All of NormanPerlmutter's Comments + Replies

Hi everybody, it looks like LessWrong made a new account for me in connection with this meetup, rather than using my existing account. Moderators, please put this event under my main account, NormanPerlmutter.

One point that is being glossed over in this essay is that teaching is a difficult skill that is not as strongly correlated with comprehensive expert knowledge of the content than one might think. I say this as someone who worked as a teacher for 6 years.

Part of the process of developing expertise in a field of study is "chunking." The expert mind sees lots of complex things together as a single chunk (which can be unpacked if necessary) whereas the beginner sees the individual pieces. This chunking helps experts to interact with other experts and to apply... (read more)

I just downloaded MS Edge so that I could use Bing AI and ask it to find me a Brazillian hammock more than 6 feet wide.  After repeated questioning, it kept giving me hammocks less than 6 feet wide (but more than 6 feet long). Even after I pointed out its error explicitly it kept making the same error and finally Bing gave up and told me it couldn't help. Like it would list two possibilities for me, state the length and width of each, and the width was less than 6 feet in each case.

Given all the amazing stuff we've seen out of AI lately, I'm kind of a... (read more)

I don't know, but I had a similar experience with chattGPT; after having heard from various sources that it is pretty decent at chess, I decided to challenge it to a friendly game. I was actually expecting to lose since I myself am not much of a chess player, and also having had conversations on topics where I am an expert, depending on your definition of that word, and being really impressed by the reasoning ability of the LLM in some of those conversations. Anyway, chattGPT insisted on playing as white. Then it said: "Ok, let's begin. You get the first move." So I had to explain to it that according to the rules of chess, it's white that moves first. So as usual it excused itself, and the game began. After each move, it printed an image of the table on the screen; I did not pay attention to that since I had my own physical board. (We were exchanging moves by stating the start square and the end square of each move, i.e. not the standard terminology for chess.) So a couple of moves in I noticed that the robot moved one of its Bishops right through one of its pawns. So I remarked that this is not a legal move. The robot agreed, and stated that it is not allowed to capture its own pawns. It made another move, it was my turn, I stated the moved but this time I also looked at the board that the robot displayed on the screen. I realized that something was very wrong, I had a knight where I should have a bishop. So I went back to trace where this error had started. But in the end it became to complicated and we decided to start a fresh game. This game also lasted a couple of moves before strange things started to happen. So we agreed to start fresh again, and now I insisted on playing as white (I was getting tired of having my physical board "upsidedown" compared to the board on screen). Anyway, the robot managed a couple of moves again, before it placed one of it's rooks outside of the board. Now I was tired so I just said "ok I give up, you're too good at this" (I don'

Since nobody has called it . . . I spotted the (intentional?) linguistic joke in one of the section headers. The Hebrew word that sounds like Llama means "why."

I've met humans who are unable to recognize blatant inconsistencies. Not quite to the same extent as Bing Chat, but still. Also, I'm pretty sure monkeys are unable to recognize blatant inconsistencies, and monkeys are intelligent.

1Jeff Rose1y
What makes monkeys intelligent in your view?  

I agree that this is a risk, but I'm not sure whether it's the main risk. Another risk is that if somebody gets access to the encrypted store, they can use it to steal all your passwords.

I read a lot of Derek Lowe early in the pandemic and regard him highly, but in this case I think he's wrong. Going through the comments of Lowe's post, I came across a link to this essay by a distinguished biologist, Stephen Salzberg, at Johns Hopkins agreeing with Zvi's perspective.

Salzberg is a computational biologist, not a virologist, but he's a distinguished professor at a prestigious school and does not seem to be on the fringe politically as far as I can tell &... (read more)

This is a fascinating essay that made me think of some of my personal experiences with having my boundaries violated in a new light. Thank you.

You pointed out that just asking for consent can be costly. I think an important social/communication/culture technology to consider is how to make consent requests less costly and/or less frequently necessary, while still allowing a strong social norm around consent.. For instance, having meta-discussions about consent with your friends or meta-rules about consent in your social group or community, that are organized in such a way that asking for consent is seen as easy. Giving close friends broad consent to a wide range of acts, and occasionally checking in on that over time. Etc.

I agree that living conditions are better today than several decades ago and worse today than 3 years ago.

That being said, I have seen a lot of mixed evidence and arguments about long covid and haven't figured out how to best think about it.

I'm confused by your use of "no longer" above. I don't believe we have ever lived in such a world., even before covid. We live in such a world to less of an extent now than we did before. Covid is one more thing that sometimes doesn't turn out ok in the end. But there are many other such thing, including many other diseases.

Compared to several decades ago, living conditions have improved. But the world is much, much more dangerous and harmful now than it was three years ago, e.g. potentially 50% of mild Covid cases permanently increase fatigue.

Could you provide more details on getting Paxlovid? My understanding was that it was only authorized for people with certain health conditions.

IIRC according to Zvi the health conditions are broad enough to encompass a significant chunk of the population, and the medication is currently way under-prescribed because people are not aware that they're eligible for it. EDIT: If I understand that website correctly (which I may very well not), being "at increased risk of severe illness" includes old age, pregnancy, or having medical conditions including common stuff like diabetes, being overweight, having a mental health condition, "physical inactivity" (?!), smoking or having a substance abuse disorder, etc.

I agree that the degree of air circulation within the terminal is an important factor. I'm not certain that the terminal is safer than the plane, but I think more likely than not the terminal is safer.

This link from my previous comment is not exactly a peer reviewed article, but it suggests that the difference in air replacement rate in a well-ventilated versus poorly-ventilated space (the terms they use for mixture of the air, not for air change rate) is only about a factor of 3. Of course, there are different degrees of poor ventilation.

I would be really interested to hear the perspective of somebody with greater expertise in the relevant engineering and physics.

2[comment deleted]2y

Airplanes do an excellent job circulating air, and are relatively safe places to be. Your risk in the terminal and the taxi greatly exceeds your risk on the plane.


I used to agree with this. But I recently realized it likely isn't true. Consider the following:

... (read more)
How much does the air near the ceiling of the terminal actually help? If the virus particles instantly circulated evenly, then they would definitely help. On the other hand, if there isn't much air current (as is the usual case in airport terminals) then the air near the ceiling is useless, since any virus particles would stay in the air near people's mouths and noses. The math might still work out in favor of the terminals, but I'm not convinced.

Regarding more people declining the second shot than the first shot, my best guess would be that people took the first shot, and either they themselves or one of their acquaintances had an extremely bad side effect, either actual or perceived, so they decided not to take the second shot. I know one person who followed this reasoning. Her husband fell ill after his vaccination with unclear causes, and she attributed it to the vaccine. She finally did decide to get her second shot and booster recently.

Another possibility could be that some financial incentives incentivized the first shot but not the second shot.

Well, the electric eraser was maybe a slight improvement over my manual eraser when I use both together, but not enough to really solve my problem. I went back to using mostly the manual one as it's more convenient.

I just found out that electric erasers are a thing. (Similar to an electric toothbrush, but an eraser.) I have ordered a high-end electric eraser, going to see whether it helps me to do better using my current paper and pencil setup.

Well, the electric eraser was maybe a slight improvement over my manual eraser when I use both together, but not enough to really solve my problem. I went back to using mostly the manual one as it's more convenient.

Anecdotal, but similar -- when I used to play in chess tournaments, I had a sense that I performed better and made fewer errors when I had more sleep, to the point of aiming for 9 or so hours of sleep the night before a tournament.

Cool idea, I like the historic and low-tech aspect. I will look into it.

I love blackboards, I was a research mathematician for many years and they have a special place in my heart along with a stick of Hagoromo chalk.  But they don't fit my purposes here for much the same reasons as dry erase boards -- they erase accidentally and don't allow for small writing.

I'm not familiar with them and am curious to learn more. My main concern would be whether they allow for fine-scale writing and erasing, since I am writing in small print with lines close together and erasing line by line. Is there a particular brand that you would recommend?

How is the older version better than the newer version (other than meedstrom's comment)?

Thanks. This is similar to what I'm looking for, but a bit too small. I'd prefer something the size of an 8.5x11 sheet of paper,  but I might give Remarkable a try. 

I just quickly browsed this post. Based on the overall topic, you might also be interested in these inconsistency results in infinitary utiliatarianism written by my PhD advisor (a set theorist) and his wife (a philosopher).

I'm curious to learn more about the thesis that caffeine or other stimulant use can completely mitigate the effects of sleep deprivation until 30+ hours without sleep. My own (subjective, anecdotal) experience with caffeine is that occasional (once or twice a week) caffeine use fairly effectively mitigates occasional sleep deprivation if I got say 5-6 hours of sleep the night before as opposed to my preferred 7-8, but is not too effective if I slept less than 4 hours the night before.  The more often I use caffeine, the less effective the caffeine bec... (read more)

What was your old job, and what is your current job?

"If we add all the percentage point increases (i.e. how many more percentage points serology positive participants experienced persistent symptoms vs serology negative participants - data from table 2) then we get 20.3%." 

I am not sure whether this reasoning is correct. It seems to be dependent on how the symptoms are categorized. For instance, suppose we divided fatigue into moderate fatigue and severe fatigue. The increased probability for each might be 5%, and then you would get 25.3% rather than 20.3%. Or suppose we combined fatigue and poor atten... (read more)

Yes, I agree that some symptoms are likely highly correlated. I didn't intend to rule out that possibility with that sentence - I was just trying to say how I did my math (although I'm not sure how clear I was!). The correct conclusion is in the following sentence: Possibly it would be better to add the caveat "0.2 persistent symptoms of those symptoms investigated".

It seems to me that the best argument against this is that there are less harmful ways to obtain an additional inoculation benefit, through additional vaccination. Either by getting additional shots of Pfizer/Moderna beyond the third shot, or by getting RadVac in addition to Pfizer and Moderna. I would imagine that there is some very large number of Pfizer/Moderna/RadVac vaccinations shots that would have comparable negative effects on health as getting Omicron once (maybe 10 or 100?), and that getting this many vaccination shots would provide much more pr... (read more)

Looks like Zvi just wrote a whole post in response to the update. In particular, January 19 was his prediction of a peak of reported cases, not of actual cases. 

This is true to an extent. Unvaccinated people are still able to attend. They just would need to forge their vaccination card. I think this is not particularly hard to do, though it's not trivially easy and many unvaccinated people would not do it for ethical reasons.

Thank you, good explanation. But see also my response to tivelen below. (the University of Washington team) released a new projection January 8, projecting that cases in the US (actual cases, not reported cases) peaked January 6. Had you seen this already when you wrote this post, and if not, does it impact your projection of a January 19 peak for the US?

(Edit: added hyperlink)

Looks like Zvi just wrote a whole post in response to the update. In particular, January 19 was his prediction of a peak of reported cases, not of actual cases. 

How is that different from what CraigMichael said? Attending that sort of event is a type of risk compensation.

I agree that attending large events is a type of risk compensation, and we may be referring to similar behavior patterns using different words here. But I'm trying to distinguish between these two types of infection: 1. Infections resulting from people going about their daily activities (e.g. getting exposed at work, in a store or restaurant, other small gatherings, etc.) Here, individuals might indeed change their behavior based on their own vaccination status and risk tolerance. But since Omicron is so widespread at this point, the probability that an infected person was vaccinated should be close to the base rate of vaccination among the overall population (although somewhat lower, since the vaccines still prevent some transmission of Omicron). In other words, P(vaccinated | type-1 infection) is a little less than P(vaccinated). 2. Infections resulting directly from attending large superspreader events where proof of vaccination was required. In this case, while P(vaccinated | type-2 infection) won't be exactly 1 due to the possibility of fake vaccine ID cards or weak enforcement of the policy at the event, I think it would still be quite close to 1.  If type-2 infections are a high enough percentage of overall infections, this could make it look like vaccinated people are more likely to get infected (which would be true at the population level!) even though getting vaccinated makes it less likely for any individual to get infected (assuming their behavior after vaccination remains the same). Apologies if much of this is obvious or redundant - I'm still trying to understand the gears behind this dynamic better myself. I agree there is likely a component coming from "vaccinated people take on more risk in general", but I hadn't considered that policies which only allow vaccinated people (to a first approximation) to attend large potentially-superspreading events could lead to increased transmission among the vaccinated relative to the unvaccinated, which coul
Suppose 50% of vaccinated people would attend this event, and so would 50% of unvaccinated people, after considering the risks (ergo, there is no risk compensation). However, only vaccinated people are allowed to go to the event. Then the vaccinated people could have increased rates of Covid compared to unvaccinated people because of being more likely to attend superspreader events, even though they did not increase their level of risk compared to the unvaccinated population. Whether this is the actual reason for the apparent negative effectiveness would depend on the actual percentages, and how common/dangerous superspreader events really are.

This is a very helpful analysis. I was independently undertaking a similar analysis, and it's nice to have this for comparison. I hadn't thought to exclude pedestrians, pedecyclists, and other non-occupants, nor of excluding single-vehicle crashes.

I think a some important pieces are missing from this analysis, as follows.

1) The final number, 548, is the number of miles that I must drive to accrue one micromort for all passenger vehicle occupants. But I am more interested in how many miles I have to drive to accrue one micromort for myself. The average (mea... (read more)

I have not used microcovid much because I am not confident in its predictions and modeling assumptions, or I don't feel they are clearly enough defined to make the tool useful. The change that would be valuable to me (which I have difficultly operationalizing) would be if Microcovid were improved such that I could be much more confident in its modeling assumptions and could use it without having to try to make lots of guesses about which scenarios are well modeled. Maybe it would be sufficient just to explain which types of assumptions make for robust mode... (read more)

Yes, I agree that this nonstandard definition is a crux for this disagreement. Good analysis.

Hmm, suppose an adult had urinary problems and wetted their bed regularly.  Which category would you say that fits into? Or somebody whose parents had named them something that they didn't like and they changed their name and didn't want others to know their original given name due to aesthetic preferences and social implications of character traits related to that name?

There would be some social harm in sharing this either of these, but would it necessarily be adversarial? Even if others were aligned with the person with the secret, they couldn't help but look at them a bit different knowing the secret.

2Alex Vermillion2y
You asking that question made me realize that I had mentally redefined "adversarial" underneath us! I feel "adversarial" is not really a good pointer to the concept I was using, which is what causes this confusion. I was reading it like it meant "referring to potential harm by person A onto person B", without any connotation of adversarial. I think that whether or not you accept this incredibly nonstandard definition is the deciding factor on this disagreement. That said, you were right! Thanks for calling me on that weird move, I genuinely would not have seen what I'd done without that last clarification.

Could you give an example of exploring un-endorsed emotional reactions? How is this related to having deeply held values?

In this case, sharing it with people who don't know her and will likely never encounter her will do minimal harm, so you might suggest that as an exception to the secret keeping.

Another such case is if sharing something would embarrass somebody. They might be embarrassed in spite of others not acting adversarial towards them. 

1Alex Vermillion2y
I think that example is either incorrect or part of a larger example class which may be weaker. It might be wrong if the embarrassment reflects fear of social harm. This would still be adversarial/harm-causing, so it is part of Dagon's framework. If that's not correct, I think the agent is simply acting irrationally and this is a larger class. An irrational agent can be mentally harmed by anything at all, so this class is much larger, but also a bit weaker to talk about (it probably best fits on a sliding scale as well).

I have been thinking about this topic a lot on my own and with friends before finding this post and was excited to see a post so related to my recent thoughts. One idea that came up in a recent discussion with a friend was that the pitfalls of the reasonable good faith effort in connection with common communication norms, especially if somebody reveals a secret accidentally and is feeling vulnerable and then asks you to keep it secret. In that case, if you say, "I'll make a good faith effort to keep it but I can't promise" it may be interpreted as "I don't... (read more)

Are you saying that you agree with William Eden's claim that vaccination does not substantially prevent the spread of covid? Or is that one of the things that you would "quibble" with him on? That point seems much larger than a quibble to me, it's a key point that's being debated currently about public health policy.  My own understanding is that while the vaccines are of course not perfect at preventing spread they do prevent spread to a significant degree, and therefore vaccination is indeed a public health issue not just an individual decision. 

I disagree with Will here.

The 97% was in case there was no other strain that comes in. So maybe the 96% is even taking into account another new strain? Or maybe it was just a typo.

There is an important practical consideration that is being left out here.  Attempting to completely devote one's life to these causes in the way described would not necessarily be effective. The best way to devote your life to a cause is not necessarily in a super-fervent way, because that's not how humans work. We need certain types of motivation, we have physical and emotional needs, we suffer from burnout if we work too hard. So if you believe that astronomical suffering risks are the most important issue to work on, then by all means work on them... (read more)

I was linking to a news article on CNBC that quoted the CDC director. in the third paragraph. I didn't take the time to track it back to the CDC directly.

For what it's worth, the CDC reported that side effects to the booster are less compared with the second dose, and that was also my personal experience.

It seems strange to link to a statement that's not by the CDC if you want to talk about what the CDC reported. In general I'd love to see data.
This would be really good  news for me, since my second shot effects were severe. I don't trust the CDC on this so went looking for papers and found this pre-print from Israel, that found booster symptoms were in the same ballpark as second shots. This release from the CDC found the same thing.

If I understand correctly, Zvi's idea is that vaccine protection against infection has likely gone down, but vaccine protection against severe infection has held nearly constant, so that the vast majority of additional infections among vaccinated people will be non-severe. 

I am skeptical of the claim that a substantially new risk profile is here to stay for the long term. The best reference case we have for this pandemic, I think, is the flu pandemic from 100 years ago. At that time we had no vaccines for the pandemic, and furthermore the flu mutates much more easily than covid. Nonetheless, the pandemic was pretty much over in two years or so.  Not because there was no flu left in the world, but because humans developed enough immunity to this especially virulent flu that it reduced back to the threat level fo the flus... (read more)

I have decided to get my booster of the Pfizer vaccine and am scheduled to get it on Monday, November 29, the soonest day that was acceptably convenient with regard to my concerns about side effects the next day.

Reasons for the decision were as follows. I'm fairly confident that I'm making the correct decision to get a booster very soon. I'm less confident as to whether I'm choosing the best booster, but I don't think that's likely to make a big difference.


Decision to get a booster very soon

The new Omicron variant increases my risk from covid over th... (read more)

Load More