Posts

A whirlwind tour of Ethereum finance 2021-03-02T09:36:23.477Z
Group rationality diary, 1/9/13 2013-01-10T02:22:31.232Z
Group rationality diary, 12/25/12 2012-12-25T21:51:47.250Z
Group rationality diary, 12/10/12 2012-12-11T11:50:26.990Z
Group rationality diary, 11/28/12 2012-11-28T09:08:11.802Z
Group rationality diary, 11/13/12 2012-11-13T18:39:42.946Z
Group rationality diary, 10/29/12 2012-10-30T18:01:56.568Z
Group rationality diary, 10/15/12 2012-10-16T05:29:24.231Z
Group rationality diary, 10/1/12 2012-10-02T09:15:30.156Z
Group rationality diary, 9/17/12 2012-09-19T11:08:39.965Z
Group rationality diary, 9/3/12 2012-09-04T09:42:59.884Z
Group rationality diary, 8/20/12 2012-08-21T09:42:35.016Z
Group rationality diary, 8/6/12 2012-08-08T05:58:52.441Z
Group rationality diary, 7/23/12 2012-07-24T08:49:25.064Z
Group rationality diary, 7/9/12 2012-07-10T08:35:27.873Z
Group rationality diary, 6/25/12 2012-06-26T08:31:53.427Z
Group rationality diary, 6/11/12 2012-06-12T06:39:20.052Z
Group rationality diary, 6/4/12 2012-06-05T04:12:18.453Z
Group rationality diary, 5/28/12 2012-05-29T04:10:25.364Z
Group rationality diary, 5/21/12 2012-05-22T02:21:34.704Z
Group rationality diary, 5/14/12 2012-05-15T03:01:19.152Z
Gerald Jay Sussman talk on new ideas about modeling computation 2011-10-28T01:29:53.640Z

Comments

Comment by cata on The aducanumab approval · 2021-06-18T02:22:17.939Z · LW · GW

There's one piece here that I don't understand. Why exactly do we expect doctors to prescribe this medication that doesn't work, thereby wasting resources? Is it a situation where we expect patients and patient representatives to demand the medication, due to being scientifically illiterate (or because they are doing Hanson-esque signalling that they care) and then doctor-shopping until they can get it? Or is there some kind of pay-per-treatment incentive that will make doctors want to prescribe it? Or something else?

Comment by cata on Reply to Nate Soares on Dolphins · 2021-06-16T06:59:27.363Z · LW · GW

I'm surprised that you are making a connection between Nate's new position and modern attitudes towards gender. To me, it's just the opposite.

For me, "dolphins are fish because they swim and look fishy" is associated with a "common sense clusters" classification strategy, which also gives a traditional common-sense idea of who is a man and who is a woman.

Conversely, "dolphins are mammals because of their genetic lineage and specific mammalian traits" is associated with a kind of "cleverly selected clusters by some intellectuals" classification strategy, e.g. "people have a separate innate gender identity which determines their gender."

Comment by cata on Covid vaccine safety: how correct are these allegations? · 2021-06-13T06:57:58.976Z · LW · GW

It's really hard to take seriously a thesis that is only presented in the form of a 3-hour YouTube hangout conversation, presumably because the participants are more interested in promoting their brand than communicating information to anyone.

If these guys give a shit maybe they should write down their opinions.

Comment by cata on What is the Risk of Long Covid after Vaccination? · 2021-05-31T22:17:04.600Z · LW · GW

In areas with a lot of vaccine penetration, case counts mostly are just going to be very low anyway; combined with the vaccine effectiveness against infection (~80-90% for mRNA vaccines) the probability of being infected will be negligible. For example, in my local area and given my demographics, my current best estimate is that I incur around 10 microcovids if I spend an hour indoors breathing right on another random person. (And I am using conservative numbers.) If that's not true in your area, it may well be true in a month or two.

The only way this is going to change is if a new variant comes around and mucks everything up (in which case the current studies won't be very informative.)

So for myself, I don't really care very much about answering this question right now.

Comment by cata on lsusr's Shortform · 2021-05-25T08:20:00.726Z · LW · GW

If there are more than 1,010 "rack stolen" credits available at a price of $0.01 then you could buy all the "rack stolen" credits for $0.01 and then steal the rack yourself for $1,000.

Since this is true, it would be irrational for participants to sell the market down to $0.01. They should be taking into account the fact that the value of stealing the rack now includes the value of happening to have advance information about stealing the rack, so thieves should be more likely to steal it.

Comment by cata on We should probably buy ADA? · 2021-05-25T06:40:49.068Z · LW · GW

Ethereum is a working system that has been improving steadily with real-world feedback and a lot of invested labor for years, whereas Cardano is effectively vaporware. Outside view says that Cardano is not going to suddenly implement smart contracts and be a perfect diamond-like jewel. I believe that they may have "something" "soon", but I don't believe that they are going to catch up to Ethereum in any sense in the foreseeable future.

I think most of the value of Cardano and similar is a hedge on the event that smart contracts are a great idea but Ethereum experiences some catastrophic problem, like a major governance schism, or some giant disaster during the Ethereum 2.0 merge.

Comment by cata on [Letter] Advice for High School · 2021-05-22T20:11:48.155Z · LW · GW

Probably if you are a programmer you will be spending thousands of hours SSHed into this or that over your life, so you had better be able to view and edit files there.

Comment by cata on Covid 5/6: Vaccine Patent Suspension · 2021-05-08T19:02:45.228Z · LW · GW

That study is exactly as useless as the totally implausible sounding result would lead you to think. It has no control group (so it's counting the 10% of children who have problems like "headache" or "lack of concentration" months after infection as having a COVID symptom), the methodology of how children were selected is a total mystery (but 5% of them were hospitalized, so it's clearly biased towards dramatically more severe cases than the general population) and the sample size is tiny.

I don't know the best way to figure out how much kids are at risk but it doesn't have anything to do with reading that.

Comment by cata on Why I Work on Ads · 2021-05-04T07:05:44.236Z · LW · GW

This really doesn't work at equilibrium though. Why would I pay in advance instead of free-riding? Why would the amount I'm willing to pay reflect the value I get? That only happens if I believe that my contribution is 100% responsible for bringing about the good, but it's not clear to me how to ever get more than 1/N. The whole thing seems to rely on charity (which I do like, but for stuff that's charitably supported you don't need ads anyway).

I'm sure you've thought about it more than me, and I agree that it's not clear this will work at mega-scale as a "literally everything that requires an initial investment runs on this" strategy. However, it also really looks to me like it can work for a lot of things. Some things working in its favor:

  • Humans have a lot of intuitively operating cooperation machinery. People understand the idea of pitching in. It makes them feel like they did a good deed.
  • People respond well to social prestige as a reward for pitching in, which these platforms are getting very good at providing, through mechanisms like special badges, public credits, and access to special preview content (technically this is a loss to not to provide to everyone, but it's tiny compared to the usual model.)
  • Removing so much waste creates a huge amount of slack for people to defect and still get results of the quality they are accustomed to.
  • Pay-in-advance models typically make visible the option to price-discriminate upwards, allowing people with huge amounts of money or grantmakers to pay for more warm fuzzies, more prestige, and more choice over what gets funded. This seems to happen often enough to make a substantial difference -- you can see examples of whales on Patreon if you look at higher tier rewards that have "N of M remaining" visible.

Your scheme is also interesting and is clearly in some sense a "less hacky" approach than trying to get people to altruistically do things which are economically irrational. (Although -- sometimes it's a lot easier to get humans to do things that seem like the socially acceptable thing to do than to get them to do the economically rational thing to do.) I would like to see that tried, too.

Comment by cata on Why I Work on Ads · 2021-05-04T05:57:31.295Z · LW · GW

I don't really see my POV represented in the comments here, so I will add my 2c.

I have two strong beliefs:

  1. Ads are net bad for people watching the ads.
  2. It's really bad to create business models that involve creating a substantial marginal price for something that has little or no marginal cost, like models where you invest a bunch of money to build something and then charge people in money or attention to have it. Ads are just a different kind of price.

1 seems extremely obvious to me and I'm sort of confused why people think differently. It's true that sometimes if I see an ad, it will communicate helpful information to me. However, this is obviously not even close to commensurate with the value of my time. If I have a problem, and I want to solve the problem, I will never, ever look for ads to help inform me about solutions to the problem, because there are much, much more efficient sources of information about solutions to all problems I ever have. This is a natural consequence of the fact that advertiser incentives are grossly misaligned with my preferences, and incidentally the fact that ads are produced by random marketing professionals who are usually not actually foremost experts on my problem. (This is also true if I want to just serendipitously learn unexpected new things.)

If the guy who writes the bottom line about the diamond in the box wants to tell you five minutes of crap about why the box has a diamond, why would you listen to him for five minutes, unless you are a masochist? That's an insane way to try to learn true things. The world is filled with people who aren't specifically trying to convince you of a particular self-interested bullshit conclusion. Why not spend your time listening to them, instead?

Regarding 2, if you create a business model like this, unless it's perfect at price discrimination, it's going to produce a deadweight loss in the form of people who are not willing to pay your price, and who then will not benefit from your product, even though you could give it to them for free. Since price discrimination is hard, this is usually a gigantic loss. In addition, with digital media, you have a huge enforcement problem, so tons of resources will be burned all over to try to extract the price from thieves, and then to work around the enforcers, and so on.

We should create coordinated investment systems where people who desire a product pay in advance in accordance to how much they desire it, and the product is then given away for free to anyone. Patreon (and clones) and Kickstarter (and clones) are clearly successful examples of this, and we should try to move more and more consumer spending into that model.

Comment by cata on Wanting to Succeed on Every Metric Presented · 2021-04-13T21:45:35.147Z · LW · GW

I am a pretty serious chess player and among other things, chess gave me a clearer perception of the direct cost in time and effort involved in succeeding at any given pursuit. You can look at any chess player's tournament history and watch as they convert time spent into improved ability at an increasingly steep rate.

As a result, I can confidently think things like "I could possibly become a grandmaster, but I would have to dedicate my life to it for ten or twenty years as a full-time job, and that's not worth it to me. On the other hand, I could probably become a national master with several more years of moderate work in evenings and weekends, and that sounds appealing." and I now think of my skill in other fields in similar terms. As a result, I am less inclined to, e.g. randomly decide that I want to become a Linux kernel wizard, or learn a foreign language, or learn to draw really well, because I clearly perceive that those actions have a quite substantial cost.

Comment by cata on IQcaptcha enters beta · 2021-04-07T05:06:54.524Z · LW · GW

Funny, but I think it would be funnier if it were less unsubtly rude. The concept is already inherently edgy enough that adding extra insults is like putting frosting on ice cream.

Comment by cata on Tales from Prediction Markets · 2021-04-04T22:49:49.823Z · LW · GW

After hanging out in the Polymarket zone for a while, it became clear to me that there are at least two pretty distinct groups of people -- the +EV veterans who have fun trying to beat the market and make money, and the gamblers who like to play the game of betting on their opinions and seeing who makes bank. The latter group doesn't really like it when the +EV people find a "clever" edge like this because it makes the game less fun and fair-seeming.

There exists no third group that really wants an accurate up-to-date prediction of the CO2 level printed on the website as of the next day.

Comment by cata on Julia Galef and Matt Yglesias on bioethics and "ethics expertise" · 2021-04-01T05:46:08.344Z · LW · GW

It seems to me that there are many different complaints being raised:

A) Experts in specific things were treated as general authorities.

A lot of what we get is adjacent expertise. So somebody who studies viruses, and maybe knows a lot about the protein structure of viruses, will opine about masks, right... They didn't have expertise in those areas, and were in fact just on a par with me, or anybody else, right? But they had the, sometimes, arrogance that comes with believing you're being asked about your area of expertise.

B) People have differing moral values about what is a good result.

We're just talking about, what is a fair and reasonable way to prioritize different people over each other?

C) The arguments and decisions being made aren't actually being made by the real experts, they are being made by pseudo-experts.

And I'm like, "Well, according to whom?" Right? Obviously in consequentialist terms, it’s good ethics. I happened to know the top expert in Kantian ethics, she thinks that's a good idea. So, who the fuck are you?

I think these complaints are not incompatible with some people being experts at reasoning about ethics. It seems like what's happening (I certainly have never observed "bioethics Twitter" so I am just guessing) is that random people are rationalizing their beliefs by appealing to random prestigious or powerful people and calling them the "experts" to whom you should defer all judgment, and then Julia and Matt are pushing back on that whole general dynamic.

Comment by cata on Speculations Concerning the First Free-ish Prediction Market · 2021-03-31T07:46:42.220Z · LW · GW

You know, Polymarket really works and anyone can use it right now, including US citizens; the only really serious problem with it is that it costs a flat fee of about ~$50 to deposit and withdraw, and you have to be able to figure out how to send cryptocurrency.

Some comments:

  • As a random smart LW person, I actually rarely have interest in participating in a prediction market, because I have to work hard to have any edge on anything, and even if I have an edge I risk betting against someone who has true inside information. I would have to value my time less or enjoy modelling things more to be an active participant. It's not like the stock market where you are incentivized to use it even if you don't know anything special.
  • It's hard to get people to participate in long-term markets, because the payoff just isn't very good. I have zero desire to work hard on a model so that I can make an expected profit of 10% two years from now.
  • $25k is a joke -- is that for real? That does not pay off someone for making a good model. It also doesn't let smart money fix a wrong price even if they magically have a perfect model. You would have to organize some kind of fund with many traders working on many models who all shared info and went in on all of each others' trades.
Comment by cata on [deleted post] 2021-03-21T22:21:35.426Z

I think if you have a good system to predict prices, it doesn't make sense to add this tweeting dynamic. If you do it, and it produces good predictions about the combination tweet+price system and makes money as a result, then what will probably happen is a ton of other people clone the idea and produce similar tweeters and/or predictors until it's much harder to produce good predictions with consumer-grade models, whereas if you kept your system to yourself you could have gone on extracting money from your good predictions quietly for longer.

Naively, the system seems similar to "good predictor" human fund managers who advertise their services, where you can see the end state I'm describing: there are so many of them that most of them don't make money anymore, and even if someone shows you good evidence of their good predictions in the past it's no longer sufficient evidence to act on their advice in the present.

Comment by cata on Open, Free, Safe: Choose Two · 2021-03-21T06:16:13.495Z · LW · GW

I agree -- I think there are many communities which easily achieve a high degree of safety without "safetyism", typically by being relatively homogeneous or having external sources of trust and goodwill among their participants. LW is an example.

Comment by cata on Conspicuous saving · 2021-03-20T21:15:26.916Z · LW · GW

The ultimate explanation for the low savings rate seems to be that people are myopic. In other words, people implicitly care more about having stuff now, rather than later. But that too can't be the full picture.

What makes you think that this isn't the full picture (or most of the full picture?)

Comment by cata on Open, Free, Safe: Choose Two · 2021-03-20T20:36:47.284Z · LW · GW

I only skimmed the article you linked, but I don't think I agree with your characterization of anonymity -- it's not on the axis of openness, but the axes of freedom and safety. If someone is already let in, they may choose to be anonymous if they think anonymity helps them express things that are outside the box of their existing reputation and identity, or if they think anonymity is a defense against hostile actors who would use their words and actions against them.

Comment by cata on The EMH is False - Specific Strong Evidence · 2021-03-18T22:48:52.034Z · LW · GW

I assume the person you're talking about who made $100K is Vitalik. Vitalik knows much more about making Ethereum contracts work than the average person, and details the very complicated series of steps he had to take to get everything worked out in his blog post. There probably aren't very many people who can do all that successfully, and the people who can are probably busy becoming rich some other way.

I could have imagined this was true a month ago, but then I spent about 15 total hours learning about Ethereum financial widgets, which was fun, and wrote it up into this post, and now I totally understand Vitalik's steps, understand many of the possible risks underlying them, and could have confidently done something similar myself. Although I am probably unusually capable even among the LW readership, I think many readers could have done this if they wanted to.

Similarly, I don't know anything about perpetual futures, but I guarantee that I could understand perpetual futures very clearly by tomorrow if you offered me $20k (or a 20% shot at $100k) to do it.

Having to think hard for a week to clearly understand something complicated, with the expectation that there might be money on the other end*, is definitely a convincing practical explanation for why rationalists aren't making a lot of money off of schemes like this, but it's not a good reason why they shouldn't. Of course, many rationalists may not have enough capital that it matters much, but many may.

*It's not like these are otherwise useless concepts to understand, either.

Comment by cata on Direct effects matter! · 2021-03-16T04:32:29.672Z · LW · GW

That's fair, I might have been being a little hyperbolic, and I don't mean to say that no other people care about kids' short-term well-being. I was more pointing at the fact that if you look for discussions or advice about parenting decisions (e.g. what school to go to, how to interact with them, what activities they should do day-to-day) the majority of the focus will typically be on medium- and long-term effects (e.g. educational outcomes, behavioral training, physical and cognitive development) while ignoring the obvious direct effects on the kid, much like the example in the post about the benefits of tennis.

Comment by cata on Direct effects matter! · 2021-03-14T07:14:54.622Z · LW · GW

Parenting advice is an interesting example that might shed some light on what's going on here. It's 100% clear that as a parent you are "supposed" to care entirely for your kid's long-term well-being, ignoring the short term, and short term considerations are only important as a kind of practical "you can only push so hard" issue. The more you manage to successfully optimize your kid for the long term, the better a parent you are in the eyes of society, and that's all there is to it.

Humans do tend to favor the short term in most domains, sometimes to what seems like a stupid degree. It seems to me that in general, many direct effects are short-term effects, and many second-order effects are postulated long-term effects. So maybe exhorting about second-order effects and ignoring direct effects are really an attempt to get people to pay an appropriate amount of heed* to the long term. I certainly see that in some of your examples.

* As Eliezer memorably wrote in his meta-honesty post:

Because any rule that's not labeled "absolute, no exceptions" lacks weight in people's minds. So you have to perform that the "Don't kill" commandment is absolute and exceptionless (even though it totally isn't), because that's what it takes to get people to even hesitate. To stay their hands at least until the weight of duty is crushing them down. A rule that isn't even absolute? People just disregard that whenever.

Comment by cata on Exploiting Crypto Prediction Markets for Fun and Profit · 2021-03-13T22:29:01.351Z · LW · GW

I wrote a little at the bottom of my post. IMO probably the main "catch" is that a lot of what you're getting paid is in governance tokens that may not hold their value, and it's expensive in gas to be constantly claiming and selling them as you receive them.

Comment by cata on [Lecture Club] Awakening from the Meaning Crisis · 2021-03-12T06:00:29.545Z · LW · GW

I didn't understand the connection he was drawing between causal modelling and flow.

  • It sounded like he was really down on learning mere correlations, but in nature knowing correlations seems pretty good for being able to make predictions about the world. If you know that purple berries are more likely to be poisonous than red berries, you can start extracting value without needing to understand what the causal connection between being purple and being poisonous is.
  • I didn't understand why he thought his conditions for flow (clear information, quick feedback, errors matter) were specifically conducive to making causal models, or distinguishing correlation from causation. Did anyone understand this? He didn't elaborate at all.
Comment by cata on A whirlwind tour of Ethereum finance · 2021-03-04T04:43:09.184Z · LW · GW

It would be even easier than that. Suppose for the sake of simplicity you put your whole million dollars of Bitcoin into Compound. You could then withdraw USD up to the point where you hit the collateralization ratio set by Compound, e.g. if the collateralization ratio is 150%, you could withdraw $666k of USD.

When Bitcoin goes up 1000x, your collateral is now worth a billion dollars, but you have still only borrowed $666k (plus interest). You have way more collateral than you need. So you could just withdraw 99.9% (minus interest) of the collateral, if you wanted. You don't even have to repay anything to do that.

Comment by cata on A whirlwind tour of Ethereum finance · 2021-03-03T23:03:28.183Z · LW · GW

Well, here's a scenario: Suppose you have owned Bitcoin since 2012 and now you have a million dollars of Bitcoin, but when you sell them you will have to pay ~$250,000 in capital gains tax, since you live in California. You still endorse owning the Bitcoin because you think it's headed for the moon, but you want to use that capital for some other stuff, like buying a Tesla.

In this case you could deposit $200k of Bitcoin into Compound, withdraw $100k of USDC, cash the USDC out for dollars at Coinbase, and buy your Tesla. Now you are still exposed to the upside on all your Bitcoin but you also have your Tesla.

In practice I actually don't know what kind of taxable event "taking out a loan at Compound" is so I am not sure this was a correct way to minimize taxes. (On the other hand, even if it's not, I bet people are doing it; the IRS actually looks if you sell Bitcoin on Coinbase, but I don't think it looks at what you're doing with Compound.) But at least it definitely preserved your exposure to Bitcoin. That's a way to "put it to good use" even when it just sits there.

This is one example. As I mentioned in the post, I'm not sure what is driving the majority of the demand.

Comment by cata on Takeaways from one year of lockdown · 2021-03-03T09:54:38.874Z · LW · GW

This post is fascinating to me because I have no idea what to say. I keep coming back to it, reading the new comments, starting to type, and giving up. It seems like the kind of thing where I should be able to think "what would I do" and generate some opinion, but I have no idea what I would do, or what should have been done.

I would be very interested to see more people reflecting on situations like this and thinking about what should have been done. This was obviously an unusual event, but not so unusual that I don't expect future events to resemble it in many ways.

Comment by cata on A whirlwind tour of Ethereum finance · 2021-03-03T04:48:15.098Z · LW · GW

I guess I expect the edge to manifest in the form of being able to look at simple, high-upside, good ideas being implemented by smart people, correctly distinguish them from uninspired hacks, and then being able to take effective action on your beliefs. (Good ideas like Bitcoin or Ethereum themselves, or like CFMMs.)

As a random example, if you go look at harvest.finance right now, you can see that there's a huge APY on investing in Uniswap liquidity pools associated with tokenized versions of public company stock, like AAPL. These tokens are designed by a new project called Mirror which I know nothing about. Do they work? Is the project credible? Is there demand for this product?

It's possible that if you spent a few hours looking into it, you could put a decent confidence interval on the probability it will succeed or crater, and if it's likely to succeed, you can make a lot of EV on these pools.

In the long term, if Ethereum (or some replacement) can continue scaling up, there's just going to be more and more stuff that is begging to be reimplemented as interoperable tokenized contracts, and it's all going to have to go from zero to sixty. If someone makes the Ethereum version of, like, Ebay, or Twitter, and you can recognize that it's any good and invest in it, that's a big edge.

Comment by cata on A whirlwind tour of Ethereum finance · 2021-03-03T03:25:37.484Z · LW · GW

I know, right?! Also, check out this article "Ethereum Is a Dark Forest" if you haven't seen it yet.

Phil once told me about a cosmic horror that he called a “generalized frontrunner.” Arbitrage bots typically look for specific types of transactions in the mempool (such a DEX trade or an oracle update) and try to frontrun them according to a predetermined algorithm. Generalized frontrunners look for any transaction that they could profitably frontrun by copying it and replacing addresses with their own. They can even execute the transaction and copy profitable internal transactions generated by its execution trace.

Comment by cata on A whirlwind tour of Ethereum finance · 2021-03-02T19:58:48.781Z · LW · GW

Thanks! I literally didn't know how any of these things worked a week ago so I'm glad if more knowledgeable people approve.

Comment by cata on A No-Nonsense Guide to Early Retirement · 2021-02-24T22:48:43.638Z · LW · GW

I made a simple spreadsheet recently to help project my future net worth based on market returns and expenses. It takes as input your current net worth (the asset class breakdown is just informational), recurring expense categories and predicted inflation rates for those categories, predicted one-off expenses, predicted future year income, and S&P 500 history, and simulates potential outcomes based on simple models like "if I earned 8% yearly" or "if I started in 1980 and owned the market."

The main reason I did it instead of using existing tools (e.g.) is that most existing tools seem to think I will constantly work until time T and then "retire" and never work again, which makes zero sense to me unless you assume there is some time T where I suddenly get a giant windfall of money, or unless you assume that I value each marginal dollar at infinity until I am unemployable or dead. I want to be able to play with questions like "what if I don't work until my kid is 10 years old, and then resume working."

Comment by cata on This is my 100ᵗʰ post on Less Wrong · 2021-02-15T07:34:03.608Z · LW · GW

Your contributions have been a highlight of the past couple of years. Please keep writing!

Comment by cata on Who should you expect to spend your life with? · 2021-01-23T21:10:56.343Z · LW · GW

If I had to guess, I bet that when children leave the home, partner time spend becomes more bimodal, with people who like their partners spending more time together, and people who were mostly co-parents with their partners spending less time together. This could net out to making it look like there's not that much change.

Comment by cata on Evaluating Opus · 2020-08-09T00:16:43.248Z · LW · GW
In our use case, however, we're not talking about sending one large recording up to the server. Instead, we'd be sending batches of samples off every, perhaps, 200ms. Many compression systems do better if you give them a lot to work with; how efficient is opus if we give it such short windows?
One way to test is to break the input file up into 200 ms files, encode each one with opus, and then measure the total size. The default opus file format includes what I measure as ~850 bytes of header, however, and since we control both the client and the server we don't have to send any header. So I count for my test file...

Based on my understanding of what you are building, this splitting is not a good model for how you would actually implement it. If you have a sender that is generating uncompressed audio, you can feed it into the compressor as you produce it and get a stream of compressed output frames that you can send and decode on the other end, without resetting the compressor in between.

Comment by cata on What can currently be done about the "flooding the zone" issue? · 2020-05-20T04:25:20.612Z · LW · GW

It seems to me that ordinary people are usually obsessed with scandals and drama, so I think Occam's Razor says that journalists are also obsessed with them and report on them because that's what they want to do. That is, I don't think all journalists are secretly wishing they could ignore Donald Trump but regrettably recognize that their incentives point in another direction -- I think they enjoy reporting on Donald Trump, so they do it.

Comment by cata on The Best Virtual Worlds for "Hanging Out" · 2020-04-27T23:31:31.799Z · LW · GW

Small correction -- AltspaceVR can be used without a VR headset using normal FPS controls, just like Hubs.

Comment by cata on Ethernet Is Worth It For Video Calls · 2020-04-13T06:27:27.041Z · LW · GW

Hey, I'm not sure what is wrong with your WiFi, but something fixable is probably wrong. I can run speedtest.net (on my roughly 70/10 connection) while pinging my router (an Archer C7 running OpenWRT which is roughly across the living room -- I have a home network with one AP at each end of the apartment) over WiFi and the ping time never exceeds 4ms. I live in an apartment with dozens of other visible networks.

You might want to check that it's using 802.11ac with 5 GHz on a not crowded channel, check that the signal strength looks good, make sure your client network drivers are up to date, check whether you have router hardware/software known for poor performance under load or bufferbloat problems, etc.

(edit: I notice you say that the router is three rooms away -- my guess is that the signal strength is just really bad and if you put an access point in between you could make it much better.)

Comment by cata on April Coronavirus Open Thread · 2020-04-01T08:30:37.688Z · LW · GW

Mea culpa, that's more of a condemnation than I thought.

Comment by cata on April Coronavirus Open Thread · 2020-04-01T02:53:54.459Z · LW · GW

It seems only skeptical of cloth masks as compared to surgical masks, which isn't really very interesting to me in the current circumstances, since most people don't have access to surgical masks.

Comment by cata on Coronavirus: California case growth · 2020-03-30T00:14:22.633Z · LW · GW
The data here doesn't give a clear idea of how the transitions from 1 to 2, or from 2 to 3 are proceeding. Nonetheless, it may offer some clues. So first, let's backtrack and think: let's say California going to level 2 or level 3 did in fact effectively stop coronavirus in its tracks. What should we see?

Ideally, we should see the number of people with coronavirus getting the test drop a lot. However, that doesn't necessarily mean that the total number of people getting the test drops, because many people who don't have the disease may also start getting tested, causing the total number of people getting tested to increase. So, more accurately, we should see one of these:

- A drop in the incremental number of tests each day.
- A drop in the confirmed positive rate on tests (but this metric is available at a further lag of 5 to 7 days).

I think it could take longer before either of these reflects the change in true cases. Here's an argument. Suppose:

- Current testing policy is declining to test many symptomatic people due to lack of capacity. I believe this is true (high-risk people, essential workers, and known contacts to existing cases are being prioritized.)
- As test availability improves, testing policy will change to test broader categories of symptomatic people, up to the testing capacity.
- The number of true cases is substantially higher than the number of other ailments that basically look the same. As a result, P(positive test | symptomatic) remains high and doesn't change much even if you halve the number of true cases. Probably variance in testing policy and test accuracy will drown out the change.

If this is right, as long as the number of true cases remains above the threshold of testing capacity, we get roughly the same number output on the metrics you mentioned, no matter whether it's 10 times above capacity or 1000 times above capacity. So if we're way above capacity right now, we won't see a decrease in true cases show up in those metrics for a while.

Comment by cata on March Coronavirus Open Thread · 2020-03-12T08:52:24.257Z · LW · GW

Here's a preprint published on March 10th testing how long coronavirus can last on a variety of surfaces.

Comment by cata on Pricing externalities is not necessarily economically efficient · 2019-11-10T06:27:20.495Z · LW · GW

What a wonderful page to come across -- I had misunderstood this for a decade.

Comment by cata on Heads I Win, Tails?—Never Heard of Her; Or, Selective Reporting and the Tragedy of the Green Rationalists · 2019-10-01T00:27:07.248Z · LW · GW

Wow, this post really got me thinking. Dealing with this kind of pervasive filtering seems super important and also difficult. Thanks for writing it.

Your proposal that you can clue other people into your filtering mechanism seems hard in practice. Any effort in this vein means that you are saying that you suspect the consensus is misleading, which means that you are saying that non-consensus beliefs are more likely to be true, and people can pick up on this. I tried to come up with ways to express "this is cherry-picked, but" without triggering that but I couldn't figure out ways that seemed plausible.

Approaches like Raemon's, where you say "I'm just never going to talk about controversial thing X" seem mentally hard to update on -- in a world where there are a million people who say "I'm never going to talk about X" and a thousand who are constantly presenting cherry-picked evidence about X, it's very difficult for my mind to interpret the filtering disclaimers as object-level evidence about X that can fight with the evidence provided by the cherry-pickers.

Comment by cata on No Really, Why Aren't Rationalists Winning? · 2018-11-05T04:51:45.430Z · LW · GW

The people I know IRL who identify as rationalists are doing great. Not a lot of people bet on prediction markets since the ones that exist are small and hard to use. Not a lot of people bet on stock markets since making money doing so is a boring full-time job.

I presume that the reason people don't post about how they are "winning" is because it's tactless to write a post about how great you are.

Comment by cata on OpenAI releases functional Dota 5v5 bot, aims to beat world champions by August · 2018-06-26T23:04:26.034Z · LW · GW

As a strong amateur DotA player, this seems absurdly impressive, albeit much less impressive than if they could beat humans in the full DotA ruleset. (Their restrictions seem to be avoiding much of the "long-term strategic planning" part of DotA, e.g. choosing heroes, item builds, and level of aggression in a way that makes sense against the enemy team.)

This seems way more surprising to me than beating humans at Go.

Gwern wrote a summary comment with some opinions here.

Comment by cata on Duncan Sabien: "In Defense of Punch Bug" · 2018-05-04T07:06:02.377Z · LW · GW

I thought just the same as you. It's extremely confusing to me that D would say "ugh" and then be unhappy when C changes their behavior. I didn't know this was a way people felt.

That said, I don't interpret "ugh" as a very strong comment. I wouldn't bother thinking to change my behavior because one person said "ugh."

Comment by cata on Are long-term investments a good way to help the future? · 2018-04-30T22:51:49.741Z · LW · GW

If you're willing to buy or own equity in some particular company, your demand is driving up the price for that company's equity. That company may own some of its own equity, or can issue new equity, and can then sell it at a higher price, or use it to pay for things (e.g. employee compensation, deals with third parties) directly. As a result, they can acquire capital to do whatever kind of presumably-profitable endeavor they can think of, presumably creating real economic value (or else it shouldn't be making a profit.) Your trade returns and dividends are a portion of that value. At least, you know, in theory.

Comment by cata on The Case Against Education · 2018-04-17T11:04:41.218Z · LW · GW

I agree wholeheartedly with Zvi's aesthetic opinion, which is that schools seem absurdly terrible, like a kind of dumb, terrible joke. I certainly wouldn't want to have to go to school again myself. I might rather select some different kind of minimum-security prison if I could.

However, I am not a child any more. When I was a child, I was different and I had a different personality. I went to a public school and a STEM magnet high school, and it was fine. It was a big waste of time, mind you. (I declined to go to college.) But it wasn't an incredibly unpleasant big waste of time, and unlike Zvi's experience it wasn't obviously worse than if my parents had just left me at home alone.

(I'm actually curious if Zvi would elaborate on his negative experience, which I didn't really understand yet. What's so unpleasant about sitting in class expending a tenth of your brainpower, if you are a smart kid? I sat in my class expending a tenth of my brainpower and I passed notes to people and programmed my calculator and read books and did math and thought about chess. That doesn't sound maximally productive, but it certainly doesn't sound torturous.)

There are many children like Zvi, who self-report hating school. Probably there are also many children who are ready to do something other than dick around in day care until they are N years old. I hope alternative arrangements flourish for those children and I am hopeful that my future children are among them. But there are definitely many, many children who are more or less OK with schools as they are, and all their stupid mediocrity. I'm wary of our crowd of outliers with our adult eyes looking with vicarious horror at something which is not actually that bad for the principals involved.

Comment by cata on An Untrollable Mathematician Illustrated · 2018-03-20T21:14:57.846Z · LW · GW

Thank you for making this! The format held my attention well, so I understood a lot whereas I might have zoned out and left if the same material had been presented more traditionally. I'm going to print it out and distribute it at work -- people like zines there.

Comment by cata on A LessWrong Crypto Autopsy · 2018-01-30T21:29:09.573Z · LW · GW

Thanks for this interesting infodump, I haven't looked hard at EOS yet (I am biased towards only bothering to learn about things which are actually "live in production" in a meaningful way, as a kind of bozo filter) but I will promote it in my attention.

I agree that Bitcoin could sort of tack on scalability with payment channels and tack on smart contracts with RSK. It just seems really unlikely that it's a better idea to do these things on top of a PoW blockchain layer that has tire fire governance. It's definitely true that some kind of worse-is-better network effect* could keep Bitcoin at the top in a way that's hard for me to understand -- I don't even have any explanation for why Bitcoin is at the top as of right now! At the end of the day, though, my original decision to buy Bitcoin and later Ethereum was based on my assessment of the technical usefulness of the work being done, and since that turned out so well, I'm going to keep riding that heuristic until I see it not work.

*Although it's interesting to speculate what the network would be. Ethereum has mostly caught up in terms of ease-of-obtaining-and-transferring; exchanges, wallets, node software, etc. Bitcoin never really hit the mainstream in terms of people transacting in it for everyday things, and due to congestion it probably won't now until payment channels are running and the kinks get worked out. One interesting thing that might be a "network effect" is the number of people with non-portable investments into Bitcoin's future, e.g. miners with Bitcoin ASICs. That's a lot of capital which is incentivized to make Bitcoin in particular succeed, and competitors don't have that.