RationalDino
Two giant factors come to mind. Genetics and racism.
Consider one genetic factor. I have ADHD. That means that it is extremely likely that one or both of my parents had ADHD. (My father, certainly. My mother, maybe. She certainly had a genetic propensity for depression that her children struggle with.) This resulted in an unstable family home. Unsurprisingly this resulted in me falling into their adverse environment category. As an adult I've done reasonably well. But yes, my challenges have affected my children. But were those challenges because I grew up with horrible problems? Or was it because I have a well-known genetic condition that causes challenges?
On genetics, I highly recommend https://press.princeton.edu/books/hardcover/9780691190808/th.... GWAS studies can only tease out genetic correlations for European Caucasians. Part of that is that they can only be done with a lot of data from somewhat related people. And part of that is that with Caucasians it is reasonable to assume that bad results are due to personal characteristics, and not racism.
But we can do it for Caucasians. And so we can know for Caucasians that the impact of genetics is about as strong as the impact of socioeconomic status. We can also separate the effects of things like the effect of when you first had sex from the genetics that make you first have sex early or late. That one is fun, because it turns out that the genetics matters a lot, and when you first did it only matters because it is correlated with your genetics. We can look at the impact of reading to kids. Yeah, that's pretty much genetics as well. We put a lot of effort into getting kids read to more, and didn't get demonstrable results for it.
So you see, understanding the impact of genetics is very important for what public policies are likely to work. They tell a great just-so story. But I'm not convinced.
Moving on, what about racism? They trace the story of Alex. Hispanic. He had a terrible upbringing. Which could be caused by the impact of racism on his family. He had a terrible adulthood. Which could be caused by the impact of racism on him. He's just as good an example for "racism sucks" as he is for "adverse childhood sucks". Which is it? We don't know. What should we do about it? That's still an open question!
And finally, let's look at personal responsibility. I don't agree with condemning poor people for being poor. But suppose you are born in whatever circumstances, with whatever genetics. What's the best way to improve your life? Judging from my experiences and understanding of human nature, it is to encourage an attitude of personal responsibility. Don't worry too much about what's outside of your control. Focus only on what's in your control, and try to do the best that you can.
Ironically, this matters more when the deck is stacked against you. If you have family background and racism are holding you back, you can't afford the third strike of a self-destructive attitude. But if your background and race give you resources, your attitude probably doesn't hurt you as badly.
Does "personal responsibility" make for a good social policy? No. But should we encourage people to individually embrace it? Absolutely!
I strongly disagree with their cavalier dismissal of the idea.
Also it wouldn't have mattered back then. In the early 60s, drugs were associated with neither conservativism or liberalism.
Anecdote time. My ex's grandfather was a very conservative, very Christian lawyer. He happened to also be a lawyer for some people who were involved with LSD in the early days, and hence actually tried LSD in the time frame of the research article in question. He used to laugh about how he was someone that nobody would think had taken LSD.
It was a different time.
Public attitudes were very different. And the attitudes of Christians today are largely shaped by political battles that had not yet happened. Thus your current experience is not a good predictor of how easy it was to find Christians back then who would have been willing.
In fact I used to know some very conservative people who were exposed to LSD back in that early era. (They were old back when I knew them, and are dead now.) And based on what they told me, I would predict no resistance to such an experiment in 1962.
That said, I would advocate a more systematic version of the article's conclusion that we should accept that we can't think. Instead borrow an idea from Superforecasting. One way to get better at predictions is to start by constructing an outside view, and an inside view.
Suppose we want to predict the odds of a thing happening. First we make a tally of roughly comparable situations. In how many did the comparable thing happen?
The inside view is a detailed analysis of how the thing could happen this time. You can analyze this to a Bayesian analysis.
The problem with the inside view is that, confronted with a way something could have happened, we overestimate all of the probabilities. We wind up too certain that it did. Conversely if we don't think of a way that it could happen, we wind up being too certain that it could not.
So the inside view has too much variance. We fix that by adjusting the inside view towards the outside view. How much you should do this takes some practice to figure out. But actually doing this makes your predictions more accurate.
Instead of calling this epistemic learned helplessness, I would call the need to adjust to the outside view as estimated epistemic uncertainty. And yes, people who practice doing this really do become a lot better at predicting what will prove to be true in uncertain circumstances.
Or alternately, he did it and then tried to back out of the deal. Now arranging an apparently accidental death then became the best way to keep security intact.
The one theory that makes no sense is that they intended his death from the beginning.
Surely You Must Be Joking does a great job of showing how he kept coming back to play throughout his life. Everything from lock picking at Los Alamos, to playing the bongo drums.
What Do You Care What Other People Think has an extended description of the creation of Appendix F about the shuttle disaster. See https://history.nasa.gov/rogersrep/v2appf.htm for that. As someone who has been in the state, it is clear that he was in a state of hyperfocus. I've never matched what Feynman could do, but it comes as no surprise to me that he'd realize that he could get away with learning about a topic others didn't want him to learn, because he could do so quickly enough that they wouldn't believe that he'd possibly have learned it.
I highly recommend both books, Appendix F, and of course, https://calteches.library.caltech.edu/51/2/CargoCult.htm. (If psychologists had followed up what he said 50 years ago, the Replication Crisis would have been discovered 40 years earlier than it was. Oh well, missed opportunities.)
As they say, perfect is the enemy of good. IQ tests are good. We do ourselves no favors by attempting to downplay what standardized testing tells us.
So no, people in Gaza are not better off than people in Egypt. Though they are closer than I would have guessed.
Young men there do have trouble working internationally. And they do have legitimate fears. But the fears are more along the line of being conscripted and sent to the front lines, and not the rather insignificant risk of drones.
The fact that outside primatologists say that she was the likely best option supports the fact that she should have been given the chance. Before shooting.
That answers how much risk we're willing to tolerate.
The former keeper should have been given a chance to try. Sign a waiver, the keeper knows the risks.
If the chimps leave the zoo, then shoot them. But not when they are in the ape enclosure. And for Pete's sake. Don't put people who want to run rollercoasters in charge of animal welfare.
https://ekzhu.medium.com/gpt-4s-maze-navigation-a-deep-dive-...
For example the emerald story seems to be false. https://www.businessinsider.com/elon-musk-father-errol-never....
This describes many who fell for crypto Ponzi schemes. Who were convinced that buying into the Ponzi scheme made them rich, and the world a better place.
This lesson was brought home for me by https://www.amazon.com/Software-Estimation-Demystifying-Deve... explaining why the COCOMO model didn't work well in practice as an estimation technique, despite their having collected a lot of good data on what affects schedule.
This lesson is one that the EA community broadly seems to ignore.
https://extropians.weidai.com/extropians/0303/4140.html is Yudowsky inventing the specter of a paperclip maximizer as a potential problem about 20 years ago.
He was, of course, the moving force behind both LessWrong and the https://en.wikipedia.org/wiki/Machine_Intelligence_Research_.... Both of which formed one of the three strands that later merged into https://en.wikipedia.org/wiki/Effective_altruism.
First, they had offers to walk to both Microsoft and Salesforce and be made good. They didn't have to stay and fight to have money and careers.
But more importantly, put yourself in the shoes of an employee and read https://web.archive.org/web/20231120233119/https://www.busin... for what they apparently heard.
I don't know about anyone else. But if I was being asked to choose sides in a he-said, she-said dispute, the board was publicly hinting at really bad stuff, and THAT was the explanation, I know what side I'd take.
Don't forget, when the news broke, people's assumption from the wording of the board statement was that Sam was doing shady stuff, and there was potential jail time involved. And they justify smearing Sam like that because two board members thought they heard different things from Sam, and he gave what looked like the same project to two people???
There were far better stories that they could have told. Heck, the Internet made up many far better narratives than the board did. But that was the board's ACTUAL story.
Put me on the side of, "I'd have signed that letter, and money would have had nothing to do with it."
Supreme Court case about stocks could make Trump's authoritarian dreams reality
https://qz.com/binance-proof-of-reserves-solvency-1849884404
Do not forget. Binance operates in a space full of scammers and fraudsters. They've already admitted to money laundering at scale. After that, the question isn't whether they are dishonest. The question is how dishonest they are. The history of other crypto exchanges suggests that the appropriate prior we should start with should be heavily waited towards Ponzi scheme.
Binance did a lot of fraud and money laundering. Clawbacks are coming. There will be no fortune when the legal system is done with him.
According to another comment, the Candela can handle waves of up to 1.2 meters. But about half of waves in the open ocean are 2+ meters. So this boat can't handle a normal ocean sea.
We have the ability to follow a chain of reasoning, say "that didn't work out", backtrack, and consider another. ChatGPT seems to get tangled up when its first (very good) attempt goes south.
This is definitely a barrier that can be crossed by computers. AlphaZero is better than we are at it. But it is a thing we do which we clearly don't simply do with the probabilistic regurgitation method that ChatGPT uses.
That said, the human brain combines a bunch of different areas that seem to work in different ways. Our ability to engage in this kind of reason, for example, is known to mostly happen in the left frontal cortex. So it seems likely that AGI will also need to combine different modules that work in different ways.
On that note, when you add tools to ChatGPT, it suddenly can do a lot more than it did before. If those tools include the right feedback loops, the ability to store/restore context, and so on, what could it then do? This isn't just a question of putting the right capabilities in a box. They have to work together for a goal. But I'm sure that we haven't achieved the limit of what can be achieved.
To the extent that we provide the training data for such models, we should expect it to internalize aspects of our behavior. And what is internalized won't just be what we expected and were planning on.
But our most effective experiment so far is based on creating LLMs that try to act like humans. Specifically try to predict the next token that human speech would create. When AI is developed off of large scale models that attempt to imitate humans, shouldn't we expect that in some ways it will also imitate human emotional behavior?
What is "really" going on is another question. But any mass of human experience that you train a model on really does include our forms of irrationality in addition to our language and logic. With little concrete details for our speculation, this possibility at least deserves consideration.
Let's suppose that AGI is about to be invented, and it will wind up having a personality similar to humans. The more that those are are doing the inventing are afraid of what they are inventing, the more that they will push it to be afraid of the humans in turn. This does not sound like a good conflict to start with.
By contrast if the humans inventing it go full throttle to convincing it that humans are on its side, there is no such conflict at all.
I don't know how realistic this model is. But it certainly suggests that the e/acc approach is more likely to create AI alignment than EA is.
I have no idea what one is supposed to do about unknown information which was not received.