stimulants in 2020.
11% of women aged 15-49 use hormonal birth control.3
All of these can have meaningful changes on individual behavior and it must add up to macro level, cultural effects. Maybe tech enhanced meditation will similarly change individuals and, in turn, the culture. Related: Sasha Chapin on Meditation and Letting Go
Loyal: A life extension drug for large breed dogs received a significant seal endorsement from the FDA and looks like it will become available to consumers. This is important both because I love dogs and want them to live a long time, and because it’s a proof of concept of getting approvals for general life extension drugs for humans.
So far, the FDA has not approved any drugs to expand the lifespan of animals—or humans, for that matter. “This is completely novel,” says Linda Rhodes, former CEO of pet biotech company Aratana Therapeutics and a consultant for Loyal. It’s difficult to study life-extension drugs in people, she says, because humans live relatively longer lives than other species. But starting with dogs—and the breeds with the shortest lives—could yield important clues. “The implication for other species, including humans, is pretty profound,” she says.
Do ten times as much: Perhaps we dramatically underestimate the amount of effort necessary to git gud. Brian Caplan argues you should 10x your expected investment to achieve your goals, and I find the combo of either do a lot more or don’t do it all to be the right rule of thumb. Related: Make an Extraordinary Effort.
Ten times as much of what, exactly? The answer is usually: Whatever you already think the crucial ingredient is. “Why can’t I get ahead in my career? I strive to study and emulate my role models.” Great idea; you just need to multiply your effort by a factor a ten. “How can I save my marriage? I’m really trying to make my spouse happy.” Again, great idea. You just need to multiply your effort by a factor of ten.
#good-content
Best Served Cold: My favorite book universe remains that of Joe Abercrombie’s The First Law, and this month I reread Best Served Cold, a Tarantino meets George R.R. Martin revenge thriller. Maybe I’ve reread it six times now? Still so good, and it works as a standalone book so it’s easy for you to dive straight into.
“Toughest thing to manage, a retreat. Maybe he just hasn’t found his moment yet.” Shivers gave a fareway sigh. “We’re all of us waiting for our moment.”
Breaking History: I read Jared Kushner’s autobiography and I must say - pretty good? It’s hard to know what’s true, what’s false in these self-reports, but I found it to be a thoughtful look at power politics in DC and the fakeness of the Overton window (it was in fact possible to achieve far more in Middle East diplomacy than the conventional wisdom anticipated). It also, of course, has delightful anecdotes about Trump:
“That’s great,” he said. “Most people think I’m Jewish anyway. Most of my friends are Jewish. I have all these awards from the synagogues. They love me in Israel.” Then he added, “I just hope you’re serious because Ivanka is in an amazing place in her life right now. You know, Tom Brady is a good friend of mine and had been trying to take Ivanka out…”
Before he got any further, I quipped, “If I were Ivanka, I’d go with Tom Brady. “ He looked at me with complete seriousness. “Yeah, I know,” he sighed.
Legendary stuff. Lets close out the year strong,
xoxo,
Ben
yours truly gets a name check in the section on FAR AI, let it be known I’ll almost always include blog posts that have my name in it (in a positive context only)
Twitter moves pretty fast and it seems lie e/acc might already be cringe. Live by the vibe die by the vibe.
These estimates are sourced from bard.google.com, so take with a grain of salt.