site banner
PaperclipPerfector  13d ago (text post)   2282 thread views

Small-Scale Question Sunday for September 27, 2026

Do you have a dumb question that you're kind of embarrassed to ask in the main thread? Is there something you're just not sure about?

This is your opportunity to ask questions. No question too simple or too silly.

Culture war topics are accepted, and proposals for a better intro post are appreciated.

3
Top Bottom Old Controversial Comments
Jump in the discussion.

No email address required.

I'm serious when I ask this: Can we have a discussion somewhere about removing the ability of people to delete or edit their posts on here after some set time period? An acquaintance found some EA stuff and was doing some digging and asking about who the good commenters other than Naraburns are, I was trying to show them some good old threads I have saved but half of the comments are responses to deleted top level posts or effort replies which makes it basically unusable.

The discussion has been had somewhere on here before. The reason we allow deletions is that some people have reasonable fears about doxxing and malicious actors using their writing against them in real life.

Having said that, if someone is quickly deleting all of their posts we look at that as troll adjacent behavior. People who want to give a minor "fuck you" to the community usually delete all their stuff after a ban.

For individual commentors I'd suggest quoting the important bits of any discussion you are having. It stays readable even if the other person deletes their comments.

For the record I know of at least one freely accessible public scrape of this site (though it isn't actively updated): https://huggingface.co/datasets/x65617379/themotte_260508

do you know of any publicly-accessible scrapes of the old subreddit?

I think commoncrawl had most subreddits at least up until the API changes?

Who would bother indexing this besides a member? 12 downloads in a month is higher than I'd expect, too. Perhaps from the same toy project.

The same account uploaded datasets for LessWrong, EAforum, EJMR and AutoAdmit. Kinda just seems like someone who did academic research on online forum culture.

What about post deletions (or mass edits that remove >30% of the content of a post) requiring mod approval and not going through without it? Or maybe temporarily going through in case things are time-sensitive like accidentally letting personal information slip, but requiring a mod appeal alongside it and reverting if the mods don't accept the reasoning.

Please remember that these pie-in-the-sky proposals would require a bunch of programming to actually implement. The administrator has not been doing much work on the site lately, and has explicitly said that the current codebase is a "dead end".

Things have been made easier with AI programming assistance. There are still changes being made to the codebase. One of the recent changes was to prevent inbox harassment from new users that haven't made it out of the filter.

If things become urgent or serious enough then code changes are still on the table. We don't feel that comment deletions are too serious as of late. It was much worse a while back and some rule enforcement changes nipped it in the bud.

Fair enough. I think the deletion of old content across the internet is a real tragedy and makes it a lot worse of a place that it otherwise could be, which is an underrated and underdiscussed issue in general. But I understand that practical limitations exist.

Do any of the more technically-inclined people here know anything about how to get past Cloudflare anti-bot verification?

I used to be able to get past — barely — if I made sure that absolutely nothing else (no other browser tabs, no background processes) was using any of my bandwidth. Then, starting a couple weeks back, even that didn't always work, and I'd sometimes get caught in the failing & retrying loop. When this happened, I'd need to stop and try again in a few hours.

As of a few days ago, that stopped working too, and I'm completely unable to get verified as not a bot. I've tried every bit of troubleshooting I could find online (on my end) — caches, different web browsers. The only thing that worked was switching over to my phone as an internet hotspot… until I hit my monthly data cap (after only a couple of hours) and got that throttled, which made it stop working too.

Is this truly an issue with my ISP or connection itself, as it appears, or is there still something I can do? From what I've read, the slow speed of my DSL connection should not by itself be an issue — that there's no "minimum speed" for passing the Cloudflare anti-bot verification. I'm in Alaska, so that may be an issue.

(I don't have much hope of contacting my ISP doing any good.)

Cloudflare should ask you to solve a captcha once and gives you a cookie when you win that should work for a while. I never have issues even when behind VPNs that are definitely being used to do mass web scraping. It may just be your network speed if you don't have some kind of anti-cookie setting.

Sorry to be the one to tell you this, but you've actually been a replicant this whole time. The tests have just now gotten good enough to detect you.

Did you use the same browser when doing the hotspot method? If yes, it means your home IP address must be hard flagged in Cloudflare's system. The solution is to get request a new IP from your ISP. If that's not an option, you can try doing a hard reset of your modem but that doesn't guarantee a new IP address in most cases since IP allocation is handled on ISP's end. You can also try some VPNs and residential proxies as well

Did you use the same browser when doing the hotspot method?

Yes, I did.

If yes, it means your home IP address must be hard flagged in Cloudflare's system. The solution is to get request a new IP from your ISP. If that's not an option, you can try doing a hard reset of your modem but that doesn't guarantee a new IP address in most cases since IP allocation is handled on ISP's end.

Thanks.

You can also try some VPNs and residential proxies as well

Aren't those more likely to set off Cloudflare?

Aren't those more likely to set off Cloudflare?

Well, in your case, it seems that your IP address is fully cooked, so it may be worth trying alternatives. Some VPNs should work and most residential proxies will work too - 'residential' just means that it's a real consumer's IP address being used to route your traffic (how companies get ahold of those is a long story, often not entirely ethical, but whatever) and those IPs are usually the cleanest.

You might have malware on your computer (or other computers/devices in your house) and that's what's getting your home IP on cloudflare's shitlist.

Try changing your IP. Ask Google what your IP address is (or at one of the top hits) and then reboot your cable modem, see if your IP changes and see if your cloudflare problems go away.

If it goes away for a bit and then gets worse again I think that means you're part of a botnet.

Or he's just behind CGNAT and someone else on that IP is scraping with it. If so he could switch services to someone who will give him dedicated or he could do something like buy a high reputation IP from digitalocean or nordvpn or wherever and configure his router to use that as a VPN. If his router can't establish VPN connections then he needs to buy a router that can ...

My understanding of cloudflare is that they aggregate information about how your browse the internet from the sites that use their service to see if you browse in a humanlike way. Having any sort of adblocking, scriptblocking, cache deletion, history deletion, vpn, cookie blocking, etc might interfere with their ability to verify that you're a human. If they fail-safe and block access... There you go.

That being said, your description of trying different browsers makes me think the above probably isn't your problem. So maybe Cloudflare just thinks your ip is from a region with a lot of bot traffic? Especially if russian, chinese, korean addresses get into america through fiber lines and DNS hops in the bering straight... Maybe try a VPN instead? You can start with one of the free ones online just to test (though fyi those will 100% harvest your data) then get a real service if it works.

So, what are you reading?

I'm picking up al-Katibi's The Rules of Logic, an apparently influential Arabic logic primer. The book itself is actually quite short, with the commentary by the translator taking up the majority of the pages.

It has recently been said of Peter of Spain’s Summaries of Logic, probably written in the 1230s, that no other book on logic prior to the twentieth century had such wide readership or, in consequence, did so much to shape Western ways of constructing well-formed sentences and putting them together in valid arguments. But perhaps one other logic text has had just as many readers, and as profound an impact on ways of formal discourse and argument. In the Muslim East, al-Risālah al-Shamsiyyah (literally The Epistle [on Logical Rules] for Shams al-Dīn) was written some forty years after the Summaries, by someone who was Peter’s exact contemporary: Najm al-Dīn al-Kātibī (d. 675/1276). Al-Risālah al-Shamsiyyah also came to figure in the education of nearly every aspiring young scholar, and is still read in traditional schools.

Started a long-run re-read (or, rather, re-listen, as I am using audiobooks for that) of Nero Wolfe series. I've read many of them (some in originals, more in Russian translations) years ago, and decided to go back. So far I got through: Fer-de-Lance, The Rubber Band and The Red Box. Skipped The League of Frightened Men for now as the audiobook wasn't available, plan to fill up later. There's a lot of them spanning 40 years of Stout's lifetime, so I expect it to take quite a while to go through all of them (and of course I'll be taking breaks for other books) but so far the first three had been fun. It's a bit slower paced than our twitter age made us used to, but that's fine. I'll probably also pick up the next Dresden book sometime next month, but Ghost Story was kind of intense (in a good sense) and I don't want to ruin the impression by getting through too much too soon, so I'll let it sit down for a bit yet.

Just started the last book in Elena Ferrante's Neapolitan quartet, The Story of the Lost Child. Having had my suspicions during the first three books in the series, I can now confirm that protagonist Elena Greco has taken an overdose of Dumb Bitch Juice™.

Senshiro, by Natsume Soseki. Not terribly far in yet, will report once I'm done.

Other than that, I recently read the Godfather, which was okay. Better-written than expected, but on the other hand it didn't really do much for me.

No Bad Parts by Richard Schwartz.

I finished The Best Minds this week. Author Jonathan Rosen published several novels before writing this, and there are times when you feel like you're reading a novel rather than a work of non-fiction. That's not solely due to the manner of its telling, but also because the story itself is so surprising and so saturated with dark irony. Indeed, if it was a work of fiction, a satire of well-meaning but misguided deinstitutionalization efforts, its basic premise (a brilliant but troubled young man is diagnosed with schizophrenia, becomes the public face of contemporary efforts to break the stigma around mental illness and stop the mentally ill from being discriminated against, sells the rights to tell his inspirational biography to Hollywood for more than a million dollars – then has a psychotic break and stabs his pregnant fiancée to death) would seem almost painfully on-the-nose.

While reading the book, it struck me that it's a clever subversion of a genre of books and films I call "TOA stories": "triumph over adversity". These are formulaic and sentimental stories about a brilliant individual who faces some unexpected major obstacle in his life (physical disability, disease, mental illness), struggles mightily against it, and summons strength he didn't know he had to overcome it. Something something mind over matter, inner human resolve, a moving testament to the power of the human spirit etc. The Best Minds starts off following that structure to the letter, but where a conventional TOA story would have ended with Michael selling the rights to his story, marrying his girlfriend and delivering an inspirational speech, Minds keeps going, showing that the real thing is nowhere near that simple. In the end, adversity triumphed over Michael, not the other way around. Part of the reason adversity triumphed over him was because so many around him chose to pretend that he'd already triumphed over adversity, and hence missed (or, less charitably, ignored) obvious red flags that his condition was very much not a thing of the past. Rosen notes that, prior to his murdering his fiancée, so many articles about Michael described him as having "a history of schizophrenia", as opposed to "schizophrenic". Amusingly, the phrase "triumph over adversity" does appear in the book, in the form of a press release Rosen quotes about the film A Beautiful Mind. That film originated as a loose biopic about Michael, but after he murdered his fiancée, the producers decided they needed a more palatable lunatic for their subject.

Interestingly, after a certain point the real John Forbes Nash gave up taking antipsychotic medication, but near the end of A Beautiful Mind, Nash claims to be taking meds. Director Ron Howard felt this artistic license was the responsible choice, as he did not want to send the message that schizophrenics should attempt to overcome their condition via sheer willpower. His decision feels even more poignant when you understand the genesis of the project: Michael's story must have been weighing heavily on Howard's mind, as a cautionary tale of what can happen when schizophrenics don't take their medication.

The Best Minds is easily the best work of non-fiction I've read so far this year, and the saddest work of non-fiction I've read in years. Highly recommended.

Magic-school time-loop litRPG progression slop: Serpent's Coil. Twist: The protagonist of the story is, not the primary looper, but just a secondary looper who can be erased by the primary looper if he ceases to be useful.

Mildly interesting Harry Potter short story: The Subtle Science and Exact Art of Asking the Right Question

Potions is presumably based on medieval alchemy more than anything else. Slughorn's antidote lesson is even called "almost alchemical". Why shouldn't it look like it?

Besides which, the other thing that comes up a lot in fanon is tables of ingredient interactions. With the hundreds, perhaps thousands of ingredients used in various potions, those could fill whole volumes like the logarithm tables of the 1600s. But worse is having to keep track of all the interactions within a potion that might have dozens of ingredients. (Hermione's antidote in that same lesson had 52.)

But the thing is, that's not really how real-world chemistry works. In chemistry, you make one (or a few) intermediate products at each step (such as Syrup of Adoneta here), and you only have to keep track of interactions with those intermediate products, which themselves may fall into a few broad categories. That problem is still fiendishly complex (ask any pre-med student who's taken Organic Chemistry), but much more tractable, as it scales linearly rather than quadratically.

Even fics that go all-in on Snape mentoring Harry to become a potions master rarely give a reason for the brewing steps, rather seemingly pulling names of ingredients and interactions out of thin air. They rarely even address the question of why the Half-Blood Prince's canonical potions modifications work. Treating potions like real-world chemical synthesis could go a long way to resolving those questions.

TL;DR, if I had more free time and a degree in chemistry, it would be eminently possible to create a potions-centric canon rewrite every bit as rigorous and innovative as the Arithmancer-Verse was for arithmancy. But since I don't have either of those things, let this one-shot stand as a taste of what's possible.

Mildly entertaining magical-girl novella: An Uneasy Stalemate

The Bane known as Decay has watched his fellows getting slaughtered, and decided he wants no part in their fate. When he's ordered to cross the rift to Earth himself, he's willing to do anything to survive. There may be no chance of victory, but, with a little effort and a few acts very much against his nature, it might just be possible not to lose.

This is a short story, mostly following the POV of the very confused and slightly traumatised magical girls as they try to figure out what to do with their rather atypical enemy.

Mildly entertaining grimdark Madoka Magica fanfiction: Fargo DX

In the frostbitten American Midwest, ragged Magical Girls vie for territory to survive the unforgiving landscape. One such girl is Sloan Redfearn, who wastes away in no man's land after being betrayed by her best friend. Only one goal keeps her alive: Revenge.

Bavitz has also been posting edited-Fargo as Magical Girl Machine Gun on RR. I…don’t know if it’s the same version as DX, but it looks like it started going up at the same time.

Mildly interesting Harry Potter short story: The Subtle Science and Exact Art of Asking the Right Question

Hey, that wasn't that bad. Interesting concept, and I like fics where characters like Snape are turned into more believable human beings without rendering them unrecognizable. Which makes it better than 99% of content of Ao3, thus, quite good actually. How do you find more of such fiction on the internet?

I've been subscribed to this author for many years on fanfiction.net (probably as a result of reading the Arithmancer series, though it's been so long since then that I can't remember), though I haven't actually read any of his work in a while. It was only upon reading this new short story, and noticing that fanfiction.net now allows advertisers to insert random links all over the page, that I checked the author's profile and realized that he had moved to ArchiveOfOurOwn.org.

Mildly entertaining grimdark Madoka Magica fanfiction: Fargo DX

If you didn't already know this, the same author (Bavitz) wrote "When I Win the World Ends", which is probably my #2 pokemon fanfic.

I can't say that I'm the biggest Pokémon fan. I have vague memories of watching a few third-generation anime episodes on television as a child, I caught all the Shadow Pokémon in Pokémon XD, and I've watched a fair number of gameplay videos (speedrun and competitive) on YouTube. That's the extent of my experience with the canon—enough that I probably could give correct labels to images of all the Pokémon from generations 1–3, but not many past those. I've read very little fanfiction for it, and none of that fanfiction interested me enough that I bothered to download it. The first four chapters of When I Win the World Ends do not spark much interest in me either.

(In fact, I'm not even the biggest Madoka Magica fan. IIRC, prior to reading Fargo DX, my knowledge of the magical-girl genre consisted of Sailor Nothing, Sailor Moon season 1, and scattered 4chan webms of Pretty Cure battles, in that order. It wasn't until after reading Fargo DX that I even bothered to watch Madoka Magica.)

It may help to know that When I Win the World Ends is specifically a fanfiction of competitive pokemon, inspired by the author's onetime league of legends obsession and smogon/pokemon showdown/vgc. Pokemon-as-sports-movie rather than pokemon-as-adventure-story, with all the megalomiania, pressure for victory, and strategic complexity that entails.

...of course if you already read and dismissed the first few chapters it probably just isn't for uou.

Who here uses Tesla FSD? I finally got a Tesla after two of my friends nagged me to for years. And holy shit, the car drives itself. How does Tesla not have like 90% market share? Why are people still buying Fords and Volkswagens?

The government should be incentivizing the crap out of these. I asked Astra and it estimates the U.S. loses 0.9 QALY's per person due to vehicle accidents. We could probably reduce this by 50% at least just by rolling out FSD.

I agree that most people would be vastly better off using FSD, and I'd be better off if they did that too. However, personally, I find it to be a worse driver than me and can't stand the sensation of letting it handle things. For example, it always starts moving into a turn just a bit later than I would, takes a bad line, and every single time I have that moment of wondering whether it's going to turn at all. It always does! But not as soon as I would.

I do use it sometimes on long drives on fairly straight freeways, like I5 up and down California.

For reference, I'm an excellent driver in that I drive a ton, have a perfect safety record, and perhaps more importantly have professional training on and off the racetrack. Also the usual high IQ reaction times and 'common' sense. Very conservative, etc. So definitely nothing like the modal murder-chimps surrounding me on the road all day.

And man have I seen it all. Women curling their hair. A man flying by at 80 MPH in a 65 zone, with bad visibility, eating a cob of corn with both hands. Too many examples to even remember any more.

By all means get everyone else on FSD ASAP. Just please don't make it mandatory.

professional training on and off the racetrack

What kind of racing do you do? I used to love doing dirt track. I have some friends who are deep into autocross these days and have been trying to get me involved, but I really don't want to buy a car for it.

I'm not as into it as I used to be but basically just driving personally-owned sports cars at Laguna Seca etc. At the moment I don't own anything interesting and am busy with other pursuits.

Tesla's FSD isn't the game changer that Tesla fans think it is; it's basically glorified cruise control. Driving a car isn't so much a physically demanding exercise as it is a mentally demanding one, and eliminating the need to operate the steering wheel only offers about as much value as eliminating the need to work the pedals. Even that wasn't as big of a game changer as automatic transmission; while both would become all but universal around the same time, for decades people were more willing to pay thousands of dollars for optional AT than to pay hundred for optional cruise control. And the demise of the manual transmission (in the US at least) was due to lack of consumer interest, while cruise only became universal because the adoption of electronic throttles on all vehicles meant that adding it was as simple as adding a button. But I digress.

The point is that until the mental aspect of driving can be automated, any autonomous system will be of limited utility. You mention the high loss of life due to car accidents. It would certainly be a good thing if autonomous vehicles could reduce this, but the dangerous nature of cars in general makes it difficult to fully automate. We currently have a system where each driver is almost 100% responsible for accidents caused by their cars. There are situations like Firestone Exploding Tires™ where car manufacturers have been held responsible, but these are edge cases, not garden-variety accidents. Removing the mental aspect from driving means removing any liability from the driver and shifting it to the manufacturer. Manufacturers will not accept this.

Consider a typical case where A & B are involved in an auto accident, with each sustaining injuries. The way it currently works is that a police officer would visit the scene and conduct a preliminary investigation and may make a preliminary finding of fault. If he determines that B is responsible, B will enter a claim with A's insurance company, and the insurance company will make an offer to settle. It's highly unlikely that they would contest the preliminary fault determination at this stage, but they may balk if A claims damages for injuries that they don't think are justified. If A is unsatisfied with the insurance company's offer, he can hire an attorney and sue B. At this stage, the insurance company will hire a lawyer to represent B, and the parties will take depositions and look at medical records. The accident report may be used as a guide, but A's lawyer will try to develop evidence to establish B's liability independent of the police report. B's lawyer may contest liability if the facts lean in that direction, but it's more likely that he'll try to challenge the damages to secure a low settlement amount. A has to prove that B was negligent. This basically means that there's a reasonable amount of caution that all drivers are expected to take, and that B failed to take the appropriate amount of caution. This could be because he was driving too fast, playing with the radio, or any number of other things that we all agree that careful drivers don't do. A trial is possible, but in normal circumstances once discovery closes both sides will negotiate a number that everyone can agree on, which will be within the limits of B's insurance policy. If A's damages exceed B's coverage limits, it's likely that A will settle for the maximum amount covered, unless B is independently wealthy and has significant assets. It goes without saying that since B is at fault in this example, he doesn't sue anyone. If he has collision insurance he will get compensation for damage to the car, but he won't get anything for medical expenses or pain and suffering.

Now let's assume A and B get in the same accident except both are in fully autonomous vehicles, A driving a Tesla and B driving a Honda. Both parties will file suits against Tesla, Honda, and whatever dealerships or subsidiaries are involved. This is not a negligence case but a product liability case. Neither plaintiff will have to prove that any defendant violated a standard of care, only that the product was defective and was a substantial contributing factor to their injuries. This will be easy for them to do; if two cars collide it is obvious that one of them did something wrong, and since neither driver had any control over his vehicle by design, the problem must be a defective vehicle. In theory, each defendant could conduct a forensic investigation and make the argument that the other car was defective and responsible for the crash. Realistically, unless it's a high value case or it's obvious that one car did nothing wrong, this is unlikely to prove anything that justifies the cost, and both parties will settle. The plaintiffs' attorneys are going to expect to get full value in these cases since insurance limits are unlikely to be an issue and, even if they are, the companies have deep pockets.

Each year, there are over 6 million reportable accidents in the United States. The average insurance payout per accident is around $30,000. My legal practice focuses on product liability, and there is no system I'm aware of where a solvent corporate defendant streamlines liability claims the way insurance companies do, so each of these accidents is going to generate a lawsuit. Actually 2 lawsuits, since both parties will be making claims now. And the effect of splitting claims among multiple defendants and not having to worry about insurance limits means the average payout will be higher. And there are legal fees to contend with, too. So assume that $30,000 becomes $40,000, and tack on $10,000 in legal fees, then double it to account for both cars involved, and we get to a nice round $100,000 per accident. (I'm keeping things simple and ignoring the effect of multi-car pileups, single car accidents, injuries to passengers, and any number of other things). More than triple the current amount, the cost shifted from drivers to auto manufacturers.

Tesla has already been successfully sued in relation to their cars' autonomous capabilities, and in one case the plaintiff was awarded an eye-watering judgment. I don't want to talk too much about those cases though, because they aren't really relevant to what I'm saying here. Those cases proceeded on a theory that Tesla failed to adequately warn the user and breached various implied warranties. As such, Tesla was at least afforded the opportunity to argue that the drivers had been adequately warned of the system's limitations and that they were responsible for those accidents by failing to heed those warnings. One can argue over whether Tesla's marketing creates a misleading impression of the capabilities of their products. What is beyond argument is that, in the case of a true autonomous system, Tesla, or anyone else, will have to end the throat clearing and market it on the basis that one can treat it like a personal taxi and ride even if they're asleep, or drunk, or don't have a license. Regardless of technical abilities, you have to add in the additional cost of taking on an additional $600 billion of added liabilities if these things become universal. That's a huge cost for one industry to bear.

One obvious argument is that self-driving cars are safer than manual cars. Okay, but how much safer? Remember, they have to take on approximately triple of what is currently paid out in insurance and lawsuit settlements for accidents based in the US. Reduce that number by 50%, or 80%, and it's still huge. And it underestimates how safe cars actually are. The 6 million+ reportable accidents a year sounds like a lot. But that's in a country that drives over 3.3 trillion miles annually. It's one accident for more than a half a million miles driven. If you drive 15,000 miles per year every year between ages 18 and 85, you can expect to be involved in two reportable accidents in your lifetime. Any assumptions about the relative safety of Tesla's current offerings versus normal cars is based on a woefully inadequate sample size. I'm sure the thing is impressive, but how many miles have you driven it so far? 10,000? 20,000? I'm guessing fewer than even that, but even if you drove it 50,000 you'd still have to drive it more than 10 times as far before we'd expect even one accident.

The final question we have to ask is what accidents are we concerned about preventing? I ignored this for the sake of simplicity, but the statistics I gave are for reportable accidents, while the average insurance payout includes all accidents. I only used reportable numbers because these are more serious, but in a large percentage of these, the police are only involved because one of the cars is undriveable. I was in two accidents in my life when my car was totaled, but other than the hassle of having to get a new car and deal with insurance, I was completely fine. I don't want to suggest that we shouldn't worry about these kinds of accidents at all, but they're certainly less of a priority than ones that cause serious injury or death. There are obviously solutions to these problems; I don't want to suggest that the liability questions are completely intractable. After all, we're already paying for motor vehicle insurance, and the cost will simply be shifted to another entity. Pay Tesla or pay Nationwide, take your pick, it's all the same. The overall point isn't that companies can't do this, it's that they won't. None of them are confident enough in their products at present to make this kind of guarantee, and consumers used to doing things a certain way will balk at the high premium they have to pay to cover the insurance premiums of a third party, even if they've technically been doing so for every company they've ever bought a good or service from.

One obvious argument is that self-driving cars are safer than manual cars. Okay, but how much safer? Remember, they have to take on approximately triple of what is currently paid out in insurance and lawsuit settlements for accidents based in the US. Reduce that number by 50%, or 80%, and it's still huge. And it underestimates how safe cars actually are. The 6 million+ reportable accidents a year sounds like a lot. But that's in a country that drives over 3.3 trillion miles annually. It's one accident for more than a half a million miles driven. If you drive 15,000 miles per year every year between ages 18 and 85, you can expect to be involved in two reportable accidents in your lifetime. Any assumptions about the relative safety of Tesla's current offerings versus normal cars is based on a woefully inadequate sample size. I'm sure the thing is impressive, but how many miles have you driven it so far? 10,000? 20,000? I'm guessing fewer than even that, but even if you drove it 50,000 you'd still have to drive it more than 10 times as far before we'd expect even one accident.

You don't have to speculate. You can just look at the numbers. 82% fewer crashes. Hundreds of millions of miles driven. Sure, a few hundred million is a lot less than a trillion, but this isn't untested technology and we have some idea of what kind of safety improvements are possible. Now, keep in mind:

  1. Not all of the remaining 18% crashes are Waymo's fault.

  2. Tripling the payout per accident and then reducing the rate of payouts by 82% means total payouts are 50% lower even assuming all those crashes are Waymo's fault.

That's Waymo, not Tesla. Their vehicles aren't available for consumer purchase, but even if they were, they would be prohibitively expensive, and they're only designed to operate in geofenced areas that have been extensively mapped with a team of supervisors on hand to navigate tricky situations. But that's orthogonal to the main point, which is that they put their money where their mouth is; if you get hit by a Waymo, the company isn't going to blame the accident on the passenger. Of course, the reason they are able to do this is they have the most advanced cars currently possible and exercise a high degree of control over them. I don't think the situations are comparable to a company like Tesla that is manufacturing cars and selling them.

Have you driven a Tesla with FSD or is this just theory? I find it much more relaxing. I'm thinking of trading in my practically brand new old Honda to get a second Tesla. But everyone's different I suppose.

I agree, lawsuits are a problem. It's probably the case that indigent people without licenses and insurance cause a huge percentage of serious accidents. But they have no money and no insurance. So there's no one to sue for wrongful death. Even for insured drivers, minimum coverage is often like $25,000. Judges can often be quite lenient in these cases. "I need my car to get to work your honor". Every once in a while you'll hear about someone getting their 15th DUI and they're still walking free. Which is crazy. I could drive drunk every day of the year and probably never get a DUI. For someone to get multiple DUI's imagine what kind of shit they must be pulling.

So, it would be a shame if trial lawyers indirectly kill thousands of people each year to perpetuate an "inadequate equilibrium" as Big Yud would say. Probably red states will do better than blue states in this regard and close the life expectancy gap a little.

For now at least, I suppose it's a minor concern. The worst drivers are unlikely to buy a $50,000 Tesla. The biggest beneficiaries would probably be seniors who have money and are smart enough to know they drive like shit. I'd get a Tesla for my parents if they weren't libtards.

Tesla has already been successfully sued in relation to their cars' autonomous capabilities, and in one case the plaintiff was awarded an eye-watering judgment. I don't want to talk too much about those cases though, because they aren't really relevant to what I'm saying here.

If you're talking Benavides v. Tesla, I think they have more relevance than that. A jury being willing to give massive punitive damages to a defendant that was not actually selling full-self-drive (the crash was 2019 and only on a car equipped for 'autopilot', which was a highway-only capability), and was not actually selling the full autonomy, where the driver was speeding, _and which presented evidence that the driver had their foot on the accelerator_ is not going to care about throat-clearing, as a matter of actual demonstrated fact.

A company faced with the revelation that, at best, they will be liable for the dumbest possible actions of their dumbest possible customer regardless of their actual product sales or familiarity with the marketing material that supposedly drove his behavior (and, while not demonstrated here, from other cases will apply even if the "customer" did not actually buy the product from the manufacturer!) has strong incentives to make a powerful product regardless so long as the AI is smarter than Mr. McGee.

((In reality, the trial happened in 2025, and if you want to pretend that politics played no role, I've got a bridge to sell you. In actual practice, PLCAA, GARA, pharma equivalents, and manufacturers moving themselves out of liability range tell you what the actual ramifications are, depending on scale.))

Court documents:

  • 2025-06-26: Judge (1) summarizes underlying facts, and (2) dismisses claims for defective manufacture and negligent misrepresentation, but (3) refuses to dismiss claims for defective design and failure to warn, plus punitive damages

    Before he reached the intersection of Card Sound Road, McGee had activated the Vehicle’s Autopilot, including the Traffic Aware Cruise Control (“TACC”), which among other things, restricts the Vehicle’s speed to 45 miles per hour when the Vehicle is not operating on a highway or limited access roadway. However, McGee subsequently manually engaged the Vehicle’s accelerator, increasing the Vehicle’s speed to 62 miles per hour and temporarily disengaging the TACC speed restrictions while leaving certain Autopilot features operational. The Parties dispute which specific Autopilot features remained active once McGee pressed the accelerator. Most notably, the Parties disagree whether Autopilot’s longitudinal control function and the automatic emergency brake function were deactivated in the moments leading up to the collision.

  • 2025-08-01: Jury instructions

    Plaintiffs claim that punitive damages should be awarded against Tesla based on its development of Autopilot and its response to its knowledge of accidents occurring while Autopilot was in use. Punitive damages are warranted against Tesla if you find by clear and convincing evidence that Tesla was guilty of intentional misconduct or gross negligence, which was a substantial cause of damage to Plaintiffs. Under those circumstances you may, in your discretion, award punitive damages against Tesla. If clear and convincing evidence does not show such conduct by Tesla, punitive damages are not warranted against Tesla.

  • 2025-08-01, later: Jury finds Tesla 33 percent liable for 129 M$ of injury (i. e., imposes 43 M$ of compensatory damages), and imposes an additional 200 M$ of punitive damages

  • 2026-02-20: Judge refuses to overturn jury verdict

    After a careful review of the parties’ respective arguments, the Court finds that the Motion must be denied. The grounds for relief that Tesla relies upon are virtually the same as those Tesla put forth previously during the course of trial and in their briefings on summary judgment—arguments that were already considered and rejected. Furthermore, Tesla does not present additional arguments or controlling law that persuades this Court to alter its earlier decisions or the jury verdict. Therefore, because the Court finds that the evidence admitted at trial more than supports the jury verdict and does not find that it committed any error affecting Tesla’s substantive rights, the interest of justice does not require the grant of a new trial or judgment in Tesla’s favor on either Count I, Count II, or the jury verdict regarding punitive damages.

The appeal still is in progress.

Yeah. Will add that I'm really not a fan of the judge's final decision spending a lot of ink on the 'hey, this is our first opportunity to bring this up, and the penalty here is well outside of the statutory bounds and separately raises 8th Amendment concerns' and then just not discussing the matter at all.

It appears clear to me that the insurance market to handle AI product liability claim will emerge very fast. Just as other methods ensuring such liability will be limited to "reasonable" sizes - either by contract, or by lobbying, or by corporate lawyering (aka "don't sue me, sue this cardboard cutout that legally is responsible for everything") so the industry will find ways to deal with it. Also: If real genuine honest FSD ever appears, I think it will drop the accident volume very significantly, because most of the accidents are caused by people being distracted, tired, drunk, stoned, or plain stupid. If we can replace Joe who just wants to get home from the pub and is absolutely sober of course after only a mere sixpack of beer, and Joan who can perfectly drive, check her email, converse on facebook and film her latest tiktok, thankyouverymuch, with robot driver that has the danger level even of an average attentive driver, the rest of the accidents - caused by mechanical failures, road malfunctions, severe weather events, etc. - I think would be quite manageable. FSD doesn't even really need to be perfect - there just needs to be enough FSD vehicles around so that the likelihood of somebody behaving in a stupid way and requiring everybody else to be perfect is low. All current stats include drunkards and idiots. How would they look without them?

Speaking of Teslas: I've probably not going to buy a Tesla anytime soon, but I've been offered Teslas by rental shops several times and I've so far declined, because driving an unfamiliar car in unfamiliar environment is kinda risky. So I wonder what's the experience of people that did drive it - is there any adjustment from driving a regular car that is needed? Can I just start it up and freakin drive or do I need some adjustments? How easy/hard is to get a hang of it from 0 experience?

Also, if I try it as a rental, what about range anxiety? I sometimes need to drive up to ~100 miles or more in such situations, of course mostly through urban-suburban landscape so there's always a gas station within 5-10 mins drive, but I have no idea about changing a Tesla. I remember some years ago I was out on a trip with friends, and the guy responsible for transport picked an EV (not Tesla, don't remember which one) and it turned out a minor annoyance to change it because a lot of nearby chargers had some kinds of "guests only" or "clients only" policies. I wonder what's the score on that now.

If you just want to get a feel for driving one, pop into your nearest Tesla showroom and take a test drive. They're usually very accommodating, they don't handle sales directly, and they don't work on commission.

The controls are a bit different, though highly automatic and generally intuitive for someone who is familiar with screen UI. For example, changing driving modes is done with a selector on the screen or overhead buttons, but is usually not necessary because simply buckling in and stepping on the brake will select an appropriate gear for you. Same with Y-turns: turn the wheel hard and brake facing a road edge, and it will automatically select reverse; reverse, turn the wheel hard and brake, and it will automatically select drive. Most other functions are either automated, like headlights, wipers, and door locking, or have pretty obvious controls. If you are a control freak and don't like automatic functions, you will probably find the controls frustrating.

The accelerator operation is much different than in a standard automatic transmission and closed to driving a manual transmission with a high-compression engine. As soon as you lift the accelerator, the car will start to slow. Lift all the way and the car will slow as much as medium braking does in an automatic, so in most driving it is not necessary to use the brake at all. It takes a little getting used to, but not having to switch pedals constantly is a pleasure. The accelerator feels much more closely related to speed than power, as opposed to a gas automatic.

On longer trips, providing you input your destination, the navigation system is very good about getting you to charging locations, so that is rarely an issue. Even the standard rage is typically well over 200 miles, so you shouldn't have to stop more than once every couple of hours, and charging from a low state of charge to around 75% usually takes barely more time than to use the restroom, get a snack, and stretch your legs a bit. The one issue I've heard of with rentals is that they expected you to return them with the same state of charge as when you left, which can be inconvenient if it was very high, as it means you have to charge somewhere close to the rental location and charging to very high states of charge takes much longer than going from low to around 75%.

I probably wouldn't rent one. The battery can degrade with non-ideal charging. I've heard horror stories of people renting them and then having to stop at superchargers every 100 miles because of cold weather and a battery degraded by abuse.

There's also a weird quirk with EV's where taking your foot off the gas triggers regenerative breaking. So you can't "coast" to a stop. Taking the foot off the gas feels like breaking. You have to keep pressing the gas, just more lightly. Takes some getting used to.

Finally, all the controls are inside the touchscreen panel which is annoying. You can use Grok to do anything and it works amazingly but I'm not sure if you'd have to set this up on a rental.

I wonder if Tesla will set up a rental program for Tesla owners where I can keep all my configs and use self-driving. That would be sick.

Can I just start it up and freakin drive or do I need some adjustments?

Do you have any experience driving an EV? If not, you need to adjust your expectations with regards to the instant torque you get with them

I drove a hybrid, many times (which made little difference), but I don't think I've driven an EV. Probably need to figure out a way to practice on one in a safe environment, just in case...

I've used it a few times. It's nice for quick trips (under 5 minutes) but for longer stuff I find it actually more mentally draining because I'm forced to pay attention but am not keeping myself busy with driving. I think it's an ADHD thing. I really want to either be completely driving or checked out and playing Cyberpunk 2077, but with FSD supervised I kinda need to watch someone else drive and it's hard not to let my attention wander.

I use it almost every time I drive. It's not perfect, but almost certainly better than me. I did manage to buy it outright just before they discontinued that option. You still need to pay a little attention, but you can definitely just stare out the windshield and zone out.

Haven't ridden in a Tesla with FSD in a while, but I have ridden in Waymos a number of times when visiting cities that have them. It's an impressively good experience IMO, though the price is not really noticeably lower than Uber etc. Naturally, supposedly super-progressive NYC (where I currently live) seems to be dragging their feet on approving them to operate here, while many redder cities seem to be much more enthusiastic about it.

Waymos are totally awesome, I always use them in cities which have them. All the convenience of Uber without some annoying guy trying to chat with you. It's a bit pricey but so far I can afford it, fortunately, and sometimes wait times are a bit long in busy places, so not ideal but quite good.

I guess right now the situation serves everyone except the customer. Taxi unions and cities don’t complain about technological unemployment, Waymo pads its margins, and Uber and the other companies don’t have to borrow billions at high rates to finance self driving fleets to compete with each other in a race to the bottom (which would follow full deregulation and which would drive down margins precipitously).

Waymo wait times are IME very high during peak hours despite the prices they charge being not much different than Uber/Lyft. It's quite clear Waymo has a car shortage.

I did happen upon this link trying to check out where they were really operating and why. Apparently, though they officially operate in ~10 cities, the total number of cars they have in most of them are "several dozen". Only in Phoenix, LA, and San Francisco do they have 500+.

People really hate paying $100/month for software products. It feels like you're getting ripped off even though it's literal Jetsons tech.

I hate paying $100/m for pretty much anything. If it's not consumable/renewable (like water, food, power) I'd rather buy it upfront. I know how much money I have now, so I can figure out if I can afford it, but the future is always risky.

Literal Jetsons tech or not, I like to own things, not rent things. Though in fairness they used to offer that as an option, not sure if they still do.

Isn’t it still illegal to actually let the car drive itself? You can’t go to the back of the car and fall asleep, you can’t even officially use your phone with both hands to watch movies or play games. You can’t even look out of the window and forget about what’s in front of you. (You can do these things in a Waymo, but not in your own self driving car).

Yes. But the experience of driving is just so much better and more relaxing. It's one of those things you just have to try. My wife is completely sold as well.

I'm waiting for the F in FSD to expand past the RoboTaxis. I'd love to just read a book while software does all my driving, but AFAIK that's not allowed yet, and if I have to watch the road anyway I might as well have fun and stay more attentive by driving while I do.

An acquaintance of mine who lives closer to the city got a Tesla last year and took me for a ride to show off its FSD functionality. It mistook a crack in the asphalt for the median line and sent us swerving into the opposite lane. Thankfully it was a back road in fuck-off nowhere so we were the only ones on the road, but it did make me hesitant to rely on the technology right now. It probably works better in places closer to the core.

I think there was a phase change at some point last year where required interventions went down by 90%, making it much more useful.

IMO, the same thing happened with coding in the last 12 months where at this point I don't even bother to check if Astra's code is correct since it always is.

I have considered it, but I like my truck.

Tesla really needs to come out with a light pickup, a minivan, and a lay flat camper van.

What is your favorite dungeon crawler OST and why is it King’s Field IV?

What is this place?

This website is a place for people who want to move past shady thinking and test their ideas in a court of people who don't all share the same biases. Our goal is to optimize for light, not heat; this is a group effort, and all commentators are asked to do their part.

The weekly Culture War threads host the most controversial topics and are the most visible aspect of The Motte. However, many other topics are appropriate here. We encourage people to post anything related to science, politics, or philosophy; if in doubt, post!

Check out The Vault for an archive of old quality posts. You are encouraged to crosspost these elsewhere.


Why are you called The Motte?

A motte is a stone keep on a raised earthwork common in early medieval fortifications. More pertinently, it's an element in a rhetorical move called a "Motte-and-Bailey", originally identified by philosopher Nicholas Shackel. It describes the tendency in discourse for people to move from a controversial but high value claim to a defensible but less exciting one upon any resistance to the former. He likens this to the medieval fortification, where a desirable land (the bailey) is abandoned when in danger for the more easily defended motte. In Shackel's words, "The Motte represents the defensible but undesired propositions to which one retreats when hard pressed."

On The Motte, always attempt to remain inside your defensible territory, even if you are not being pressed.


New post guidelines

If you're posting something that isn't related to the culture war, we encourage you to post a thread for it. A submission statement is highly appreciated, but isn't necessary for text posts or links to largely-text posts such as blogs or news articles; if we're unsure of the value of your post, we might remove it until you add a submission statement. A submission statement is required for non-text sources (videos, podcasts, images).

Culture war posts go in the culture war thread; all links must either include a submission statement or significant commentary. Bare links without those will be removed.

If in doubt, please post it!


Rules


Recommended Posts And Communities

Recommended Realtime Chats