majromax's recent activity
-
Comment on Why are Americans fighting over no-fault divorce? Maybe they can’t agree what marriage is for. in ~life
-
Comment on Climate deniers don't deny climate change any more. They do something worse. in ~enviro
majromax Link ParentIt's less expensive to fix one scratch than 100 scratches, yes. However, the marginal cost of fixing the 100th scratch given that 99 have already happened is nearly nil: you're already repainting...No? It's less expensive to fix one scratch than a hundred, that's common sense.
It's less expensive to fix one scratch than 100 scratches, yes. However, the marginal cost of fixing the 100th scratch given that 99 have already happened is nearly nil: you're already repainting the whole car door.
In the climate context, going from 0° to 4° is of course more costly than going from 0° to 2°, but the 2-4° part is worse than the 0-2° part. The "climate is doomed" perception falsely applies car-door logic to the planet, thinking that since so much damage is already done there's not much more left to do on the margin.
-
Comment on What's the matter with men? They’re floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex. in ~life.men
majromax Link ParentWith hindsight, I don't think it's all that surprising. We saw the end of the "A woman? With a middle-class job? Because she likes it?" plot lines on TV by the middle to end of the 80s, and a...I would have expected to see that narrowing more over the last two decades. (December 2003 was 86.4% for men and 71.5% for women, a 14.9 point difference.)
With hindsight, I don't think it's all that surprising. We saw the end of the "A woman? With a middle-class job? Because she likes it?" plot lines on TV by the middle to end of the 80s, and a woman who turned 18 in 1985 would be right in the middle of the 18-55 bracket by 2000. From the early 90s onwards, women were fully normalized in the workforce, and remaining barriers that force a family-versus-career choice (notably parental leave policies and daycare subsidization) are more slowly-changing.
-
Comment on What's the matter with men? They’re floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex. in ~life.men
majromax Link ParentLooking at the various series available, I think you can construct a chart showing stagnation if you look at the aggregate earnings per week, per person, essentially multiplying inflation-adjusted...Looking at the various series available, I think you can construct a chart showing stagnation if you look at the aggregate earnings per week, per person, essentially multiplying inflation-adjusted $/wk by either the employment/population ratio or the prime-age employment ratio.
In both cases, you see an upwards trend halted by Covid, with rough stagnation thereafter as inflation has approximately compensated for nominal wage gains. The overall series looks smoother to my eye than if you exclude the employment ratios; that sees a post-covid bump and an inflationary drop.
-
Comment on Climate deniers don't deny climate change any more. They do something worse. in ~enviro
majromax Link ParentIf you accept this as fact, then in a fun twist of irony this makes your individual action even more important. Climate change is not a binary "safe/doomed" set of outcomes. It's a range, with the...A little bit afraid to say this, but while I am by no means a doomer (by his definition), I certainly am very afraid that this is not going to be solved in anything even closely resembling good time.
If you accept this as fact, then in a fun twist of irony this makes your individual action even more important.
Climate change is not a binary "safe/doomed" set of outcomes. It's a range, with the cost of mitigation / damages from non-mitigation increasing with the amount of global warming. One study for Canada modeled an $80bn(CDN) cost to Canada by 2100 if global warming were 3° rather than 2°, with an additional $100bn(CDN) for 4° over 3°.
This breaks our intuition. Most things in our life work the opposite way: the first scratch on the car door is "worse" than the 100th, for example. For climate economics over a global scale, however, as warming increases more and more human systems (small things, like city placement) change from "able to cope with climate change with some investment or loss of productivity" to "largely untenable, needing wholesale replacement." Diminishing returns don't set in until civilization-ending levels of climate damage, at which point there wouldn't be many people left to pay the bill.
This fundamentally makes doomism a self-defeating argument. The first bit of emissions reduction accomplishes the most long-term good, no matter how it's achieved. If a set of policies only brings the world from 4° to 3.5°, that's still more damages saved than bringing the world from 2.5° to 2°. That first lone voice in the wilderness reducing carbon emissions does the most direct good.
-
Comment on What's the matter with men? They’re floundering at school and in the workplace. Some conservatives blame a crisis of masculinity, but the problems—and their solutions—are far more complex. in ~life.men
majromax Link ParentYes, but the demographics of today's United States don't match that of the 1980s: the country is older, overall. Older people are more likely to take early retirement, so even with the same...- Exemplary
Yes, but the demographics of today's United States don't match that of the 1980s: the country is older, overall. Older people are more likely to take early retirement, so even with the same employment-rate-per-age the overall employment rate appears to suffer.
My preferred labour statistic is the prime-age employment rate: the employment ratio of people between 25 and 54. This helps control for the demographic shift.
In this case, the current prime-age employment ratio matches other cyclical peaks like 2019, 2007, and 1990, but it's a couple of percentage points below the late 90s peak – highlighting that period as a particular boom.
As another argument against a long-term weakening of the job market, the employment rate for 55-64 year olds is also near a record high. While more people are aging into the early-retirement bracket, fewer are in fact taking early retirement.
-
Comment on A peer reviewed journal with nonsense AI images was just published in ~science
majromax LinkNote that the journal, Frontiers in Cell and Developmental Biology, is listed among predatory journals, and the publisher (Frontiers) does not have a good reputation. This is not so much a failure...Note that the journal, Frontiers in Cell and Developmental Biology, is listed among predatory journals, and the publisher (Frontiers) does not have a good reputation.
This is not so much a failure of peer review as its absence, and it should not be surprising that a predatory journal publishes nonsense.
-
Comment on Battery life of AAA batteries that come with the original products seem unusually long in ~tech
majromax Link ParentYou'd notice it as a suspiciously short battery life, being half or less the battery life with single-use alkalines. Some devices that were particularly sensitive threw up a 'low battery' warning...You'd notice it as a suspiciously short battery life, being half or less the battery life with single-use alkalines. Some devices that were particularly sensitive threw up a 'low battery' warning with even fresh rechargeable cells.
Fortunately, this problem has become much less prevalent with time. Warnings against rechargeable cells were much more common in the 90s, right around the intersection of "lots of devices are using electronics that need fixed voltages" and "very low-voltage electronics or flexible power supplies are too expensive for many consumer devices."
(Edit to add:) Also, rechargeable batteries themselves have gotten better. Before nickel metal hydride batteries, the most common rechargeable cell type was nickel cadmium, with a similar cell voltage. Those cells, however, tend to have a large self-discharge rate, making them much less suitable for 'occasionally used' devices like remote controls where steady power output isn't the point.
-
Comment on Battery life of AAA batteries that come with the original products seem unusually long in ~tech
majromax Link ParentNon-rechargeable alkaline batteries provide a nominal 1.5 volts per cell, so with four cells in series the device expects 6 volts of power. In the meantime, electronic devices expect to see a...Huh, I honestly, have no idea that some devices works only with non-rechargeable batteries.. Interesting, but why.. Is there some spikes in power then only non-rechargeable batteries supports..
Non-rechargeable alkaline batteries provide a nominal 1.5 volts per cell, so with four cells in series the device expects 6 volts of power.
In the meantime, electronic devices expect to see a fixed voltage, and they often use linear regulators for that purpose. These regulators take a variable supply voltage and lower is to a fixed target, but they need some headroom to perform that regulation. Modern electronics might require 3.3 volts, and a cheap regulator might require an extra volt of headroom, so on net the device could work with anything above 4.3V. 6 volts from four batteries in series would be plenty.
However, the battery voltage drops as it is discharged. For regular alkaline batteries in our hypothetical device here, this isn't much of a problem: the alkalines will be essentially dead before four in series stop providing 4.5V.
Unfortunately, rechargeable cells aren't so generous. Nickel metal hydride cells, the current generic rechargeable in AA-battery format, has a nominal cell voltage of 1.2V rather than 1.5V. Four of those in series will only provide 4.8V nominal voltage. That's still sufficient for our hypothetical device, but now there's much less headroom for the voltage to fall as the batteries are discharged. These cells have a 'flatter' discharge curve than alkalines, but if a device needs a voltage near the peak of that curve it can "run down" while rechargeable cells are barely discharged.
For very modern electronics, this phenomenon is less of a problem. Built-in rechargeable lithium batteries are more common, standardizing on a rechargeable format with fixed and known behaviour. In addition, switch-mode voltage regulators that can decrease or increase the supply voltage to meet a device's needs are also more common, making it easier to design a device that can handle a wider range of supply voltages.
-
Comment on Is fandom.com actually getting worse? in ~tech
majromax Link ParentRemember that humans aren't the only things browsing the web. All of those inter-links also matter to Google and other web-crawlers; they increase the discoverability and connectivity of other...I can't really believe they would be introducing so many inter-wiki links/menus if they weren't trying to get users to browse around the pages more after landing there from Google.
Remember that humans aren't the only things browsing the web. All of those inter-links also matter to Google and other web-crawlers; they increase the discoverability and connectivity of other pages for the purposes of search rankings.
One significant part of SEO is reputation laundering. If a lot of users search for "redstone," visit the first link returned (Google gives me the fandom Minecraft 'redstone dust' page as the top result), and are evidently happy with the result (not returning to the search page to look at other results), then search engines can infer that the linked-to page is good and thus its outgoing links should carry some weight.
-
Comment on A man blew up his apartment in Japan while trying to kill a single cockroach with way too much insecticide, police say in ~life
majromax LinkA 2004 Mythbusters Episode covered a similar situation, whereby… considerable overuse of "bug bombs" could create an explosive fuel/air mixture. The episode also found a documented US case of such...A 2004 Mythbusters Episode covered a similar situation, whereby… considerable overuse of "bug bombs" could create an explosive fuel/air mixture. The episode also found a documented US case of such an explosion.
-
Comment on 'The Marvels' ends box office run as lowest-grossing MCU movie in history in ~movies
majromax Link ParentI think that point #3 (dependency chains) is a problem even if the movies are less frequent. Movie series take place over time, so each new movie might be someone's first (in the franchise)....3 is not such a problem if the movies are less frequent and better.
I think that point #3 (dependency chains) is a problem even if the movies are less frequent.
Movie series take place over time, so each new movie might be someone's first (in the franchise). Saying "to understand this movie, you first need to watch these movies from X years ago" is a hurdle, even if the raw number of movies is low.
Disney got this right with Star Wars 7-9. For all the flaws of that trilogy, the movies didn't depend on knowing any details of movies 4-6 or 1-3.
For very long-running series, the whole dependency chain becomes a problem even more than the dependencies for any particular movie. If a proper understanding of this movie depends on the single prior movie, but understanding that movie depends on the one before it, and so on back to the series-originator, jumping into the franchise – essentially the only way that the studio can build its audience over time – would require new patrons to complete a daunting amount of homework.
-
Comment on 'The Marvels' ends box office run as lowest-grossing MCU movie in history in ~movies
majromax Link ParentIt's more like Calvinball: the rules really aren't important, it's about who you're sharing the experience with. If you find the rest of the experience engaging, with enjoyable characters and a...Without proper definition of ideas and powers, it's basically just playing action figures with my 5 year old nephew:
It's more like Calvinball: the rules really aren't important, it's about who you're sharing the experience with. If you find the rest of the experience engaging, with enjoyable characters and a pleasant environment (good special effects, good pacing, good music/set design/locations), then it's fine. If not, then your conclusion comes right back to the fore: what's the point of it all?
In other visual media, sitcoms and soap operas work this way, where consistent limitations aren't really the point. Homer Simpson's dramatic shifts of competence from episode to episode don't detract from the overall series, and the show's generally-regarded decline has more to do with it becoming formulaic.
On the other hand, if rules aren't the point, then this needs to be a self-aware point. The movie can't make a Big Deal of "but we can't!" in one scene, only to have the same rule casually flaunted an act later. Contradictions can still be okay, but they have to support the surrounding tone and plot, rather than drive it.
-
Comment on US taxpayers paid for 5.8 billion of Lockheed's 7.9 billion dollar stock buyback in ~finance
majromax Link ParentAn expected stock buyback does not change the price of shares very much. While the business purchases shares on the market and retires them (reducing the overall availability of its shares), in...Increasing the price of stocks isn't a side effect, it's the purpose. To increase the wealth of investors by making the remaining shares more valuable.
An expected stock buyback does not change the price of shares very much. While the business purchases shares on the market and retires them (reducing the overall availability of its shares), in executing the buyback it also spends money, making itself a modestly smaller company. The equity:share ratio remains constant.
Announcing a dividend payment of 8 billion distributed throughout all the shares would have a similar impact on the stocks suddenly getting more valuable,
Dividends function in a similar way, but issuing an expected dividend reduces the stock prices. After issuing the payment, there's now less company but the same number of shares. An example of this effect is in the six-month chart of JPST, a short-term income ETF that issues dividends monthly (and thus predictably): at the start of each month the share price drops by more or less the dividend amount.
Unexpected dividends and share buybacks, however, can have an influence on stock prices. In this case, however, the effect isn't from the mechanical purchase/dividend, it's from investor expectations. Suddenly, this company has become one that gives (more) income back to shareholders. That income stream has value, particularly if strong within-company growth is unlikely thanks to a mature industry.
-
Comment on Not all porn is created equal - is there such a thing as a healthy pornography? in ~life
majromax Link ParentI think there's a problem of implementation. As a society, we can ban things for children because they are physical things, with fundamental limitations on distribution. We can check ID before...- Exemplary
Horseshoe theory really applies here... I entirely disagree with those in the article that want to ban books like this, outlaw porn etc., but I also disagree that porn is great and good and should be a free for all. I think there is absolutely a discussion to be had on age limits. As they say in the article, the barrier to entry has been lowered and I think it has been lowered way too much.
I think there's a problem of implementation. As a society, we can ban things for children because they are physical things, with fundamental limitations on distribution. We can check ID before selling alcohol or cigarettes precisely because the sale is somewhat centralized.
For the past 25 years or so, however, pornography is no longer "a physical thing" but rather abstract information, distributed digitally over a channel where people are fundamentally anonymous. It's hard to imagine how to control pornography distribution without radically changing the Internet to attach an identity to every online action. The costs of the latter might not outweigh the benefits of the former.
It's unhealthy for young kids like I was when I discovered it (about 12 or so) to consume it. Period. […] It completely skews your frame of reference.
I think it doesn't help that kids at that age don't have a frame of reference for sex. For Very Good Reasons we shelter kids from in-person sexuality, but that just leaves a vacuum to be filled with rumour, imagination, and (over the last 40 years or so – don't forget the old tropes of boys stealing Playboy mags from adult figures) pornography.
Like, maybe one would start thinking that things porn has normalized - take anal sex
Nevermind that, I suspect that pornography of earlier decades went a long way towards normalizing oral sex, a task accomplished so thoroughly that the idea is no longer controversial (see for example an academic study on prevalence by age, or a 2008 Slate article).
I think the inflection point isn't even the Internet, but rather mass media in general. We're in the middle of a generations-long shift where cultural attitudes are no longer conveyed from the community to its children, but from the media to citizens writ large (including children) without strong intermediation by traditional authorities.
Sociologists a couple of hundred years from now might be able to come to a consensus about this change, but sitting in the middle of it I don't think I have any particularly privileged perspective.
-
Comment on Not all porn is created equal - is there such a thing as a healthy pornography? in ~life
majromax Link ParentThe 'harms of pornography' seem to fall into two categories: Harm in the production of pornography, running the spectrum from undisputed harms like violent abuse to more controversial harms like...- Exemplary
For all of real porn’s faults, I find it deeply disconcerting the recent drive to divorce it from real people entirely. Whether that’s through hentai, 3d animated cam girls, or in this case random generation.
The 'harms of pornography' seem to fall into two categories:
- Harm in the production of pornography, running the spectrum from undisputed harms like violent abuse to more controversial harms like the 'inherently degrading nature of sex work'.
- Harm in the consumption of pornography, again running a spectrum from 'giving children harmful misconceptions about sex and relationships' to 'it promotes lustful thoughts outside heterosexual marriage'.
Porn-without-people – not just AI-generated porn, but also written/drawn/animated works – mostly eliminates the harms of the production side. For people primarily concerned about exploitation in sex work, this is a good thing. However, it doesn't do anything about consumption-based harm, so the relative merits will depend on your analytical framework.
-
Comment on The problem with ChatGPT is that all of these websites like W3Schools and TutorialsPoint will go bankrupt in ~tech
majromax Link ParentThis has been postulated as the Waluigi effect. The basic idea is that any simulated persona is a superposition of two opposite states: The persona as intended, and More or less the exact...Well, sort of. Here’s what I mean by “learning the pattern:” an LLM is always imitating some writing style, based on context. Once it starts making mistakes, the writing style it’s imitating becomes that of someone who makes mistakes.
This has been postulated as the Waluigi effect. The basic idea is that any simulated persona is a superposition of two opposite states:
- The persona as intended, and
- More or less the exact opposite, pretending to be the persona specified.
In this theory, if the LLM ever 'slips up' and responds according to the anti-persona, then the superposition collapses and you're left with just that one.
Specializing this to the case of a 'smart, helpful code generator', the personas would be:
- The smart, helpful, and correct code generator, or
- A persona that acts like it's smart and helpful but really knows little.
If the LLM begins making mistakes, then it's increasingly more likely to settle on the "incompetent at code but good at bluster and ego" version. Once that happens, the "incompetent at code" persona is unlikely to rediscover technical aptitude.
This hypothesis is not without its critics (see the comments of the above link for more discussion), but I consider it a neat way to break our human-based intuition† and grapple with LLM simulators on their own terms.
† — Namely, that humans do have a consistent persona, such that it's possible to learn about the 'real you' with increasing fidelity.
-
Comment on Twitter is now X as the little blue bird disappears in ~tech
majromax Link ParentI think it's because "extreme free speech" isn't really a draw for the average user. Those drawn in are disproportionately people who have had their hobbies kicked off the existing popular...I find it fascinating that this pattern hasn't been successfully replicated on all the extreme free speech replacements for Reddit or Twitter.
I think it's because "extreme free speech" isn't really a draw for the average user. Those drawn in are disproportionately people who have had their hobbies kicked off the existing popular networks, so the new network is seeded with the "wrong crowd." This is exactly the story of Voat, after all.
To me, the competing theory is that the pattern is one of semi-benign neglect. A community that starts for anodyne reasons develops its own shady corners with questionable content, but in moderation that content turns some users of the site into dedicated users. It's still tickling the dragon's tail, since exposure is bad news for the site as a whole.
Another instance of this pattern would be Tumblr, where many of its users stayed precisely for the kind of adult or adult-adjacent content that the site sought to restrict.
-
Comment on Twitter is now X as the little blue bird disappears in ~tech
majromax Link ParentIn addition to the problems you mention, I think there's one to add: image hosting marked Reddit's change in focus from a link aggregator to a social media platform. A link aggregator's job is to...2016 - Reddit image hosting. You can still use imgur to host images of course, or some other hosts, but reddit started their own hosting and made it the default so most casual users use that. But the problem is twofold
In addition to the problems you mention, I think there's one to add: image hosting marked Reddit's change in focus from a link aggregator to a social media platform.
A link aggregator's job is to connect the user with an outbound link. The user leaving the site is the point, since the link (hopefully) directs them to interesting content elsewhere on the Internet. However, that's a failure state for a social media platform, which values (and monetizes) "user engagement" measured by time spent interacting with the main site.
By hosting images and then video, Reddit hoped to keep users engaging with its own comment threads, driving up engagement and hopefully exposing the user to more ad-monetization. Most of Reddit's user-questionable decisions (such as built-in chat) also seem to have a similar focus†, keeping user eyeballs somewhere on reddit.com rather than an external provider.
The redesign also has nods in this direction. Under the redesign, the title of a link-post does not open the link itself, it opens the comment section. Opening the link requires clicking on the URL itself, which has much less visual weight than the title, comparable to the domain on old.reddit. It's a subtle nudge towards the idea that even for link posts, users should treat the link as optional while feeling free to engage in the comment section.
† — In my opinion, the biggest feature introduced here wasn't image/video hosting or even chat, it was turning user pages into content-hosting places equivalent to subreddits themselves. As far as I can tell this hasn't had much impact, but it seems to me to have been designed as an answer to Facebook walls, turning users themselves into content-display points.
-
Comment on Why is Elon Musk doing what he is to Twitter? in ~tech
majromax Link ParentWere that the case, he could have played hardball during the contract-writing stage. Unfortunately, he was so eager to have the contract signed that he didn't perform pre-contract due diligence,...But then Twitter said "yeah ok go on then" and he tried his hardest to not actually buy twitter.
Were that the case, he could have played hardball during the contract-writing stage. Unfortunately, he was so eager to have the contract signed that he didn't perform pre-contract due diligence, which could have let him walk away for some of the same grounds (e.g. "bots") that he tried to use during the lawsuit stage.
His purchase of Twitter may have started as a joke and ended as a legal compulsion, but somewhere in the middle he was earnestly trying to own the thing.
I don't see the contradiction here. Contracts can generally be broken by either party on short notice. Even if the breach goes against the term of the contract itself, general damages for breach of contract relate to actual damages, which could include (judicially reasonable) penalty clauses.
The family law system does an adequate job of implicitly performing this calculation, namely through the division of marital property and (depending on jurisdiction and reasons) awarding spousal support. Couples who want an even more contractual marriage can arrange one through a pre-nup agreement.
Instead, I think the outrage about marriage being "unilaterally broken" comes from an implicitly religious framework. If marriage is a voluntary agreement between two people for mutual benefit, it's reasonable that it could be ended when it no longer provides a mutual benefit. If instead it is a sacrament, then God itself is a third party to the contract who ought to be satisfied. This kind of logic has little place in a secular legal system, and it ought to remain in the chapel.