By bias, I mean the framing effects described in this SlateStarCodex post.
Is there an accusation of violation of existing norms (by a specific person/organization) you see “The AI Timelines Scam” as making?
It’s unclear to me whether that post makes such an accusation.
How about focusing on the evidence, and on demonstrating good epistemics?
The styles encouraged by peer-review provide examples of how to minimize unnecessary accusations against individuals and accidental appearances of accusations against individuals (but peer-review includes too many other constraints to be the ideal norm).
Compare the paper When Will AI Exceed Human Performance? Evidence from AI Experts to The AI Timelines Scam. The former is more polite, and looks more epistemically trustworthy, when pointing out that experts give biased forecasts about AI timelines (more biased than I would have inferred from The AI Timelines Scam), but may err in the direction of being too subtle.
See also Bryan Caplan’s advice.
Raemon’s advice here doesn’t seem 100% right to me, but it seems pretty close. Accusing a specific person or organization of violating an existing norm seems like something that ought to be kept quite separate from arguments about what policies are good. But there are plenty of ways to point out patterns of bad behavior without accusing someone of violating an existing norm, and I’m unsure what rules should apply to those.
I liked most of this post a lot.
But the references to billions of dollars don’t feel quite right. The kind of trust that jessicata and Benquo seem to want sometimes happens in small (e.g. 10-person) companies, and almost always gets destroyed by politics well before the business grows to 1000 people. The patterns that I’ve seen in business seem better explained by the limits on how many peoples’ epistemics I can evaluate well, than they are by the amount of money involved.
LessWrong and the rationalist/EA movements seem to have grown large enough that I’d expect less trust than exists in a good 10-person group, based purely on the size.
I recommend starting with the original greenhouse effect forecasts that were made over a century ago (by someone who expected global warming to be desirable). That model still looks pretty good, except that CO2 emission forecasts of that time weren’t very good.
I wrote a long review with some comments on the extent to which I trust the book.
Metformin has a bunch of undesirable side effects. I don’t see an easy way to quantify the importance of those side effects, so I try to evaluate how likely it is that the benefits of metformin will apply to me.
The obvious way in which metformin might cure or prevent age-related diseases is by curing insulin resistance. Some educated-sounding people have been saying that insulin resistance contributes significantly to Western disease (aka diseases of civilization, including cardiovascular disease, diabetes, dementia, and some subset of cancer). Doctors have arguably been undertreating insulin resistance, because it doesn’t produce immediate disease-like symptoms. Maybe if metformin were still patented, the patentholder would be pushing the FDA to get insulin resistance classified as a disease.
Insulin resistance seems to be non-existent in cultures that most resemble our pre-farming ancestors, so it sure looks like it’s avoidable via lifestyle changes (the best guesses are diet (high fiber, whole foods), exercise, and sleep). That suggests it’s possible, although maybe hard, to get the benefits attributed to metformin without the side effects.
I pay close attention to my insulin resistance, via blood tests, and have avoided metformin for now because it looks like my lifestyle is good enough that I have few insulin-related risks so far. If my A1C gets above 5.6, I’ll think pretty carefully about getting metformin.
Daily supplements:
10mg pregnenolone: sometimes a bit more, aiming for blood levels around 60-80 ng/dL. Seems to increase my mental energy.
25mg dhea: I’m uncertain about the dose, and I’m experimenting a bit to see if I can detect effects.
~3000 IU vitamin D: aiming for blood levels just over 50 ng/mL.
100mcg vitamin K2 mk7
1mg folate: I’m homozygous for MTHFR C677T; it seems to improve my mood.
1mg B12: it’s keeping my blood levels around 800 pg/mL, compared to < 500 before supplementing.
1g omega-3 (currently in the form of Bulletproof Omega Krill Complex).
Stem Cell 100, once per day
500mg N-Acetyl-L-Cysteine: possibly helps lower my homocysteine, but my results are pretty noisy.
ashwagandha: I had very interesting thyroid effects when I was getting too much iodine from kelp, but now that I’ve fixed my iodine levels, I don’t detect any effects.
ResveraCel (Thorne), contains Nicotinamide Riboside, Quercetin Phytosome, Trans-Resveratrol, Trimethylglycine.
curcumin (Life Extension Bio-curcumin): I suspect I have inflammation due to a temporary problem, and will likely go back to taking this less regularly in a few months.
caffeine: from tea, green tea extract pills, or occasionally a 100mg pill.
Supplements taken less than once per day:
creatine, 1 or 2 grams (no clear effects).
melatonin, typically 300mcg, timed release version.
liver, and organ meat pills, from Ancestral Supplements.
green mussel
kratom: typically 0.25 teaspoon of Maeng Da, mildly stimulating with reduced anxiety?
phenibut: taken about once a month, for the anti-anxiety effect, with a bit of stimulation.
A majority of these choices are influenced by Bredesen’s book The End of Alzheimers, or by a prior source with similar advice.
Quinn: “Hold it right there. Regardless of whether that’s true, it’s bad to say that.”
Carter: “That’s an appeal to consequences, well-known to be a logical fallacy.”
The link in Carter’s statement leads to a page that clearly contradicts Carter’s claim:
In logic, appeal to consequences refers only to arguments that assert a conclusion’s truth value (true or false) without regard to the formal preservation of the truth from the premises; appeal to consequences does not refer to arguments that address a premise’s consequential desirability (good or bad, or right or wrong) instead of its truth value.
Meditation is action, in some important sense, and mostly can’t be demonstrated.
It is hard to reliably distinguish between the results of peer pressure and actual learning. I think CFAR’s best reply to this has been it’s refund policy: last I knew they offered full refunds to anyone who requested it within one year (although I can’t find any online mention of their current policy).
The idea that CFAR would be superfluous is fairly close to the kind of harm that CFAR worries about. (You might have been right to believe that it would have been superfluous in 2012, but CFAR has changed since then in ways that it hasn’t managed to make very legible.)
I think meditation provides the best example for illustrating the harm. It’s fairly easy to confuse simple meditation instructions (e.g. focus on your breath, sit still with a straight spine) with the most important features of meditation. It’s fairly easy to underestimate the additional goals of meditation, because they’re hard to observe and don’t fit well with more widely accepted worldviews.
My experience suggests that getting value out of meditation is heavily dependent on a feeling (mostly at a system 1 level) that I’m trying something new, and there were times when I wasn’t able to learn from meditation, because I mistakenly thought that focusing on my breath was a much more central part of meditation than it actually is.
The times when I got more value out of meditation were times when I tried new variations on the instructions, or new environments (e.g. on a meditation retreat). I can’t see any signs that the new instructions or new environment were inherently better at teaching meditation. It seems to have been mostly that any source of novelty about the meditation makes me more alert to learning from it.
My understanding is that CFAR is largely concerned that participants will mistakenly believe that they’ve already learned something that CFAR is teaching, and that will sometimes be half-true—participants may know it at a system 2 level, when CFAR is trying to teach other parts of their minds that still reject it.
I think I experienced that a bit, due to having experience with half-baked versions of early CFAR before I took a well-designed version of their workshop. E.g. different parts of my mind have different attitudes to acknowledging my actual motivations when they’re less virtuous than the motivations that my system 2 endorses. I understood that pretty well at some level before CFAR existed, yet there are still important parts of my mind that cling to self-deceptive beliefs about my motives.
CFAR likely can’t teach a class that’s explicitly aimed at that without having lots of participants feel defensive about their motives, in a way that makes them less open to learning. So they approach it via instruction that is partly focused on teaching other things that look more mundane and practical. Those other things often felt familiar enough to me that I reacted by saying: I’ll relax now and conserve my mental energy for some future part of the curriculum that’s more novel. That might have led me to do the equivalent of what I did when I was meditating the same way repeatedly without learning anything new. How can I tell whether that caused me to miss something important?
You use math as an example, but that’s highly focused on System 2 learning. That suggests that you have false assumptions about what CFAR is trying to teach.
There are many subjects where written instructions are much less valuable than instruction that includes direct practice: circling, karate, meditation, dancing, etc. Most of those analogies are fairly imperfect, and some have partially useful written instructions (in the case of meditation, the written version might have lagged in-person instruction by many centuries). Circling is the example that I’d consider most apt, but it won’t mean much to people who haven’t taken a good circling workshop.
A different analogy, which more emphasizes the costs of false assumptions: people often imagine that economics teaches something like how to run a good business or how to predict the stock market, because there isn’t any slot in their worldview for what a good economics course actually teaches. There are plenty of mediocre executive summaries of economics, which fail to convey to most people that economics requires a pervasive worldview shift (integrating utilitarianism, empiricism about preferences, and some counterintuitive empirical patterns).
The CFAR handbook is more like the syllabus for an economics course than it is like an economics textbook, and a syllabus is useless (possibly harmful) for teaching economics to people who have bad assumptions about what kind of questions economics answers. (This analogy is imperfect because economics textbooks have been written, unlike a CFAR textbook.)
Maybe CFAR is making a mistake, but it appears that the people who seem most confident about that usually seem to be confused about what it is that CFAR is trying to teach.
Reading the sequences, or reading about the reversal test, are unlikely to have much relevance to what CFAR teaches. Just be careful not to imagine that they’re good examples of what CFAR is about.
I don’t like the way this post hints that schism and splitting are the same thing.
E.g. the split between CFAR and MIRI, and the creation of multiple other AI risk organizations (FHI, FLI, BERI, etc), fit the archipelago model, without being schism-like.
Hostility begets hostility, but agreeing to specialize in different things doesn’t have the same tendency to beget further specialization.
High short-term interest rates constitute relatively good evidence of a coming recession. The hard part is deciding what qualifies as high. “Higher than long-term rates” seems to get attention partly due to its ability to produce a clear threshold, rather than to anything optimal about the signal that’s produced by long-term rates.
Note that if recessions were easy to predict, they would be easier to avoid than they have been.
Here are some, based mainly on a search for which ones I’ve linked to before:
Distinguish Info, Analysis, Belief, Action
Two Types of People (foragers and farmers)
And, since I think he published better writings outside of his blog, here’s a longer list of those:
Shall We Vote on Values, But Bet on Beliefs?
How YOU Do Not Tell the Truth: Academic Disagreement as Self-Deception
He Who Pays The Piper Must Know The Tune
Burning the Cosmic Commons: Evolutionary Strategies for Interstellar Colonization
Long-Term Growth As A Sequence of Exponential Modes
Economic Growth Given Machine Intelligence
Fear of Death and Muddled Thinking—It Is So Much Worse Than You Think
When Do Extraordinary Claims Give Extraordinary Evidence?
WHY HEALTH IS NOT SPECIAL: ERRORS IN EVOLVED BIOETHICS INTUITIONS
I’ll suggest the main change is open threads get fewer comments because the system makes open threads less conspicuous, and provides alternatives, such as questions, that are decent substitutes for comments.
Note that the rationality community used to have the Singularity Summit, which was fairly similar to EAGlobal in its attitude toward high-status speakers.
The risk/return curve hides some important problems.
Many investors have more income when the stock market is high than when it is low.
For example, donations to MIRI are likely heavily influenced by the price of tech stocks and by cryptocurrency prices. MIRI needs to pay salaries that fluctuate much less than those prices. If MIRI invests in assets that have much correlation with those prices, MIRI will likely end up buying disproportionate amounts of those assets within a year after they peak, and buying less (sometimes even selling) within a year after they reach a low.
Maybe that works well for short-term market fluctuations, but whenever there are significant market fluctuations that last for several years, that would mean that MIRI would be buying mainly at the worst time. I expect that would cause them to underperform the relevant benchmark by at least a few percentage points.
So I expect MIRI will be better off minimizing the volatility of their investments.
Note that this conclusion holds even if MIRI is risk-neutral.
I expect that a fair number of individual investors have income fluctuations that imply the same conclusion.
They have a good argument that the Baumol effect explains a significant fraction of cost disease.
Much of their point is that it’s not surprising that some industries have better productivity growth than others.
But they don’t have a full explanation of the extent to which productivity growth varies by industry. Why do medical productivity trends appear much worse than auto repair productivity trends? What caused the pharmaceutical industry to become less productive?
Their partial answer seems to be Hansonian:
A key distinction is when labor is an input into the good or service and when labor, in essence, is the service. Consumers do not care how much labor is used as an input into the production of a car, but they do care how much labor input is used in a massage, artistic performance, or doctor’s visit.
That’s pointing in the right general direction, but I expect a full explanation would also involve consumers being unwilling or unable to switch to companies that offer cheaper services.
a decentralized internet versus an internet under the central control of something like AOL.
Bitcoin energy usage.
electrical systems that provide plugs, voltages, and frequencies which are incompatible between countries.