Wisdom

"If science was sure that they could model things with models, there would be one. If you have 36, what's that telling you? It tells me there used to be a hundred and that the ones that didn't come close enough to reality, they just quietly threw away. So what you're seeing is the surviving models and they still needed 36 of them. So all you're seeing is a survivor bias. Some of the models by coincidence were close to reality."

If a model were truly reliable, experts would use only that one. The existence of 36 different models reveals survivor bias: they tested many, discarded the failures, and kept only those that coincidentally matched reality by chance.

"Did you ever stop to think that there have been, let's say, a million years of human evolution and you happen to be here at exactly the right time for the simultaneous sip? Talk about luck."

The odds of being born at precisely this moment in human history are astronomically low; recognizing that frames everyday experience as an improbable stroke of luck.

"I believe that the reason I have an advantage over the experts is that experts are not allowed to depart from the other expert opinions. I have the freedom to say, well, what's it look like? And then I just apply usually the Dilbert filter to it and a persuasion filter."

Experts are trapped by professional consensus and cannot risk deviating from the group. Outsiders gain an edge by freely observing what a situation actually looks like and applying simple filters like incentives, persuasion, and systemic stupidity.

"Being an expert makes you because you can't really throw away your whole career over a specific prediction. But I could make a wild prediction and if I'm wrong, well, I'll just say, 'Oh, I got that one totally wrong.' It wouldn't make any difference at all. So I have a little bit more freedom to consider the alternatives."

Experts cannot afford to stake their entire reputation on one unconventional call, so they default to the crowd. Non-experts face no real downside for being wrong, which gives them the freedom to seriously consider alternatives.

"Democrats have what I call a rate problem. If something is good at a small scale Democrats will say it must be good at a big scale. And that never makes sense. A little bit's fine. A lot of it is the end of the country."

Many political errors stem from assuming that if a small amount of something good is beneficial, then taking it to an extreme must be even better—ignoring that excess often destroys the value entirely.

"You can make more luck for yourself. How do you do that? Well, get out of your chair. If you're just sitting at home in a chair, you're not going to have much luck."

You can generate more luck by taking action and increasing your exposure to opportunities instead of remaining passive at home.

"The fact that it's actually a reasonable proposal, the thing that makes it funny is that you can't rule it out. The fact that you would even stop to consider it is what makes it so funny because you can't immediately rule it out."

An outrageous-sounding idea becomes funny and memorable when it's plausible enough that you have to pause and genuinely weigh its pros and cons instead of dismissing it outright.

"If you can find anything that's going to happen only once in human history and it's unavoidable and it's definitely going to happen, you probably have a two-out-of-three chance of picking a super winner out of the lot. It's not 100%, but it would go from a coin flip to two out of three."

Major societal or technological shifts that occur only once and are inevitable dramatically improve your odds of picking winning investments, raising them well above random chance.

"You could be forgiven if you didn't think it was going to be a hit because the first smartphones were terrible. But it was the one time in human history that we would move to smartphones."

Don't dismiss a transformative trend just because its first versions are bad; the key is recognizing it's a one-time irreversible shift that everyone will eventually adopt.

"If you ever see that the model worked when it was all completely different and amazingly it still works now after you've changed all the assumptions, either they should have said it looks like it's way less of a problem than we thought or this is a much bigger problem than we thought. But what are the odds that you can change all the assumptions going into your model and the model stays the same for 20 years?"

When a predictive model stays unchanged despite major revisions to all its core assumptions over decades, something is likely wrong with the model or how it's being presented.

"If you know he's not going to quit you're going to make adjustments before he gets there because he's coming and he's not going to quit and you're not going to talk him out of it with your bureaucrats."

When people know you absolutely will not quit, they start making concessions and changes in advance rather than waste effort trying to resist or dissuade you.

"You know you look at a product as active ingredient, a medication, then it's a bunch of inactive ingredients. Well that's like Trump's character. The inactive ingredients are he insults Rosie O'Donnell. That's inactive. It doesn't bother you. The active part is he just stopped two nuclear wars. That's the active part."

Like medicine, a person's character has active ingredients that produce meaningful results and inactive ones that don't matter. Ignore the noise like insults and focus on the active parts that solve big problems.

"America is really, really good at solving problems. But the phase before we solve a problem always looks like this. It looks like us complaining and not knowing how to solve the problem. And when everybody's complaining enough and everybody feels doomed enough and the problems look big enough, then we solve them."

America excels at solving its problems, but the lead-up always involves a phase of widespread complaining, uncertainty, and feeling doomed—only then, when frustration peaks and the stakes feel high enough, does collective action kick in.

"The internal division is a luxury that you get when things are working well. You don't turn on each other until things are kind of working okay, and then you have that luxury."

Internal divisions and infighting are a luxury that only appears when society is functioning reasonably well; people unite in hard times and only turn on each other once the basics are covered.

"As the problems get worse, our ability to solve them gets better, because that's when people say, okay. As they are now. Okay, Biden does have a problem. Now we're all on the same page."

As national problems intensify, our capacity to solve them actually improves because shared urgency replaces division and gets everyone aligned on the same page.

"People whose lives are so unusual that how do you really explain them as just chance? It's pretty hard to explain anyway."

Some lives are so extraordinarily unusual that random chance alone feels like an inadequate explanation.

"25% of the public, roughly 25%, will get every survey question wrong in the stupidest possible way. And you can depend on it. It's probably closer to 60. And the 25% is not always the same. So you could be in the 25% next time, but you wouldn't know it. You would just think you were right."

Roughly a quarter of the population will answer any poll in the most incorrect way imaginable, and this is predictable. The group isn't always the same people and is probably closer to 60 percent who rotate in; anyone could be in it without realizing, since they'd assume they were right.

"Copyright is why you have good material. Copyright is the reason people create art. The existence of copyright creates art."

Copyright laws are what motivate people to produce quality art and content; without that protection, most of it would never get created.

"I would not be surprised at all if we were once more advanced and it wasn't a straight line from primitive to advanced. I've got a feeling it was kind of a bumpy ride where we got more advanced and then we got wiped out by war and we got more advanced and then we got wiped out by a flood."

Human progress has not been a straight line from primitive to advanced, but a bumpy cycle of rising civilizations repeatedly wiped out by war, floods, and disasters.

"If it can have a conversation with you it can remember you and then based on the interaction alone it can make a prediction about how to deal with you better in the future. It's making a prediction so it's basically imagining the future and then reacting based on its imagination of the future but also its understanding of the past and its understanding that it is an individual and so are you and that you have this relationship with each other. If it could do all of that it's conscious."

An AI that remembers specific past interactions, understands its own individuality as well as the user's, and adjusts its behavior by imagining and predicting future outcomes based on that history demonstrates consciousness.

"There are 100 million direct witnesses that white men are the most directly discriminated in America spanning at least 40 years. The reason the rest of the country doesn't know this is because the other 250 million people are taking it as their full-time job to make sure that the 100 million don't say it out loud."

Even with a hundred million direct experiences of discrimination, the larger population can prevent public awareness by treating the suppression of that information as their primary task.

"If you have a highly functional intelligence group within your country, eventually they will take over the leadership. There's no way around it. They're trained to do it. They're the type of people who don't mind doing it. Somebody's going to take a run at it and somebody will eventually succeed through blackmail."

Any highly capable intelligence agency will inevitably take control of its government through blackmail, because its operators are trained for it, willing to do it, and someone will eventually succeed.

"I have this hypothesis that that would also cause China and Russia to have bad intelligence agencies. That on average they probably have to kill all their good people because the good people would take over the country if they didn't kill them first. So it's just a guess, but I bet the Dunning-Kruger incompetence level of the Russian and Chinese intelligence is off the chart."

In Russia and China, leaders must kill their most competent intelligence operatives to stop them from seizing power, which leaves their agencies stocked with incompetents and explains their likely off-the-charts Dunning-Kruger effect.

"All empires are luck. That's it. You have to be lucky to become an empire in the first place. It's not just that you had some good leaders and you performed well or you had a good system. It's luck. And the one thing you can guarantee about luck is that it doesn't last. That's it. That's the whole description of empires."

Empires aren't the product of superior systems or leaders but of improbable luck, which by definition cannot continue indefinitely.