Previous: 2011, 2012-2013

“It cannot be gotten for gold, neither shall silver be weighed for the price thereof. / It cannot be valued with the gold of Ophir, with the precious onyx, nor the sapphire. / The gold and the crystal cannot equal it: and the exchange of it shall not be for vessels of fine gold. / No mention shall be made of coral, or of pearls: for the price of wisdom is above rubies.”

Another 477 days are past, so what have I been up to? In roughly topical & chronological order, here are some major additions to gwern.net:

Statistics:

QS:

Black-markets:

Bitcoin:

Tech:

Literature/fiction

Misc:

Site:

  • I began A/B testing my site design to try to improve readability:

    • no difference between 4 fonts
    • no difference between lineheights
    • no difference between the null hypothesis & the null hypothesis
    • a pure black/white foreground/background performed better than mixes of off-colors
    • font size 100-120%: default of 100% was best
    • blockquote formatting: Readability-style bad, zebra-stripes good
    • header capitalization: best result was to upcase title & all section headers
    • tested font size & number size & table of contents background: status quo of all was best
    • BeeLine Reader: no color variant performed better than no-highlighting
  • anonymous feedback analysis (feedback turned out to be useful)
  • deleted Flattr, trying out Gittip for donations; Gittip turns out to work much better
  • I began a newsletter/mailing-list; the back-issues are online:

New Comment
32 comments, sorted by Click to highlight new comments since:

Evolution’s attempted blackmail of me

The linked article expresses doubt that it was really Evolution, but this summary doesn't. I'm not sure if that matters to you or not, but I thought I'd note it.

For what it's worth, the "Sand" page doesn't seem to be linked anywhere on your homepage.

Yeah. I intend to add it to the index whenever I get the survey responses properly analyzed. The rest of it is still interesting though.

With regards to your nootropics and evolution article, something to note is that we in fact do appear to have a system for regulating dopamine up when sleep deprived (dopamine is imitated by drugs such as amphetamines), to keep us going. That's a system that wouldn't be specific to sleep deprivation if there weren't strong negative side effects.

Ultimately when you have a complex analog circuit network with feedback loops and gain, it ought to be pretty obvious what may be the negative side effects from amping up the gain (Runaway positive feedback. There's some exciting new research in computational neuroscience on this).

With regards to your nootropics and evolution article, something to note is that we in fact do appear to have a system for regulating dopamine up when sleep deprived (dopamine is imitated by drugs such as amphetamines), to keep us going.

We do? As far as I've ever seen, sleep homeostatic pressure is pretty relentless, there is no off switch for it. As far as I know, modafinil's mechanism hasn't been pinned down, but modafinil's dopamine effects (while much touted in the media as indicating risk of addiction despite the abundant real world evidence against addiction being a substantial risk) don't seem to be large and some of the followup research was weak.

I'm on the phone, google sleep deprivation dopamine. A built in equivalent of fighter pilots taking amphetamines to stay awake.

edit: this should get you started . Ultimately, some important functions having to do with adjustments of synaptic strength (e.g. synaptic renormalization) seem to be done during sleep, and it makes sense that animals evolved a mechanism by which they can partially compensate for the lack of maintenance and post-pone such maintenance until next night.

With regards to the whole nootropics issue, none of the very high performing people I know regularly takes some wonder drug (other than caffeine, and even that in moderation), nor is any wonder drug common in programming contests or other intellectual sports. You can make a case that IQ hits diminishing returns. But human head size was incredibly expensive due to the risks of childbirth (and head trauma too, common cause of death in prehistoric societies), and the returns on shrinking the brain while maintaining the performance are anything but diminishing.

I'm on the phone, google sleep deprivation dopamine. A built in equivalent of fighter pilots taking amphetamines to stay awake. edit: this should get you started .

I don't know what that means. All that shows is in one small sample, dopamine levels went up after staying up. Well, that's not surprising: they were either going to go up or go down, so you might as well flip a coin to choose...

With regards to the whole nootropics issue, none of the very high performing people I know regularly takes some wonder drug (other than caffeine, and even that in moderation), nor is any wonder drug common in programming contests or other intellectual sports.

Why do you think you would know and on what basis do you claim they are not common? Most people carefully conceal any drug use, especially when it's something like modafinil or Adderall they either are obtaining illegally or have fraudulently gotten a prescription for. Modafinil usage is notorious in Silicon Valley, amphetamine use (in the form of Adderall and Vyvanse) is even more notorious in the Ivy League and higher ed in general, and surveys like the Nature survey suggest that quite a few top scientists are very quietly abusing stimulant on the side.

Alcohol use is very notorious in the higher ed in general too, so is all sorts of drug use. Hollywood is notorious for drug abuse too.

Are you referring to this survey by chance? Around 20% ever used, 10% regularly.

It's perfectly consistent with my opinion that gains are a: generally small, and b: likely occur only in a portion of the population. (a random small change to brain parameters should be expected to improve performance in about 50% of people, assuming brain parameters are not precisely at the local maximum for performance)

You can contrast that with physical sports, where if you just let people dope, people who aren't doping wouldn't stand a chance (and people who are would die young).

Not entirely sure we even substantially disagree.

Alcohol use is very notorious in the higher ed in general too, so is all sorts of drug use. Hollywood is notorious for drug abuse too.

The privileges of wealth and intelligence and self-discipline - you get to afford drugs and avoid the bad parts and get the good parts, like the possible longevity boosts of moderate alcohol consumption.

Are you referring to this survey by chance? Around 20% ever used, 10% regularly.

Just one of many surveys. Like iodine surveys of pregnant women, there seems to be a cottage industry of surveying college students and academics to find out how much Ritalin/modafinil/Vyvanse/Adderall/etc they use lately.

It's perfectly consistent with my opinion that gains are a: generally small, and b: likely occur only in a portion of the population. (a random small change to brain parameters should be expected to improve performance in about 50% of people, assuming brain parameters are not precisely at the local maximum for performance)

'I don't know anyone smart who uses them.'

'Here's a ton of real-world evidence that shows tons of smart people use them, think they benefit, and also the obvious explanation for your ignorance'.

'That's all perfectly consistent with my opinion...'

You can contrast that with physical sports, where if you just let people dope, people who aren't doping wouldn't stand a chance (and people who are would die young).

Pro athlete dope all the time in highly organized circles; do BALCO or Lance Armstrong ring any bells? While it's hard to make any overall comparisons because pro athletes are a very selected part of the population and one might expect them to either live much longer than average (because they're selected for good health and strong bodies and get lots of exercise) or much less (because sports can break down bodies, sometime grotesquely so in the case of the NFL recently admitting brain degeneration is endemic among pro football players and setting aside hundreds of millions of dollars for their treatment), the little research I've seen doesn't show any big decreases in longevity as one would expect if doping were really that bad.

doesn't show any big decreases in longevity as one would expect if doping were really that bad.

"Doping" is a very wide category -- ranging from pretty harmless diuretics to force-feeding kids hormones (East Germany). I don't think one can speak about health effects of "doping" in general. Is typical use of anabolic steroids bad for longevity? Schwarzenegger looks pretty healthy :-/

'Here's a ton of real-world evidence that shows tons of smart people use them, think they benefit, and also the obvious explanation for your ignorance'.

I don't hang out specifically with the folks that use those drugs, or transhumanist crowd. If 10% of people use those drugs regularly, and I am talking of 10 top people or so who i know personally and about who I would know if they used those drugs regularly, then I need not know of a single individual, need I? Out of how many high performing people you know (independently of drug use), how many use those drugs regularly?

Re: athletes, the point of doping regulations is to limit doping. With no regulations, you'd have some utterly insane cocktail of hormone substitutes delivered by an implant pump, especially for burst-of-activity sports such as short distance running. And it'd be mostly a biochemistry contest.

If 10% of people use those drugs regularly, and I am talking of 10 top people or so who i know personally and about who I would know if they used those drugs regularly, then I need not know of a single individual, need I? Out of how many high performing people you know (independently of drug use), how many use those drugs regularly?

My crowd is very different from most people's, so I prefer not to generalize from it and instead focus on surveys when I want to know whether stimulant use is widespread particularly among high-performers.

With no regulations, you'd have some utterly insane cocktail of hormone substitutes delivered by an implant pump, especially for burst-of-activity sports such as short distance running. And it'd be mostly a biochemistry contest.

I don't believe every sport has an effective anti-doping regime, and we have not seen any of that... (Which is better, a biochemistry contest, or what we have now, a genetics and luck contest?)

Which is better, a biochemistry contest, or what we have now, a genetics and luck contest?

The genetics-and-luck contest is better in Yvain's Moloch sense -- it provides comparatively less incentives to sacrifice long-term utility for short-term advantage.

Worse in the positive externalities sense, though.

You mean you get volunteer guinea pigs to run research on?

Well, presumably you'd want to be doing your actual research on something other than the athletes you're working with; pro-level athletes are neither common nor cheap. But getting better at sports biochemistry sounds like it'd do more outside its domain of application than getting better at, say, water polo coaching.

pro-level athletes are neither common nor cheap

Under the let-the-better-potion-win system what makes you pro-level is that you work with a good lab which makes effective potions. The incentives for second and lower-tier athletes are to get to the bleeding edge and push -- that's the only way for them to get to top tier. I doubt there will be a lack of willing test subjects.

It's probably still better to just ban the doping from human sports and allow them in, say, pig races.

My crowd is very different from most people's, so I prefer not to generalize from it and instead focus on surveys when I want to know whether stimulant use is widespread particularly among high-performers.

Well, what's "widespread" in your book, 10% ?

For an illegal, expensive, and according to many people immoral, practice, I'd say >10% is widespread, yes.

Dunno if legality even has much an impact. Many things are a lot less illegal in some places, you know, without dramatic increases in use. Expensive, like, for high performers? Are we talking again of the college students, hung over and sleep deprived due to partying all night, taking stimulants?

Many things are a lot less illegal in some places, you know, without dramatic increases in use.

Those other things are in different circumstances and are other things. In general, making something illegal is... probably going to reduce how many people do it. Odd, I know. (If modafinil had been invented 200 years ago and had never been regulated, its usage in the general population would probably be similar to caffeine and nicotine, which is a much higher figure than its actual current usage.)

Expensive, like, for high performers?

Prescription modafinil is something like $300 a month or $3.6k a year. It's worth it, especially for high performers in the highest-wage countries in the world like the USA, but it's still a nontrivial financial expense on top of the reputational problems, unease with the concept, and worries about the legal risk and unknown unknowns.

BTW did we ask people whether they drank on the LW Census/Survey?

Looks like no; the 2013 survey doesn't include the words "alcohol" or "drink". It doesn't ask about recreational drugs either, although it does ask about prescription psychiatric medication and dietary supplements.

The 2012 survey asks about nicotine, modafanil, and caffeine, but not alcohol or other drugs. Earlier surveys don't ask about any of the above.

Wow, I just realized that the 2014 survey is probably coming up quite soon. That's exciting.

Unfortunately, it probably won't ask about the things we really need to ask, like the basilisk.

That would certainly shut some people up, but good luck selling it to Yvain. It's certainly not the path of least resistance.

It'd be like a scientology forum asking about thetans.

Speaking of which, http://en.wikipedia.org/wiki/Academic_doping#Potential_for_addiction references studies that found strong links with other drug use. Also, interestingly, it's 4.5% ever used in Germany.

1 question following the nootropics update: has there recently been any more data on side effects and risks of the top-rated noots in that link (e.g. Armodafinil)?

I haven't seen anything particularly noteworthy about their risks. I'd say I'm a little more skeptical of fish oil than I used to be, and a bit more favorable about adrafinil.

[-][anonymous]00

no difference between the null hypothesis & the null hypothesis

Just as I expected.

[This comment is no longer endorsed by its author]Reply