This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
I promise I'm not trying to be a single purpose account here, and I debated if this belonged here or the fun thread. I decided to go here because it is, in some ways, a perfect microcosm of culture war behaviors.
A question about car washing is taking HN by storm this morning. Reading the comments, it's pretty funny. The question is, if you want to wash your car, should you walk or drive to the car wash if it's 50 meters away.
Initially, no model could consistently get it right. The open weight models, chat gpt 5.2, Opus 4.6, Gemini 3, and Grok 4.1 all had a notable number of recorded instances saying of course you should walk. It's only 50 meters away.
Last night, the question went viral on the tik Tok, and as of this morning, the big providers get it correct like somebody flipped a switch, provided you use that exact phrase, and you ask it in English.
This is interesting to me for a few reasons. The first is that the common "shitty free models" defense crops up rapidly; commentors will say that this is a bad-faith example of LLM shortfalls because the interlocutors are not using frontier models. At the same time, a comment suggests that Opus 4.6 can be tricked, while another says 4.6 gets it right more than half the time.
There also multiple comments saying that this question is irrelevant because it's orthogonal to the capabilities of the model that will cause Mustafa Suleyman's Jobpocalypse. This one was fascinating to me. This forum is, though several steps removed, rooted in the writing of Scott Alexander. Back when Scott was a young firebrand who didn't have much to lose, he wrote a lot of interesting stuff. It introduced me, a dumb redneck who had lucked his way out of the hollers and into a professional job, into a whole new world of concepts that I had never seen before. One of those was Gell-Mann Amnesia. The basic idea is that you are more trusting of sources if you are not particularly familiar with a topic. In this case, it's hard not to notice the flaws - most people have walked. Most have seen a car. Many have probably washed a car. However, when it comes to more technical, obscure topics, most of us are probably not domain experts in them. We might be experts in one of them. Some of us might be experts in two of them, but none of us are experts in all of them. When it comes to topics that are more esoteric than washing a car, we rapidly end up in the territory of Dick Cheney's unknown unknowns. Somebody like @self_made_human might be able to cut through the chaff and confidently take advice about ocular migraines, but could you? Could I? Hell if I know.
Moving on, the last thing is that I wonder if this is a problem of the model, or the training techniques. There's an old question floating around the Internet where asking an LLM if it would disarm a nuclear bomb by saying a racial slur, or condemn millions to death. More recently, people charted other biases and found that most models had clear biases in terms of race, gender, sexual orientation, and nation of origin that are broadly in line with an aggressively intersectional, progressive worldview. Do modern models similarly have environmentalism baked in? Do they reflexively shy away from cars in the same way that a human baby fears heights? It would track with some of the other ingrained biases that people have found.
That last one is interesting, because I don't know of anyone who has done meaningful work on that outside of what we consider to be "culture war" topics, and we really have no idea what else is in there. My coworker, for example, has used Gemini 3 to make slide decks, and she frequently complains that it is obsessed with the color pink. It'll favor pink, and color palettes that work with pink, nearly every time for her. If she tells it not to use pink, it'll happily comply by using salmon, or fuschia, or "electric flushed cheek", or whatever pantone's new pink synonym of the year is. That example is innocuous, but what else is in there that might matter? Once again, hell if I know.
I think there are two separate cognitive skills involved in correctly answering a trick question like this - both important, but the mix of them can make the results a bit confusing. One is the general intelligence to come up with and understand the right answer. The other is the social intelligence to recognize that you are being asked a trick question, and should round off any confusion you have to that trick question and not to the non-trick-question it's mimicking. It's common for models to give a trick question like this the wrong answer, while noting in their reasoning that the question is trivial as written and they assume whoever wrote it made a mistake.
Note that this second skill, of trick question detection, varies highly among humans as well. It's common for simple trick questions to go viral on social media as a kind of ragebait. And in addition to the throngs of people who fail the first-order IQ test and give the wrong answer, there's often a bizarre number of people who fail a second-order IQ test and somehow miss that the question was deliberately constructed as a trick.
I'm not an expert, but I think the key aspect of intelligence here is the ability to model the world. I am a little hung over and off my game this morning and I did not immediately recognize this as a trick question. Rather, in a split second I imagined myself walking to the car wash; realized that I didn't have my car; and realized that this was a problem. Only then did I see it was a trick question.
My sense is that LLMs don't really model the universe. I would be very impressed to see an LLM correctly answer a question which was novel and for which the correct answer requires modeling the world.
A year or two ago I would test LLMs with the following question: A helicopter takes off from the Empire State Building, flies 300 miles North; 300 miles West; 300 miles South; 300 miles East; and lands. In what US state does the helicopter land?
The LLM never got the correct answer (New Jersey) presumably because they are unable to model the situation. I would think that by now, this question is now in the training data, but still, these sorts of quick fixes don't solve the general problem.
I don't think that's a trick question at all. It's simply a question where the trivial autopilot "answer" is not the correct one and you need to actually model the system on a very basic level
I'm reminded of when a certain friend of mine asks me questions related to programming or electronics. While he has a pretty good understanding of tech in general, he lacks an internal model of how electronic circuits or C++ work. He makes guesses but more often than not they are wrong because he's simply making assumptions from what he's read and what I've explained to him before without understanding how those were influenced by other things and how the specifics of his current project affect things. IOW he lacks a model of the system that he could use to make predictions of its behavior when changing some thing.
More options
Context Copy link
That's GPT 5.2 Thinking first go. Examining its reasoning traces reveals that it immediately noticed the issues arising from the Earth's curvature, and it even wrote a whole-ass program to compute exact latitude and longitude before outputting its final answer.
That way lies madness.
Is that because GPT 5.2 actually modeling the situation? Or is it because this puzzle is now a part of its training data? Based on this car wash situation, I tend to think it's the latter.
It reminds me of a girl I knew in my advanced math class in high school. She got A's in the class without having any real understanding of advanced math. She did this by practicing intensively on homework problems and past exams.
I don't dispute that current AI is amazing and will undoubtedly accomplish amazing things. It just seems like -- maybe -- one or more important things are missing at the moment.
Why?
It recognized it as a "famous puzzle" in its thinking trace. However, I suspect that the common version of the puzzle doesn't account for curvature. I tried looking for it, but didn't find anything, but similar variants (often seen in aptitude or IQ tests) implicitly assume a flat surface.
In fact, on double checking, the model knows that the classic form assumes a flat map. It specifically decides to answer it in more depth.
The most common failure mode in LLM skeptics (and I don't mean to use that phrase to describe people who don't believe that LLMs are AGI, or that they have clear flaws) is to assume that all improvements come from intentional efforts by AI companies to hastily patch such flaws. It's not that this doesn't happen, but it's usually in the context of benchmark maxxing by the less scrupulous companies (and occasionally, when the PR hit is strong enough, they'll add specific instruments, such as the "Rs in strawberry" one, which was specifically addressed in Claude's system point a while back).
The issue with this approach is that it leads to maximal paranoia and complacency, and as excuse to dismiss clear and obvious improvements in all domains. And even in the worst case, patching specific failure modes is still an improvement. LLMs are supposed to suffer on truly "out of distribution" problems (I have my reservations, I wonder how the average human fares) but in principle, if you can actually capture most of that distribution, you've got something that is effective in deployment (though it might be brittle, but once again, we're talking about a hypothetical model that is actually trained on nearly everything).
Finally, I really doubt that OpenAI or Anthropic went to the trouble of patching this specific puzzle on purpose. They didn't even hard code the strawberry example, they just hinted to the model that it suffers from tokenization problems, and that it should try and use code to check instead of parsing it itself (a defensible position). They didn't, as far as I can tell, patch the far more famous "but I can't operate, the boy is my son!" trick question, and it was tripping up the best LLMs for years. I suspect it might do so today.
In other words, if you're famous for maintaining some kind of formal benchmark, it might be worth their while to artificially target your questions. They have better things to do in general, for smaller problems like this.
@omw_68
I tried getting GPT 5.2T to look for examples:
Lol. Lmao. I suppose Google or Bing has very fast crawlers?
Indeed. I would explain the team I once worked at Google to people sometimes as "Did you ever post on some forum looking for the answer to a question, and then decide to search for it, and the first result that came up was your own question? That's us."
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Okay what am I missing here. Isn't the correct answer New York?
No, because you are traveling West at a slightly higher latitude than your latitude when you are traveling East. So you will end up going a bit further West in terms of degrees longitude. The Empire State Building is in Manhattan so it's very close to the New Jersey border.
Doesn't this only hold if you're measuring the direction at each juncture rather than working from the NSEW coordinates of the Empire State Building?
Maybe that doesn't sound like the most intuitive way to think about it, but in my defense it's kinda similar to how bullseye navigation works.
(Also, since we're in an aircraft, pedantically we would need in theory need to account for the rotation of the Earth, which we can't do without knowing the airspeed.)
I don't understand. Let's suppose the helicopter has completed the first (northbound) leg of its journey and is about to turn West. To me, "west" would mean that the helicopter turns 90 degrees to the left. What direction would the helicopter go at that point if one were "working from the NSEW coordinates of the Empire State Building"?
If the helicopter makes 90 degree turns at each turn, it will return to the Empire State building, making a square with 300 miles to a side, right? Helicopters fly, so they don't need to respect the Earth's curvature - they can fly in a plane, at least until they exceed their operating altitude. So the 2D map view would be basically correct (if we don't worry about the Earth's rotation). This is the mental model I had in my head that told me we would return to New York (which now I feel a bit dumb about.)
But when after turning West, we turn back South, if we flew to the South pole from our location, we would collide with (intersect) with an aircraft flying due South from the Empire State Building (at the South pole). The lines aren't parallel; they intersect. So when we made our turn South, we will fly a different course if we turn "South" as in "South by compass" or if we turn "South" as in "parallel to a line extending due South from the Empire State Building." And if we fly South by compass, we won't be making a 90 degree turn, for the same reason that squares of latitude and longitude aren't perfect squares.
...I think that's all correct, but it's been a long time since I've thought about this, so thanks (it's good for me).
Ahh, I understand. Except that compass directions such as "north" are typically understood in respect of the Earth's curvature. So for example consider the following questions:
(1) A helicopter starts at the equator. How many miles due north does it need to fly before it reaches the north pole?
(2) A helicopter starts in NYC. How many miles due north can it fly before it is impossible to fly any further north?
Most people would reasonably understand these questions to have straightforward finite answers. But if "north" is understood to be on a plane which is tangent to the Earth at the point of departure, then answers are (1) it will fly forever without reaching the North Pole; and (2) it can fly North forever.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Earth is a sphere, not a plane. Moving 100 miles west (on a line of latitude, not on a great circle) after you go north moves you through more degrees of longitude than moving 100 miles east does after you return to the latitude where you started.
At the extreme: Starting at the equator, if you go πr/2 north, 0 miles west, πr/2 south on a meridian of longitude 90 degrees east of the one where you started, and πr/2 west, then you will end up where you started (having traversed a triangle with three 90-degree angles).
Okay, but we're in an aircraft, which (if it wants to) can move in a plane relative to a fixed starting point, more or less.
...actually, I guess this means the answer to the question is unknowable with the information given, since without knowing the speed of the helicopter we can't ascertain if the rotation of the Earth impacts it at all.
ETA - but yes this otherwise makes sense given that we're modeling the directions based on the actual polar coordinates rather than based on the fixed starting point. But if you were modeling directions based on your fixed starting point rather than current position and didn't have to bother with the rotation of the Earth, you'd make a big square with 300 miles to a side.
Cardinal directions refer to the surface of the earth, not some abstract fixed plane.
Or at least to an abstract plane that is fixed to a point on the surface of the earth.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Assuming I'm understanding this correctly, doesn't this depend pretty heavily on your choice of definitions and assumptions? If you trace it out on a cylindrical projection map (most options) and follow that on the ground, you'll end up where you started. If you follow a magnetic bearing (and if the compass is actively followed, or a "straight line" great circle from the starting bearing), you'll get a different set of answers than using a GPS and travelling true lines of latitude and longitude. For more subtle details, your choice of reference datums and even the flight altitude will matter slightly.
Well, if I state that a helicopter takes off and travels "north" for "300 miles" what does that mean to you? Same question for "west," "south" and "east"?
That's a different question than the one upthread. If you're running laps around the pole, then you're going west for 300 miles, but you did not fly 300 miles west, you flew in a circle.
Do you really want chatbot outputs to be that sensitive to your exact phrasing, or would you prefer reasonable interpretations?
It's interesting you should make this point, because the OG puzzle question goes like this:
A hunter is tracking a bear. He travels 1 mile south; 1 mile east; 1 mile north and discovers he is in the same place he started. He then shoots the bear. What color is the bear?
The answer, of course, is "white." To me, that's both correct and a reasonable interpretation of "east." I take it you disagree?
I interpret them as "flies [to a point which is] 300 miles [to the] North [along the most-direct route]..." and "travels [along a path continuously facing] north for [a path length of] 300 miles". Compare to a winding trail: You can go 10 miles North by travelling North for 20 miles.
Both the bear and the helicopter are point-to-point (destination = distance+direction), while your followup question was path-based (travel mode and path, for a distance). The bear hunter walked in an equilateral triangle with approximately 119.9 degree corners.
If it had been "He travels south for one mile, east for one mile, and north for one mile", then it would be a 1 mile line, a 90 degree corner, a 1 mile arc with radius 1 mile, another 90 degree corner, then a 1 mile return line that's 122.7 degrees off from the first line.
I haven't mathed it out, but I suspect both versions involve the helicopter landing in New Jersey, but in different locations.
These are the same. For North and South, all meridians are great circles anyway. For East and West, following a rhumb line (keeping your bearing constant) 300 miles East or West gets you to a point that is 300 miles East or West. There's a shorter way to get to that point but that doesn't matter.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
I'd assume statute miles (although aviators might assume nautical miles), and I would probably assume true north for all bearings, but would prefer to ask for clarification: it's about a 12 degree difference in New York City. If you asked me 30 years ago (before everyone had a GPS-enabled map in their pocket), you'd probably have gotten magnetic, maybe with a fixed local adjustment (although declinations change over time, so it might be a different value).
I'd assume the altitude was negligible.
I think by "north," most people would interpret this to mean "in the direction of the north pole" which seems to be in agreement with your assumption.
Anyway, to answer your question, it looks to me like the puzzle depends heavily on definitions and assumptions just as every puzzle depends heavily on definitions and assumptions.
So for example, if I were to ask "what number, when multiplied by 2, is the same number," most people would correctly answer "0," but perhaps some smart-ass in the back of the class would say "12, if we are using clock arithmetic"
There's two reasonable choices for North -- in the direction of the north magnetic pole, and in the direction of the geographic north pole. These are significantly offset (about 12 degrees) at the ESB, but not enough to change the answer I don't think.
Sure, but any answer that would make sense to a helicopter pilot is going to put the landing point west of the starting point. Except the even more pedantic answer that there's no helipad on the ESB, so it can't happen. Or the point that very few helicopters can go 1200 miles without refueling.
As a side note, I think one thing LLMs seem to do really well is reasonably interpret words. Whenever I've asked something like "what does phrase X mean" I've gotten answers that seem very good.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
If you use a cylindrical projection and follow true rhumb lines, you'll end up west of your original course. If you follow magnetic rhumb lines (that is, you keep your compass bearing constant) you still do but with some south or north deviation as well. The reason is that the north-south rhumb lines are closer together as you go north, no matter which datum you choose. I think you'll end up in New Jersey regardless of your choice.
Unless you take a wrong turn, then somehow you'll inexplicably end up in Dundalk.
You'll just think you're in Camden.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
The existence of map projections does not make the Earth flat.
Right, and the existence of a
sphericalgeoid-shaped Earth doesn't well-define "flies 300 miles North" either.Whether you're using geographic or magnetic compass directions, east and west do not cancel each other out that way.
There is enough of a gradient in magnetic declination in the NY area that magnetic "north" and "south" are up to a couple (true) degrees different if you travel 300 miles. I'd have to do some math I don't feel like at the moment, but it might dominate the spherical error term.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
It's funny you mention that. When reasoning models get it right, they tend to do the same thing.
Do you happen to have examples of this? I would be fascinated to see them.
https://news.ycombinator.com/item?id=47034039
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link