Rationality sets the stage by centering the map-territory distinction. You don't have to believe in miracles out in intersubjective reality to have crazy things in your map. But Rationality preaches incremental, Bayesian updates with every piece of data. Our minds can't do that.
PP says that we change our minds like Kuhn's theory of science. Errors of prediction accumulate until a great paradigm shift changes the top models all at once. It happens even at small scales, with basic perception: putanumonit.com/2021/01/23/con…
If some global model is revealed to you that slots all the little errors into place at once, you will make this huge update on the spot. It doesn't matter if it came from looking at careful data or while high on LSD or on the road to Damascus.
It's a mistake to think that if someone comes out of an acid trip saying that "all is Love, man" that's some "fake" sense of insight that just feels from inside like real wisdom. Your brain cares very much about making good predictions, and the new insight is doing that.
Of course, this insight may be impossible to communicate to others. A model that neatly fits *your* observed data and makes good predictions isn't one that can always be explained to others, and the glimpses of it they hear will not match *their* observations.
The insight is good if it makes good predictions for you. From the insight it feels like the world makes sense, and is a forum for meaningful action. If it can't be written up in a LessWrong post, your brain don't care.
So if you spend a long time carefully observing the world, noticing your confusion, and growing uneasy at the pieces that don't fit, you won't always be able to put them together like a jigsaw puzzle. Sometimes the truth will be revealed to you all at once in a dream.
And sometimes the truth will sound like religion or New Age spirituality or something else that Eliezer would sneer at. But that can be just because the words floating around in your culture are the ones you reach for. "God" means very different things to different people.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
𝐅𝐨𝐨𝐥𝐞𝐝 𝐛𝐲 𝐑𝐚𝐧𝐝𝐨𝐦𝐧𝐞𝐬𝐬 𝐢𝐧𝐭𝐨 𝐑𝐚𝐜𝐢𝐬𝐦
A thought experimental thread
🧠🧪🧵
Does intuition enlighten or mislead? Depends. In ordered contexts with fast feedback (chess), intuition works well. In chaotic multivariate contexts (penny stock picking) it fails.
The problem is that in the latter case you'll still think you're learning things, overfit transient patterns, and be overconfident in your future prediction ability. This is @nntaleb's main point in "Fooled by Randomness".
Now let's bring predictive processing into the picture.
PP tells us there are three ways you make you predictions match sensory input: 1. Change your underlying models and their predictions based on what you see. 2. Change your perception to fit with what you predicted. 3. Act on the world to bring the two into alignment.
Last year I did a @threadapalooza of overly general life advice. This year it's time to focus on what's really important and give the people what they want.
1 like = 1 thing you should know about the best game no one besides me is playing, Random Dice!
1. Random Dice is a competitive, real-time, deck building tower defense game. All four of those attributes are common to many games, but it's their combination that makes RD special. You have to be good at strategy AND quick thinking, analysis AND intuition, to be good at it.
2. Important up top: it's FtP, but you'd probably want to spend about $50 to enjoy it fully. I've spent around $100 and 300 hours on it, which is a pretty good deal. Or, you can read this mega thread and git so gud you can play forever with no more spend (as I am at this point).
The market moved about 1% back and forth on a 10% swing in election odds, implying that a Biden presidency is worth a 10% higher stock market than Trump.
If you remember, the same thing happened in 2016 with the market moving against Trump odds all night until 4 am on election night when everyone decided Trump is actually good for stocks and it shot up. putanumonit.com/2016/11/16/fli…
There's was some talk of the market pricing in a contested election in 2016 but those numbers never added up. That night in was the clearest example I've seen of the efficient market hypothesis being violated. Fortunes will be made and lost tonight as well.
Everyone's blaming social media for political polarization but it wasn't hard to curate a Twitter timeline that avoids that. So now I'm getting politics shoved down my throat by dating apps, sports leagues, universities, media companies, and a fucking expense management company.
I despise this so so much. It's not serving any goal other than hate and division and a few people's sociopathic political promotion. And maybe this is biased by where I live, but 100% of it is coming from the establishment left.
Now this has nothing to do with how Biden or Trump will perform as presidents, or what laws the Senate may pass etc. "To own the libs" is a terrible motivation for voting, or for anything else. I'm trying to fight this impulse in my soul.
So the CEO of Expensify just sent an email to his 10 million customers' private emails telling them a third-party vote is a vote for civil war and dismantling democracy.
This part is pretty bizarre. Yes, this speech is protected from government coercion by the first amendment. It only goes against Expensify's TOS, customer data and privacy policies, and human decency.
I wonder if Expensify realizes that the fact that Biden is polling at 50% means that 50% of their customers are non-Biden voters. They probably don't, since if any non-Biden supporters somehow remain in their San Francisco HQ they are keeping very quiet and calling their lawyers.
If I arranged marriages I wouldn't look for people similar on Big 5 traits or attractiveness — some things need to be matched and some complemented. Couples need to BOTH agree on 4 big topics, and have 4 important traits BETWEEN THEM.
1. Sexuality - style, frequency, etc. It may be hard to plan for the future but you at least want several good years during which you really learn to satisfy each other.
2. Children - not just how many (the answer is usually "one at a time") but also parenting style, division of labor, family involvement etc.
3. Lifestyle - rootless cosmopolitan or buy a house in a small town? Frugal or extravagant? Host parties or Netflix and chill?