46 Matching Annotations
  1. Jun 2022
    1. Context modeling. Literals are coded using an order 1 model plus the low bits of the current position. The position is useful for modeling binary files that have a repeating structure that is a power of 2, for example, an array of integers or floating point numbers

      🧐

  2. Dec 2021
  3. Aug 2021
    1. Germany was the first—one month, maybe two months before everything went bad, saying ‘Hey we’re gonna evacuate 300 interpreters and staff.’ Italy had more people, and they flew their people out. Then Britain came along and said ‘Hey we’re gonna move as many interpreters out as possible,“ said Miervaldis. “NATO allies started this process earlier and the question is, why hasn’t the United States jumped on that effort, or followed their lead

      no one could truly believe this is an accident

    1. candidates have to be prepared very early on to go on the offensive

      but being agressive is what they were attacking you for. The challenge is that this is a democratic primary and people are partisan

  4. Jul 2021
  5. May 2021
    1. Franklin Genomic Advancements

      too many big cap ones and big cap pharmacuticals have underperformed for a long time. I think that's because it is better to let the free market decide how much to allocate to each individaul project where each project is represented by a company. This is superior to a big hierarchical corporate bureaucracy deciding everything.

      Their largest holding is a chinese company and I'm bearish on China's pharmaceutical industry considering that their vaccine was not nearly as trusted as the american ones. https://www.nature.com/articles/d41586-021-01146-0. Trust matters. They also have a culture that discourages independent thinking and rather than attract the best and brightest from around the world they lose talent to america. They also don't have the same ability to enforce their IP protections because of their weak naval power. Seems like their pharmacutical industry could go the way of huawei.

  6. Apr 2021
    1. elite overproduction

      these "elites" are really just narcissists. Like people with bullshit degrees in stuff like art. If they were actually smart they would get a degree in something useful and have no issues getting whatever job they want.

    1. While some strategically minded thinkers, such as Andrew Bacevich, advocate redeploying U.S. assets in the Middle East to the Pacific, as opposed to the comprehensive retreat their more isolationist counterparts espouse, the general message remains the same: That there is no longer much value in securing geographically strategic points in the Middle East, and that U.S. security does not depend on it. This perspective is mistaken. Foreign Policy’s Dec. 13 article “RIP the Carter Doctrine” is correct that a stable Persian Gulf benefits the United States indirectly, by safeguarding a global economic and security interest in the steady supply of Middle East energy. In a parallel but opposite direction, an unstable Middle East benefits Russia’s interest in higher energy costs.

      so now that ISIS is defeated and people say this must mean we can go home now, they go and tell us what the real objective was

  7. Mar 2021
  8. Feb 2021
    1. "$TSLA and $BTC correlation coefficient is 0.951967 over the last six months," Burry said. "@elonmusk going for perfect unity? Nah, Elon dreams the impossible. He is determined to break unity. Correlation > 1. And he has history on his side. $TSLA and $BTC investors can make anything happen."

      2 factors have created a world where these bubbles are possible:

      1. Super exponential money printing
      2. Lack of engineers means there is a shortage of real investment opportunities. This shortage of investment opportunities is exacerbated now that the stock market seems pretty maxed out. There is also an increased demand for high return opportunities because the finance sector has gotten so big and so many of the smartest people are working in it now. All these highly paid people on wall street need high returns in order to justify their salaries and the increased size of the finance sector. This is a double whammy because many of those people now chasing bubbles in the finance sector are people who could have been great engineers

      This has created an incentive for people to invest in these bubbles and work tirelessly on pumping them up by spreading good news about them. As more people buy bitcoin more people will have incentive to pump it up.

      I would expect all of these factors to continue but what would make it stop at some point? 2 theories:

      1. maybe it won't but the bitcoin will depreciate relative to consumer goods because of rising prices. This is more so what I expect to see with the stock market
      2. At some point bitcoin will start flat lining or at least slow down. There is no such thing as exponential growth. What looks like exponential growth is always just the beginning of an s-curve because there are physical finite limits. On the other hand things could be different this time since the total amount of available dollars depends only on numbers in a computer. But if it does flat line then many people including me will pull out pretty quickly. In fact I temporarily pulled out before when it was oscillating between 30k and 40k because my loans were collecting interest during the period. So it could be people's impatient expectation of fast returns that causes a chain reaction of people pulling out.

      But in the short term I think bitcoin will do well. It still has a lot of room to grow. In the near future, the money printing will continue and there aren't a lot of other good options for where to put that money

    1. November

      According to official numbers real GDP growth has been consistently less than 3%. I suspect they are overstating the growth by underestimating inflation. Then the true inflation rate should be well above (19000/4500)^(1/21)-.03=1.04%. S&P 500 has been doing 1.04737 per year since 2000 (https://dqydj.com/sp-500-return-calculator/). No wonder gold blows everyone out of the water at (2000/350)^(1/21)=1.08 per year.

      From 1989 to 2000 we had (4600/3000)^(1/21)=1.02 per year rise with high real growth.

      From 2000 to 2021 we had (19000/4500)^(1/21)=1.07 per year rise with low growth.

    1. relationship between electoral competitiveness and news-paper circulation size and House campaign coverage

      based on measuring pixels it seems that the largest outlets have just 1-(54/91)*58/95=63.7% fewer stories mentioning both candidates

  9. Jan 2021
    1. it’s simply because people understand it’s a bad deal

      This can also be explained if you just use log($) as your utility function. Simulating in python we get an expected change in utility of sum([math.log(100 (1.5**ii) (.6 (10-ii))) * (.5 ii) (.5 ** (10-ii)) binom(10, ii) for ii in range(11)]) - math.log(100) = -0.526802578

      But the expected change in dollars is sum([100 (1.5**ii) (.6 (10-ii)) * (.5 ii) (.5 ** (10-ii)) binom(10, ii) for ii in range(11)]) - 100 = $62.889

    1. I care most about currency-related shifts in interna-tional competitiveness.

      But I just want to buy gold. Gold should actually go up by more than his prediction because everyone is printing and therefore the currencies he's comparing too will also depreciate relative to gold!!!

    2. As long as China stays the course of structural reform—shifting from manufactur-ing to services, from investment- and export-led growth to consumer-led growth—and embraces a further liber-alization of its financial system, the case for further RMB currency appreciation remains compelling

      this is a bad assumption. They literally made Jack Ma disappear for asking for reforms to the financial system. That happened after he wrote this. Nevertheless they will probably grow at least more quickly than the US

    3. personal saving rate has already begun to decline—from 33.6 percent in April 2020 to 14.1 percent in August

      accounting for our increase in personal savings is crucial. He's done his homework

    1. deepening rift between China and the US

      it all makes sense now. They want tensions with U.S. because they need an excuse to supress development. It is like how north korea uses US as a boogeyman

    1. Credit markets are pumped-up in order to supply the average consumer with more capital or buying power without increasing wages/decreasing profits

      just like the declining interest rates and super exponential growth of m2 money

    2. 77% of corporate profit growth between the dot-com bubble's peak in 2000 to the American housing bubble's peak in 2007 derived from wage deflation

      this was necessary because of we are on an s curve so rate of returns for capital is slowing down. Like I pointed out in https://hyp.is/bh954F4dEeuJFpf5V6i1Ow/journals.plos.org/plosone/article?id=10.1371/journal.pone.0237458, you have to either decrease relative worker wages to stop increasing size of the pie for elites. political stress indicator goes up either way.

      Or as marx would say this is a consequence of the tendency for the rate of profit to fall

    1. assuming value is tied to the amount of labor necessary, the value of the physical output would decrease relative to the value of production capital invested. In response, the average rate of industrial profit would therefore tend to decline in the longer term

      So his argument is fundamentally based on the assumption that everything is relative just like the sciodynamics paper. That is problematic because with that logic it is a zero sum game and there can never be any real productivity increase whatsoever.

      I would go even further and say that the rate of increase in use value will also decline because the technological process is on an s-curve

    1. In fact, the oscillations between relatively peaceful and violent periods are not cycles with fixed periods. Rather, these somewhat irregular oscillations arise as a result of dynamical feedbacks that affect the functioning of social systems.

      Then they are not really cycles. They are just markov chains. Violence leads to more violence until there's some resolution somehow. Peace and bad memories of violence leads to more peace. We are currently in a period of escalating divisions since the sides are refusing to communicate with each other and the left wants to punish the right. I suppose this is what he means by "dynamical feedbacks"

    2. ε simplifies to

      you could decrease w to increase \epsilon and thus decrease EMP. But then that increases MMP. So you are screwed either way. Either you hurt the elites or you hurt the common people. In Joe Biden's case he is caught between his base and his donors

    3. declining living standards are often a contributing factor to the social pressures for instability, reversing this trend does not end instability until elite overproduction is also reversed

      it does not matter that you can still make a good life by moving to a rural area and not go to college. It does not matter that Americans are incredibly well off compared to the rest of the planet. It is all relative

    4. w0

      this relates to the discussion on how many people are better off than their parents. Millennials come off as spoiled because they are the first generation to have a 50/50 chance of being better or worse off than their parents. So half of them are worse off (https://www.brookings.edu/blog/up-front/2018/07/25/fewer-americans-are-making-more-than-their-parents-did-especially-if-they-grew-up-in-the-middle-class/)

      So we have already reached the zero net rate of social mobility he's talking about.

      This also reinforces a theory that I used to have about china. In the early years of China's integration into the global economy it was common to say that they would start democratizing as a result but yet they have become more authoritarian. I personally speculated that this is because the economy is improving so why fix what isn't broken. The children are better off than their parents. They have high upward mobility. Once they have a strong middle class, the middle class will want more freedom and political power and therefore bring about a peaceful gradual transition to democracy. That is what would normally happen but I think China can actually prevent this by taking advantage of:

      • AI
      • modern understanding of history, psychology, and cliodynamics
      • They can simply prevent the formation of a middle class and rely on foreigners to provide new innovations and consumer demand
    5. father-and-sons cycle is expected to exacerbate the social pressures for instability

      constructive interference is coming. and this is on top of our debt cycle too!

    6. T is the proportion of the population expressing trust, and (1 –T) is the proportion expressing distrust in the state institutions

      we need trustworthy institutions

    7. s doesn’t change too much

      the loophole could be that protectionist policies and breaking up of monopolies could actually increase s. In our world it is really about corporate jobs more than government jobs

    1. What was extra ordinary in each instance was the recovery in the benchmark indices that followed. The market would halt its decline on a dime, then immediately reverse course and recover all of the lost ground with the same speed that characterized the decline. In my 22 years of market observation, I have never seen such powerful and immediate recoveries on a recurring basis. It does not typify normal market behavior, yet each time it happens I hear that it is the proverbial "money on the sidelines" coming back into the market, as investors see a buying opportunity. I find this hard to believe. Often the majority of the gains during these market rebounds occur overnight in the futures market.

      this explains why the stock market is doing so well now

    1. Gold has far and away the strongest positive relationship with both Global and U.S. Political Risk, correlating at 0.72 and 0.94

      This seems to suggest buying gold when you expect political risk. But I suspect this is more correlation than causation because there is the confounding variable of money printing and debt

    1. 37.3%

      China's secret to exponential growth is simple. They spend 37% of all imports on machinery. More money-> more machinery -> more productivity -> more money

    1. in 2009,ChinabecameJapan'slargestexportdestination,takingoverthat positionfrom theUnitedStates.Cur-rently, more thanhalfofJapan'sexportsto Chinaconsistofmachin-ery, includinghighlysophisticatedindustrialmachinerythat supportsChina'smanufacturing

      the real secret to China's growth