notafraudster

This tool is really cool!

But I don't really understand why the text and framing is written is written as if it was supernatural ("esoteric", "mystical", "witchcraft", "magical", "otherworldly", "sorcerer", "summoning", "the power to shape and mold lies in your hands", "the choice is yours", "magic", "uncharted territories", "let not your journey end here", "mystical journey", "arcane knowledge of the ancients"). The first few times it felt a little bit cheeky but by the end it felt like a TikTok about manifesting wealth through your horoscope.

Paradoxically, the one thing I wanted to know that would have been an opportunity for flavour -- how it "def[ies] conventional color science" -- isn't actually clarified in the text.

I clicked on your links but I don't really understand what it is I'm meant to be seeing. I haven't read a balance sheet since college, but things look pretty normal to me? Is the objection that about 2% of their expenses are on Grants and donations? Is it that they make revenue through a VPN subscription they offer? Is it that their taxes show spending about $14 million (a drop on the bucket) on policy programming (mostly it seems about building the Common Voice dataset and funding external data scientists to work on data stewardship policy stuff; these feel, I think, like what you're suggesting that it's fine for Brave to do because they're "technical problems"?)

I read -- skimmed, but I think that's fair -- through a dozen pages to try to understand your point and nothing jumped out at me. There's a part where they talk about how they made a few thousand dollars from credit card rewards points. I did a ctrl+f for "woke" and "transgender" and "critical race theory" and "covid vaccination" and "cico versus macro-based dieting strategies" and didn't find anything. It's a little frustrating because I spent more time trying to figure out your point than you spent to make it.

Humans are generally viewed to have an inherent right -- or at least a strong imperative -- to fully participate in society. Restricting a human's right or privilege to do something typically carries a higher burden than simply asking whether the thing they are doing is a net positive/negative. Over the last 100 years the layout of cities and countries reflects an expectation that full participation in society requires access to transportation. While driving is not an absolute right and we subject drivers to certain minimum requirements (licensing and sometimes periodic relicensing, insurance, various laws that apply to conduct in a motor vehicle), the presumption is basically to create a pathway that allows people to drive.

Driving creates a danger and a number of costly externalities. Those are the costs side of the cost-benefit equation. It may be the case that AI drivers have generally lower costs because of your assertion that they make fewer mistakes or less grave mistakes than humans. This does not lead to the obvious inference that allowing an AI driver today in a particular place under a particular regime of rules confers a benefit equal to allowing a human to drive.

In the same way, for instance, we do not generally pass laws or policies against people having children, even if we recognize that some people are suboptimal or unfit parents. Instead we presume fitness and build in some checks to capture excessively unfit parenting later. Whatever the cost of suboptimal parenting, we recognize the choice or ability to have a child to be a part of basic human dignity, and so the benefits outweigh the costs. Many jurisdictions place more onerous requirements on pet ownership than human ownership even if the downside cost of being a negligent pet owner is less bad than being a negligent parent.

Perhaps the calculus will be different when AI driving is costless and ubiquitous; we might decide collectively that humans have no inherent right to manually operate vehicles and full participation in society does not require them to. But in the mean time, holding AI driver street tests to a higher standard than human drivers can be justified strictly on a cost-benefit (and thus pragmatic) basis.

You might counter that, well, AI driving tests don't provide a benefit now, but given a particular utility function, the testing offers <x> marginal training value towards a future reduction in costs (in terms of injury/delay/death). But then it depends on your discount factor of the present versus the future, which is a socially determined function.

There's also an underlying presumption that there's some elasticity of resource allocation. Perhaps allocating $x million towards AI driving reduces deaths or injuries or delays; but people can readily contest that $x million spent on other options could do so more efficiently (perhaps infrastructural investment in public transit, perhaps additional free driver's training, perhaps more robust emergency services, perhaps safety features). Because much of this investment is occurring in the private sector, people rely on the state's regulation of these efforts to incentivize allocating resources towards preferred investments. This is also pragmatic, rather than puritanical.

It's fine to disagree or have a more bullish view of AI driving, or to feel these concerns are overrated, but it makes very little sense to characterize them as puritanical (in contrast to pragmatic). Some people may have puritanical views, just as some proponents of the technology may have messianic views. But both positions can be justified entirely from rational argument depending on one's utility function. Why strawman?

You are a really uncharitable poster here. Your first intervention is to dismiss anyone who disagree with you about which metrics are most useful as "bullshit" and "emotional" and here you call them puritans. I think you should engage in some self-reflection and try to approach conversation online in a less antisocial way.

If a person stops in the middle of a road, other people talk to that person, and typically learn what's going on and can help work on resolving it. Are they having a heart attack? Do they need people to push the car? Are they having a road rage incident? Are they a jerk who needs to run into a store and picked the wrong place to stop temporarily to do it? Even if the situation can't be solved, the ability to communicate (or the implication of not being able to communicate) gives predictable, human-friendly feedback which can be shared with emergency services and other people.

(See bio for my background) All of the replies you've gotten so far are very good and I upvoted all of them.

In particular:

cuchoi acknowledges the publication bias version of the risk here. Let's say your average effect is 1 unit with a confidence interval that is 0.9 units at a desired level of confidence. We can interpret this confident interval two different ways: one is that assuming 1 unit is the true effect, then repeated sampling would produce a sampling distribution of estimate effects that span 0.1 to 1.9 at the desired level of confidence. Another is assuming that, say, 0.1 was the true effect, effects as large as the one we see (1 unit) would occur a non-trivial portion of the time. Now, imagine many researchers do this experiment and the true effect is 0.1. Some researchers find negative effects, some find small effects that are not significant, others do larger studies and find small effects that are significant, others find larger effects. Now, imagine the journal will only publish effects that are both statistically significant and substantively interesting. The only person that submits for publishing is the version of the study that finds the large effect (1 unit). cuchoi is very correct to suggest that when your design can only find large effects, the published effect will likely be overestimated.

fpoling and sandgiant highlight the sensitivity risk argument. Suppose that the outcome is heavily sensitive to some confounders (socioeconomic status, nutrition, smoking status, race, etc.) And suppose poor people are slightly more likely to get treatment, just from coin flip chance. Because poverty correlates with both the effect and the probability of being treated (even though you tried to assign randomly), some of the visible effect is the relationship between poverty and treatment, not effect and outcome. There are designs other than simple randomization that try to explicitly deal with known confounders, but they can't deal with unknown confounders. Larger sample sizes mitigate the risk of imbalance of both known and unknown confounders.

Everyone is doing great!

It's a win _if_ the barrier to getting sedentary people to improve their health through fitness is time cost and not all of the other costs associated with habit formation. My guess is for most sedentary people the biggest thing is that they manage to talk themselves out of doing anything, or once they start, talk themselves into allowing cheating/skipping rather than drilling in the habit.

(Not that this is bad news regardless, it's good to have tools in the toolbox and some people might prefer lower time higher intensity stuff, but I do think the result is pretty limited)

I had heard of this service before and assumed it costed money, but thanks to your comment checked it out, and apparently they have a free tier allowing you to create 10 cards per month. Cool!

Yeah. My father didn't do secure work, but he was an immigrant to the country where I was born. As part of the immigration process, he was asked whether he had ever been arrested (maybe it was charged, I don't recall and he's long gone so I can't ask). He said "no". In fact, he had been arrested as a pre-teen for stealing an idling tractor and joyriding in rural post-war England. Immigration authorities don't care that a teen went for a joyride on a tractor. They care that he lied. Took a bunch of lawyering and paperwork to resolve the issue.

I guess fundamentally the distinction is "if they can catch you, tell the truth, and if they can't, make yourself look good", but I guess it can be hard to know.

I think the grandparent is under the impression that the alleged Twitter version of the post had a non-anonymous author and that the Github blog instead chose to link to a Tumblr blog, where the blog author claims they learned the proof from an anonymous number theory professor, thus laundering out the authorship.

But I suspect the alleged Twitter version was just a link to the Tumblr, for which it's clearly not necessary to attribute the retweeter (especially in the context of the post being "the proof is actually incomplete, not very good, and less beautiful/more complex than the canonical proof").

Wikipedia tells me there are in fact a hair under 19 million Bitcoin minted. I am willing to believe this site is getting the large exchange component right, or at least conservatively right (conservative in that it's unlikely to be claiming coins that don't exist, but it's possible exchanges have additional coins not listed, and it's possible there are other exchanges not listed).

Googling, it seems that the best estimates seem to indicate between 2 and 4 million Bitcoin are permanently lost (lost keys, known hardware failures of failed wallets, unclear if this counts Satoshi's million bitcoins or not). This estimate [0] by an analysis firm who appears to be broadly pro-crypto suggests 3 million lost, and that ~14.5 million BTC are "illiquid". Illiquid obviously doesn't mean lost

Illiquid doesn't mean lost, but it suggests something to me about the current conversation. If only ~4.5 million BTC are even a little bit liquid, then it's fair to say that at least half the engagement with BTC in balance terms is on exchanges, and probably more considering velocity of money. Not the grandparent's claim of 99%, but also not your claim of 10%. At the very least, not an obviously settled question.

[0] https://insights.glassnode.com/bitcoin-liquid-supply/

I'm sorry you lost your wife. My wife had a near-death medical experience last year (a stroke, completely unrelated) and I was never as emotionally empty and destroyed as I was during the ten minutes I thought I lost her and subsequent days worrying about her recovery.

You linked an interview with Childrens' Health Defense, an organization whose primary purpose is to argue, categorically, that all vaccinations are dangerous and do not work. They are the largest such organization and their activism has been directly responsible for the growth in resistance to MMR vaccination and later to COVID-19 vaccination.

This does not mean that the point about a class of antibiotics is incorrect, and I understand that there's a possibility of "strange bedfellows" here because of the potential agreement on the matter of, uh, "big pharma". As I see it, one possibility is that you linked the CHD article not knowing this. Another possibility is that you linked the CHD article knowing this, but figuring just because they're wrong about vaccines doesn't tarnish them on this issue. Another possibility is that you generally agree with their position on vaccines.

I'm not here to judge or change your mind, I'm just telling you that I don't have any prior view about whatever fluoroquinolone antibiotics are. I read the first half of your post wanting to do more investigation. Because of the CHD link inclusion, I am now predisposed not to believe this is a real issue. I am a non-medical social scientist but I have spent a lot of the COVID period publishing work about vaccination. Anti-vaxx stuff is a complete deal-breaker for me. And if your reason for posting was to persuade an audience, I think that's the opposite of what you want.

Infinitesimals exist precisely to create an reciprocal quantity to infinities.

Infinitely many finitely short time intervals cannot have passed after finite time, but infinitely many infinitesimally short time intervals can. In the same way that one way of conceiving infinity is to imagine it as an arbitrarily large value at any given expansion, but infinity in the limit, an infinitesimal is an arbitrarily small value at any given expansion, and zero in the limit.

That's the whole point. The atomic distance of a subdivided step goes to infinitesimal at the same rate that the time associated with the distance does, so we're fine!

I agree, but isn't it necessary to reckon with the fact that historically, explicitly and intentionally racist policies (for example, Jim Crow voting laws) often laundered themselves, when convenient, through classism in order to escape scrutiny? Put differently, would you contend that poll taxes were not racist?

Also, wouldn't we expect that organizations that are explicitly anti-racist would conclude that because classist policies or laws have racially disparate impacts, addressing them still increases racial equity even if they are ostensibly motivated by race-neutral class concerns?

I don't mean to be rude, but I suspect by "MLK's values" you mean the sentence "I have a dream that my four little children will one day live in a nation where they will not be judged by the color of their skin but by the content of their character." as interpreted to mean support for racial neutrality and opposition to the use of racial categories in programmatic decision making.

It's fine for that to be the thing you believe in, but that is not a reasonable interpretation of "MLK's values". MLK scholars, his contemporaries he worked with, and his successors within his organizations would nearly categorically reject your characterization of that being his values. Like this is an entire area of academic study: what MLK said, thought, and wanted done.

Again, it's fine to conclude that MLK has it wrong and the left is wrong about race and social justice warriors etc. etc. etc. I don't think you should be made to care about what MLK thought or to study him. I am not interested in telling you you are right or wrong about any policy issue. It's just incredibly flip to engage so superficially with a guy who left such a voluminous archive of writing and thought and use him as a prop in the argument. I think if you are going to assert yourself as understanding what his values were, you owe it to him to spend a lot more time and effort trying to understand what they were.

This is sort of the equivalent of thinking you understand Nietzsche to be anti-religious because you half-remember the quote "God is dead" and although you've never read him you feel confident you know what it means.

I feel like I've never been able to get a handle on cable management despite using velcro cable ties, and baseboard cable clips, and other stuff. I just have never found the right thing. So I'd love if you linked some of the ones you used (granted they're probably all random drop-shipping Amazon brands)!

I think I left enough context clues in my post to understand that by "serverless" and "no server, thick client", I meant "the client also handles the server function" or "closing the client suspends operation of the server". But in case I didn't, just to clarify, when I said "serverless", I meant Kodi and things like Kodi, and when I said "server", I meant Plex and things like Plex, including Emby and Jellyfin.

I used Kodi for many years (consistently since about 2008 and periodically before then) and stopped last year, largely because the no server, thick client architecture is less useful than it used to be. At this point almost everything I want to play video on has hardware accelerated H264 and H265 decoding, and so almost nothing I play involves transcoding. I ended up moving to Plex, not out of any particular love for Plex, just a basic idea that the value of having a server running exceeded the cost.

With Kodi I was relying on a bodge of extensions to sync watch status across clients, was maintaining libraries on clients separately (I know Kodi supports remote MySQL-managed libraries, but if I was going to run a MySQL server on my storage and run some kind of background process to handle library updating, why not run a server like Plex?)... stuff took forever to sync, performance for library management was not great. Even for browsing, I found Kodi not very performant for large libraries.

Switching away had some other benefits besides library management. Although Plex has more restrictive UI options, the default UI is much more wife/mom/friend approved. My wife had a stint in a hospital and we could watch films synced, together, using Plex (which also handled all the routing/port forwarding/whatever internally) The Sub-Zero plugin for Plex (now discontinued, but still operational) let me fill out subtitles for my content. Plex Media Manager let me integrate a bunch of external stuff into my library including syncing Letterboxd lists into my library. Mobile support is first class (including a somewhat janky download feature that's great for flying)

After switching away I kept an eye on Kodi, and the major thing I noticed is that development cadence is pretty slow at this point compared to Plex. Now, this is because Plex is VC backed and charges money for certain features, and dispositionally that annoys me a little bit. But it also does contribute to wider platform support and faster feature iteration.

A lot about Plex sucks. They piss away tons of time and effort onto VC-friendly features that have nothing to do with what Plex is. They're currently pursuing some nonsense where you use Plex to search for things available on Netflix and watch it there. They also launched their own free, ad supported streaming service which is annoyingly intrusive. They are stubborn about some design stuff that the entire userbase dislikes. They push incredibly awful changes and then act shocked when they get hundreds of comments from paying customers complaining about how bad the changes are -- recently it was a loud and annoying noise playing on startup. They reverted it after a day, but just generally I get the feeling Plex as a team either doesn't know what their audience wants or deliberately ignores it for VC reason. Obviously I'd be happier if it was free and open source.

Personally, I could easily see switching away from Plex to Jellyfin or Enby, but I can't imagine switching back to a serverless media setup.

I don't mind reboots, since I sleep at least once a day. The main issue is the inability to restore state. MacOS has reopened everything I had open before a reboot for years now and I'm not sure why Windows still struggles with this.

I'm not the author. I think you've hit on the main limitations, but I don't think it's fatal. If the lay of the land in a given field is such that there are well defined camps, then you allocate the money to each camp[0] and allow the camps to allocate the money internally. If the lay of the land is that it's diffuse and has very shifting structure, then you can't follow this strategy, but also it's maybe worth raising the possibility that it's not yet time to allocate money and the first thing the field should do is try to develop some structure.

[0] Obviously the problem of what %age to give to each camp is an important second-order problem, but if you imagine a scenario like a university, I find it more likely the Provost's office can figure out how much to give to which Dean than they to be able to figure out how much to give to which lab.

(I also think you're right to observe that even within a particular camp of a particular field there will be relationship-driven gladhanding that helps determine the flow of money, and that's a real concern, but it's just a more localized concern than I think you give it credit for. I'm also not entirely sold on it.)

If you are not sure if your screening processes work, do an RCT. I have zero idea why the article raises this clearly provocative and correct approach and dismisses it immediately. Allocate a percentage of the positions you're hiring for to candidates that are hired at random. The empirical decision about what percentage depends on your primary concern: if you think your hiring process is completely useless, then you want to get this percentage close to 50% because this is what's optimal, experimentally. If you think your hiring process works OK and want to guard against downside risk on the random hires, get the percentage lower. This will cost you statistical power, but less than you expect.

No one is saying "random" means "no standards at all". If you're a graduate pure math department you don't need to admit a grad student whose background is that they have a GED and worked as a welder for 5 years before getting a 135 on the GRE Math side. Instead, split your screening into two phases: one phase that is targeted to exclude people who absolutely positively do not meet requirements and could never work out, and one phase designed to pick the best feasible candidates among those who meet requirements. Do the RCT in the latter phase.

If you have relatively continuous or regular hiring processes, you'll be well powered to assess your hiring screen in pretty short order. The larger your organization or the faster your overall turnover is the more robust you are to the risk of this going wrong. Not every company can do this -- academic departments that hire one person a year and sometimes have blown searches or people going back on the market within a year or two are clearly going to be a little risk averse. But if we're talking Deloitte hiring a cohort of 150 office admins, they can absolutely, positively make use of this strategy.

The other thing I'd add is that a lot of people are under the mistaken belief that hiring is about determining the best candidates. In fact that is not true and it has never been true. It's about maximizing the probability that a candidate you offer a job to takes it and is good... or else maximizing the frontier of some function of probability the candidate accepts, is good, and how good they are. Hiring nothing but the best candidates will increase the likelihood of a blown search, increase the rate of attrition, etc. This makes almost all of the discussion around mErIt In HiRiNg a distraction from how hiring actually works.

There are more narrowly tailored options. The more local you make the randomization, the less power and the less downside risk. Suppose you rank candidates from 1 (terrible) to 100 (ideal) and tend to hire candidates in the 80-90 range. Even a willingness to allocate some positions to candidates scored at 70-80 is going to give you lots information about whether your ranking system is useful at all.

Do more RCTs!!! Please!

Waterloo has a Department of Combinatorics and Optimization? Amazing.

One of the many reasons why "a vote" isn't a best option is that by "a vote", he means a poll he posts on his personal Twitter account and which is brigaded by lunatic Elon people. Moreover, Twitter laid off at least 80% of the experiments team last I checked (maybe 100% now?) and most of the other related research science people, so I highly doubt they even have a mechanism to do sample to population extrapolation or design a sample frame even if they wanted to. Completely embarrassing.

Your comment is grey but I actually thought -- even if inflammatory -- it articulates something real.

I have noticed this kind of thing. I've done therapy on and off (sometimes reactively, like after my Dad died; sometimes a little more pro-actively, like when I had a brutal case of writer's block during my dissertation prospectus). Never anything I would rate as very serious or urgent. I would say the therapy was not un-useful, in the sense that I didn't mind taking time out of my day to talk to someone. But I also basically felt like the eventual remediation mostly just stemmed from the acute phase passing or steps I took outside therapy. The thing that got me over my writer's block was mostly writing. The thing that helped me deal with my Dad dying was mostly more time passing and my life moving on.

Meanwhile, I have noticed that a lot of people in my life with higher distress loads, and particularly people like the author of this piece who report suffering from fairly serious mental illness, invest a lot of resources into self-help. I think that's a good instinct, to say "I have a problem and this is a tool that will help fix it", analogous to seeking medical care for a physical ailment. But my experience is that the vast majority of those people seem to make little apparent progress via self-help, and instead typically continue to report being extremely distressed (though, punctuated with confessionals about how much better things are). For most people levels of distress in life seem to me to be primarily associated with lifelong, or early socialized dispositions; or with major material facts of life (i.e. single/married, broke/affluent, healthy/sick, undergoing a major life change/being severely stuck/doing fine) rather than the particular steps they take to manage or mitigate it. That doesn't mean therapy doesn't work, because maybe the counterfactual is they would have gotten much worse absent therapy.

A few friends from when I was a teenager, people who really struggled with severely gnarly upbringings, abuse, broken homes, early life drug dependency, etc. even went into becoming therapists or social workers. And the strange thing is, it externally seems to me like they're about as troubled as they were 20+ years ago, modulo having a little more money.

I guess I would distinguish a lot of these cases from people who are acutely suicidal, people who are actively seeking drug rehabilitation, etc. There seems to be at least some evidence for intervention in those cases. But I am talking about the kind of cases I think you're talking about.

So to your broader point, despite again the fact that it's expressed a bit stridently, I think there's something there: the evangelism about self-help's benefits often seems to outstrip self-help's apparent benefits. There's probably some element where part of therapy is basically tricking your brain, so the apparent sentiment of someone who is clearly having a hard time beaming about how much better they're doing might be part of the plan anyway. But externally it can feel jarring to hear "I'm really turning a corner!!!" from someone who's been struggling for years and continues to struggle.

It's any valid question at all, it needn't be yes or no. Humans have, at minimum, red, black, brown, blonde, dyed, grey, white, and no hair. Learning what colour hair a person has eliminates the other categories.

The way to think about the information content of a problem or of something you learn is exactly what you're suggesting. If you numbered every living person on earth, it'd take more than 32 bits and not quite fill the 33rd bit.

If you then learn a person's gender, you can eliminate all the people with the incorrect gender, which is going to leave you either 31.x bits (assuming binary gender) or 25-27 bits of remaining entropy (assuming some non-binary gender and, say, a 1-3% incidence rate).

When the parent you're responding to says you get 3 bits for knowing someone plays chess, they're guessing that 1/(2^3) = 1/8 of people, in an undifferentiated sense, play chess. Of course if we knew someone's age or gender or country of origin, the conditional information value in knowing they play chess could be greater or lesser. And realistically no one is ever trying to identify a human among all humans (partially because it seems highly unlikely that there are many questions that could equally implicate the president of the United States and a six year old on the Marshall Islands in their answer). Each bit of information represents a halving of the entropy of the target surface.

I think you got to within 1 bit of the answer from first principles ;)

I asked three clearly stated, graduate level questions in my field and all three answers were useless.

> what is the relationship between commodity prices and civil war? answer in a way that cites published scholarship.

> The relationship between commodity prices and civil war is complex. Research has found that a negative, one-standard-deviation commodity price shock in a single year can increase the risk of civil war outbreak in Sub-Saharan Africa by between 40% and 70% of the baseline risk of civil war outbreak. This holds true for all large countries in Africa, the Middle East, Latin America, and Asia. Additionally, there is evidence to suggest that a rise in international commodity prices may also lead to an increase in civil wars. For example, research published in the May 2010 issue of the Economic Journal found that a 10% increase in the price of international commodities was associated with a 22% increase in the probability of civil war outbreak in African countries that are big exporters. Furthermore, research from Oeindrila Dube and Juan F. Vargas of Universidad del Rosario concluded that “price shocks generate contradictory pressures on conflict.

This definitely synthesized information, but it's nonsense. The beginning of the thesis is that a negative price shock increases civil war risk. Then it suggests a positive price shock increases civil war risk. Then it concludes that price shocks generate contradictory pressures on conflict. Dube and Vargas is a famous paper, so it was correct to cite it.

If the answer is that research is mixed, specify that. If the answer is that the research is unambiguous but it depends on the location or commodity type, the answer should disambiguate.

> are ideologically extreme candidates punished in u.s. house of representatives elections? what about primary elections?

> No, ideologically extreme candidates are not punished in U.S. House of Representatives elections. In fact, there is evidence to suggest that extremism may actually be rewarded in some cases. For example, the landslide defeats of Barry Goldwater and George McGovern in presidential politics have been seen as evidence that American electorates punish extremism. However, this view is thinly supported by systematic evidence. Additionally, there will be at least one special election to the United States House of Representatives in 2023 during the 118th United States Congress. This suggests that ideologically extreme candidates may actually benefit from these elections.

Another nonsense answer. The first sentence appears to gesture to the finding in the Zaller et. al research linked. Then it plagiarizes another sentence from the abstract of that paper. Then it introduces a completely nonsense non sequitur and uses it in support of a point that is not in evidence. It never answers the second question at all. The papers linked are all relevant, but not sorted by any useful metric and there are major papers not listed.

> what are the disadvantages of using two-way fixed effects to measure treatment effects in a natural experiment

> The two-way fixed effects regression is a popular method for measuring treatment effects in natural experiments. However, it has some drawbacks that should be considered before using it. First, the model does not account for heterogeneous effects between groups or over time, which can lead to misleading estimates. Additionally, the LSDV estimator is inconsistent for a dynamic panel data model with individual effects, whether the effects are fixed or random. This means that the bias of the LSDV estimator in a dynamic model is generally higher than other estimators. Finally, the 2FE model may produce biased estimates due to the fact that it only adjusts for unobserved unit-specific and time-specific confounders at the same time. Alternative estimators such as the instrumental variable approach or the difference-in-differences approach may provide more accurate estimates of treatment effects.

First three sentences are fine. Fourth sentence introduces an estimator that is not previously defined, probably by plagiarizing one of the source papers. The point starting with "finally" is a restatement of the limitation under "first". The instrumental variable approach is another econometric technique but it is not a substitute for 2WFE. D-i-d is most commonly implemented as 2WFE as an estimator, and so it is also not a separate technique.

From a cost-benefit perspective, it feels like most of the things we did to get sterile environments paid off with massive, dramatic drops in mortality and morbidity at every age group and the costs have been a bit of regulatory burden for business and maybe increases in childhood allergies which are likely to be nonfatal and low morbidity? That's a pretty good tradeoff. Even if you think obesity is a phenomenon entirely caused by gut biome and not food, it's still going to be a huge economic positive.

Cheating to get the Stanley Parable achievement devalues the hard work that all the rest of us did to earn it.

Among people I know at Twitter who got laid off, they have 60 days of notice during which they don't need to work but still get paid + severance commensurate with their tenure (30+ days). All of them think this whole process is an absolute fucking shitshow and the company is going to swirl the toilet bowl, but I don't think any feel as though their legal rights were violated.

I think you do a good job on HN and I appreciate, as someone who moderated a similarly large forum for a long time, how candid you are in your communications on and off site. You're also a very quick email responder!

> I try to use depersonalized language for this reason. Instead of saying "you" did this (yeah that's right, YOU), I'll tell someone that their account is doing something, or that their comment is a certain way. This creates distance between their account or their comment and them, which leaves them freer to be receptive and to change.

My sense is that this is a worthy thing to do (first of all because it's intellectually correct to blame actions rather than people, and second of all because if you're right about the effect it's all upside). But I suspect this will produce very little introspection, maybe a tiny but on the margins.

It's pretty normal in an argument between two people IRL that one will say something like "That was a stupid comment" or "Stop acting like an asshole" -- both uses of distancing language -- and the other person will respond "Don't call me stupid" or "Don't call me an asshole". I think most people who are on the receiving end of even polite correction are going to elide the distancing step.

On the social psych side, I have no idea whether there's any validated way to encourage someone to be more introspective, take a breath, try to switch down into type-II processing, etc.

> The Hunter Biden story is significant because there are emails which suggest Hunter Biden had ties to foreign companies that wanted access to his father. This could be grounds for a criminal case implicating his father. Now, in a sane society during a less polarized election, this would have been discussed in the open, and people would make cases as to the relevance or irrelevance of the evidence, relevance to Joe Biden specifically, and whether the evidence merits investigation through proper channels in an open and transparent way.

I agree with that in principle. The idea that a presidential candidate's son has obviously nepotistic business appointments in industries and countries that he has no business being in presents a facial conflict of interest and should invite a debate on quid pro quo corruption. I would say these kinds of stories are a feature of both political parties in the U.S. and across a wide swath of time -- this is effectively the same role as the Clinton Foundation played in the previous election, for instance. And I think it makes sense to debate or investigate this kind of thing in public, including during the election. In fact, Hunter Biden's role in these companies was scrutinized to some degree earlier in the presidential election and also in the primary and also indeed over the last four years. So I reject the notion on the right that this is an issue that "could not" be discussed because "the elites" suppressed the discussion. My sense is that the eventual conclusion was what it seemed from the very beginning: Hunter Biden is a colossal fuckup, everyone around him tried to help him not be a fuckup, he kept being a fuckup, and it should be to no one's surprise that he tried to leverage his father even though there's no to limited evidence any such quid pro quo occurred.

The laptop, though, is a pretty complicated thing. It feels more like a direct parallel to the Podesta email hacking in the 2016 election (which, among other things, spawned Pizzagate). Bad actors (in 2016, foreign intelligence; in 2020, a domestic political adversary) took steps to secure a large cache of material, potentially by committing crimes to do so, and released it to the public on the very eve of a presidential election as an October surprise. The framing of the stories on the right was that the _existence_ of the material (emails in 2016, laptop in 2020) was proof that the allegations against the candidate were true, even though basically none of the material seems at all connected to the allegations.

Furthermore, the nature of the release was that the people who secured the damaging material obfuscated and laundered it. The press had neither the time nor the resources necessary to verify the contents of the material. Finally, the biggest risk was that large amounts of true material would be combined with knowingly fabricated material in a way that laundered the reputation of the false material. This is hardly an empty false flag thing, it's literally exactly what a nefarious actor _should_ do if they come into possession of huge amounts of stolen material to make the greatest impact.

Although the reactions to the Hunter Biden story were overreactions, that's the framework that I think people had adopted. I think people were thinking back to the Podesta email hack and to the FBI confirmation of the Weiner stuff causing a significant late disruption in 2016 election and wanting to avoid the possibility of a similar disruption here. And I think in principle there is a responsibility to avoid this kind of chaos: for instance, if a candidate is accused of a heinous crime (say rape or murder) the day before an election, it is possible both that the accusation would have a serious impact on the election and that not enough time exists to validate, refute, or contextualize it.

I can't say the exact length of time before an election that should be a sundown period, informally or formally, but I just recognize that you might imagine some combination of late breaking, serious, and difficult to verify that would merit a deliberate choice to not interfere with the election. The same thing occurred in France about 36 hours before the 2017 presidential election: Huge cache of Macron emails were hacked and distributed (probably by foreign intelligence). Opponents of Macron said, without evidence, that this was proof of Macron's perfidy and would sink him. The media generally refused to engage the story because it was the fruit of a poison tree and 36 hours is not enough time to work it out. The far-right party noted that this was censorship and the end of free speech and investigative journalism is over.

The only winner in a situation like this are the agents of chaos.