“AI is powerful and uncontrollable and could kill all of humanity, like seriously” is not a complicated message.
Anything longer than 8 morphemes is probably not going to survive Twitter or CNN getting their hands on it. I like the original version (“Literally everyone will die”) better.
In what sense? The problem of potential helium scarcity has been (effectively) solved in the last few years by just looking for more helium deposits.
In a sense, we are “running out” (because there’s only a finite supply of Helium), we’re just running out very, very slowly.
Actually, maybe I should clarify this. This is standard practice when you hire a decent statistician. We’ve known this since like… the 1940s, maybe?
But a lot of organizations and clinical trials don’t do this because they don’t consult with a statistician early enough. I’ve had people come to me and say “hey, here’s a pile of data, can you calculate a p-value?” too many times to count. Yes, I calculated a p-value, it’s like 0.06, and if you’d come to me at the start of the experiment we could’ve avoided the million-dollar boondoggle
that you just created.
You’re completely correct! However, it’s worth noting this is standard practice (when the treatment makes up most of the cost, which it usually doesn’t). Most statisticians will be able to tell you about this.
So I think I have two comments:
It’s actually pretty neat you figured this out by yourself, and shows you have a decent intuition for the subject.
However, if you’re a researcher at any kind of research institution, and you run or design RCTs, this suggests an organizational problem. You’re reinventing the wheel, and need to consult with a statistician. It’s very, very difficult to do good research without a statistician, no matter how clever you are. (If you’d like, I’m happy to help if you send me a DM.)
Could be marketed as a medication for hypersomnia, narcolepsy, chronic fatigue, and ADHD.
There’s an EA cause area in this, but it’s not tooth-related. Our jaws are perfectly fine. The problem is bad incentives in the healthcare system; dentists get paid to take wisdom teeth out, not to leave them in, so about 90% of wisdom tooth extractions are unnecessary.
The mortality rate of anesthesia is not quite negligible; a couple hundred people have died because of anesthesia during wisdom tooth removal, and there’s also the risk of infection.
With regard to differing forms of , the article you linked notes:
Despite these minor differences in processing, each of these forms is probably equally effective when equal doses are given.
Which is understating things a bit: All the minor variants will be identical after dissolving in the body—anyhydrous creatine will become hydrated, and particle size is 0 after the creatine has dissolved. The only (minor) difference is that micronized creatine will dissolve slightly faster if you mix it with a drink, because of the smaller particle size. This can be mildly convenient, but I’ve never been bothered by the monohydrate being too slow to dissolve (it’s almost instant for both forms).
With respect to the impacts on cognitive performance, Gwern completed an in-depth look here. Most important is the section on publication bias; the authors of these papers failed to replicate their results several times, but these failures were never published. Gwern concludes (and I agree) that creatine almost certainly doesn’t affect intelligence in nonvegetarians, and has at best
The study you quote at the end about vegetarianism has a sample size in the low dozens, and I wouldn’t put much stock in it. The large meta-analyses of this all find better life expectancy and health markers in vegetarians/vegans. The one important exception is B12 (for which the study you cite finds a deficiency in vegans). Luckily, th
is can easily be fixed by supplements.
I’d rephrase that, maybe. Whether we’re asking people to make a change doesn’t matter—that’s assigning an unearned privilege to old ideas. The reason we have to justify this claim is because it’s a claim. If you’re saying “I know vegetarianism is healthier,” you should be able to explain how you know that.
If you’re wondering whether the above studies leave some room for including fish in your diet: My guess is yes, pescetarianism is likely to be slightly healthier, because of the aforementioned effects of polyunsaturated fats (including omega-3 fatty acids). This is why doctors also recommend the Mediterranean diet a lot, although it’s not clear how much of the effect comes from fish and how much from vegetables. Omega-3 supplements definitely help.
The quality of research is already there: See my comment here.
I’m not aware of any newer studies on creatine after the original one that spawned this belief; I will point out the original studies were likely aggressively p-hacked, showed impossible effect sizes, and make very little sense because brain creatine levels are not substantially different between vegetarians and vegans.
Still, it’s not completely out of the question there’s a real effect. We still want to avoid the fallacy of the one-sided bet. It’s definitely possible that vegetarianism causes a drop in . But given vegetarianism seems to lower blood pressure, LDL cholesterol, and obesity rates, I’d be willing to put money on the exact opposite. A healthier kid is a smarter kid, and vegetarianism seems to have positive effects on health in general.
If you’re still skeptical, over-the-counter creatine is extremely cheap (and has benefits for physical performance anyways). Trying to get rid of LDL cholesterol, on the other hand, is pretty hard.
With regards to protein: studies consistently show getting enough protein to prevent sarcopenia or other diseases is not especially difficult, and in fact most people get substantially more than the recommended daily allowances. Also important is that consuming animal protein is positively correlated with all-cause mortality, while plant protein is negatively correlated. From what I can tell, though, this is probably because fiber is confounding the results.
Moving on from observational studies and to randomized controlled trials, it’s unclear whether protein even has a meaningful effect on muscle strength, let alone something like lifespan, with some studies finding little or no effect from supplementation. It’s really not clear protein would help with sarcopenia from the studies, but we’d honestly expect null results from theory alone. We know from surveys that the limiting factor in most people’s muscle strength is physical activity, not protein. Most people get more than enough protein, and excess protein consumption turns into fat, not muscle. For most people, protein supplementation just can’t help because they aren’t moving enough.
The primary dietary determinants of health seem instead are mostly:
Polyunsaturated fatty acids (found in fish and vegetable oils)
Fiber (typically supplemented with psyllium husk)
I’m busy, so I’ll stick to looking at the evidence base we’ve acquired just on fiber supplementation. Here’s a bunch of insanely overpowered observational studies with sample sizes stretching into the hundreds of thousands, all showing reductions in all-cause mortality. Meta-analyses of RCTs? We’ve got those in spades.
There is a reason every doctor’s dietary advice is exactly the same: “Replace refined with whole grains, avoid saturated fats and replace them with unsaturated fats.” The reason is because we have million-person studies showing these interventions reduce heart disease. Note that heart disease is by far the leading cause of death among the elderly; sarcopenia doesn’t really kill, although it does make life substantially worse. And importantly, both of these are substantially more common in plants than in animal foods, which offers a pretty good explanation for why vegetarians seem to live longer, even after controlling for confounders. And also for why vegetarian diets seem to improve health markers in RCTs.
The value of both can both rise or fall, but real estate is only an investment when rented out. Otherwise, it’s a durable consumption good. In particular, the EMH* implies the expected value of buying real estate and renting it out must be equal to the expected return on stocks. Otherwise, people would stop sell stocks (driving their price down, and therefore the rate of return up) and then buy real estate to lease it out.
*While it’s entirely plausible the EMH doesn’t hold, no analysis arguing this is presented, and I don’t think that placing bets on certain sectors of the economy is a particularly good idea for a charity. Notably, arguments against the EMH almost all fall on the side of suggesting the housing market is currently overvalued because of structural deficiencies (like the inability to short housing) and subsidies that make buying cheaper for individual homeowners (but not charities)
.
Not sure, I don’t know all that much about them, unfortunately.
Is Wytham Abbey being used 365 days a year, though? A couple thousand a day is a perfectly reasonable cost for a conference. But why would you need to own a space that large permanently? I’d be just as shocked at the idea of renting out Oxford College’s conference hall 365 days a year. Nobody is having that many conferences.
The calculations there are completely correct under the assumption that the space is being used 365 days a year, which strikes me as wildly implausible. I was working on the assumption the is space used a few days each year. If this space is actually being occupied 100% of the time, I’d gladly retract my criticism.
The question here isn’t whether building a new castle from scratch would cost £15,000,000, or whether the real estate is properly valued; I assume it is. The question is whether you need to buy a castle to host yearly or monthly conferences—to which the answer is clearly not.
Investing £15,000,000 would yield roughly £1,000,000 a year on the stock market. If you are spending a million pounds on the venue for a conference, you are not doing it right. A convention hall runs in the tens of thousands of dollars, not the millions. This is a 10-100x markup.
£15,000,000 sounds like a lot to a researcher because it is.
I refuse to believe that renting out a conference hall would actually have cost more.
Investing £15,000,000 a year would yield roughly £1,000,000 a year on the stock market. If you are spending a million pounds on the venue alone for a 1,000 person conference, you are not doing it right. A convention hall typically runs in the tens of thousands of dollars, not the millions. This is a 100x markup.
A typical researcher might make £100,000 a year. £15,000,000 is roughly £1,000,000 a year if invested in the stock market. So you could hire 5 researchers to work full-time, in perpetuity.
Conferences are cool, but do you really think they generate as much research as 5 full time researchers would? As a researcher, I can tell you flat-out the answer is no. I could do much more with 5 people working for me than I could by going to even a thousand conferences.
Certainly makes sense to bring in PR resources, but let’s be clear that much worse ethical scandals than “One rich guy who donated to this turned out to have committed fraud” have frequently failed to destroy social movements. This will probably hurt EA, but is unlikely to permanently ruin its reputation unless it leads EAs to disaffiliate themselves from the movement to avoid negative PR—i.e. if most of its supporters are more concerned about optics and signaling than doing good.
The only way EA can be destroyed is if EAs, not non-EAs, abandon it.
The empirical track record is that the top 3 AI research labs (Anthropic, DeepMind, and OpenAI) were all started by people worried that AI would be unsafe, who then went on to design and implement a bunch of unsafe AIs.