domain:drrollergator.substack.com
SubstantialFrivolity
I'm not even supposed to be here today
1hr ago
I can't say I agree with that. Someone who is so violent and unhinged that they might shoot you if you look at their girlfriend wrong is not meaningfully more dangerous with a gun, in my opinion. They're going to get the job done no matter what, even if they just have their bare hands.
Tree
1hr ago
I don't know what you guys are talking about, "Viande de boeuf, viande de poulet" is very common french.
Shrike
1hr ago
How is your semi AR15 with a ten rounds mag going to fare against a predator drone or a tank? In the very best case, you would be fighting a protracted war against the federal government. If you win, it looks like Mao winning his civil war, if you lose, it looks like Hamas in Gaza.
First off, as a technical point, the ARs will have a lot more than ten rounds (30 round is the standard magazine, lots of people run drums with 50 or 100 rounds).
Secondly, gestures to Afghanistan the US army is capable of losing a war to an opponent with small arms and IEDs! I've never understood the "the US military would crush an armed populace" line of arguing because it had a chance to do that in the last two decades and failed. (And of course laying the blame on Iran or Pakistan or whoever is cope – do you think China or Russia would fail to arm insurgents in the US if there was a civil war?) What I find much more questionable an assumption is that the US armed populace would act like the populace in Afghanistan (or Northern Ireland) but if they did, it seems likely from history that the US armed forces would in fact lose. Wars are political endeavors and technology does not change that.
Thirdly, in most civil wars, the military and national security apparatus is not actually monolithic. Let's say that it's true for the sake of argument that the "armed populace" is not capable of "beating the US military" (I actually agree this is a fantasy because even if the "armed populace" could beat the US military on a giant featureless plain that's...not how real wars work.) In many, perhaps most civil wars, the military fragments alongside the rest of the populace. In which circumstance, it can be really helpful to have an armed populace even if there is no irregular warfare because they are likely to be better marksmen, more likely to be able to contribute to arms stockpiles, etc. In a prolonged civil war situation, the side with the support of the armed populace will be favored to win all else being equal. Which means there's a certain incentive for ideologies to promote firearm ownership (on their own team) and to attempt to convince the other side to disarm.
(As an aside, for this reason widespread firearms ownership is actually extremely beneficial to the US state. The US military recruits disproportionately from certain areas for reasons that are not but are correlated to firearms ownership.)
For example, I imagine that hand grenades are much fun. Or landmines. Watch the stupid coyotes explode when they trespass on your property. Contact poisons are fun. Radioactive substances are fun. So is building your own nuclear reactor.
All of these are, at least in the right circumstances, legal in the US, but the hand grenades and nuclear reactors at a minimum require paperwork.
MaiqTheTrue
Renrijra Krin
1hr ago
Maybe it’s because I live in rural Midwest but I just don’t get that worried about the guy with a pistol in a holster on his side. I’ve never once seen anyone pull a gun like that in public. Those guys are generally the responsible ones, the guy prone to shooting at people is not going to open carry because he wants to surprise people with the gun. Open carry doesn’t lend itself to sudden shooting or crime because as you mentioned everyone notices the gun.
Corvos
1hr ago
The gun massively increases your danger, surely? Firstly because it so hugely reduces the amount of effort he needs to put into damaging you, and secondly because it makes it so hugely more likely that the damage will be lethal.
SubstantialFrivolity
I'm not even supposed to be here today
1hr ago
But you weren't objecting to volatile drugged-up gangsters. You were objecting to guns. And of course, there is a major difference between a pet tiger and a gun in terms of whether you need to watch them carefully for danger.
And yes, I realize that part of your argument has been the inability to know whether any given gun-owner is unstable. But the unstable people are always a threat to you. The volatile gangster can quite easily stab you or beat the shit out of you, even were he to not have a gun. I don't think that him having a gun meaningfully increases the amount of danger you are in, so seeing a gun should not (imo) make you any more nervous than you would be around any crowd of people.
Corvos
1hr ago
I think he’s arguing that the argmax you run over the logits is not technically part of the LLM neural network so the LLM is just ‘an algorithm that produces math’ (ie produced a probability distribution), but that seems tendentious and also kind of weirdly put because it sounds like describing a tokenizer.
wayfarer
1hr ago
The problem that commentary was not even interesting, but dealing with fiction invented there. (fiction propped only by fake or worthless credentials)
celluloid_dream
1hr ago
Is there so much of a difference between a pet tiger that could maul you if you accidentally trigger its prey drive, and a volatile drugged-up gangster who thinks you were chatting up his girl?
TequilaMockingbird
Brown-skinned Fascist MAGA boot-licker
1hr ago
Because I read DeepSeek's 2024 paper.
Shrike
1hr ago
Man, don't shoot the messenger here.
Ha! No, like I said, that's definitely not my ethos. But I hear ya.
Either way, the thing is that the rule against killing is, again to a first approximation, fairly absolute; and to someone who actually believes in an absolute rule, asserting that you actually want to break it in a fairly broad special case is not persuasive.
Sure, this makes sense. And of course Americans often don't believe in this at all (even when it comes to executions and the like).
Can you muster the theory of mind to understand that some people actually believe that there are no "bad guys" who it is a good thing to kill?
Yes. And I think you're right, there's an incommensurability problem that plausibly is only worked out on civilizational timescales.
I'm not so convinced that they are strongly correlated at all - East Asia has ubiquitous AC but no guns and an atrocious free-speech situation compared to Europe as well, Russia flip-flops but at least intermittently had quite liberal gun laws with no relation to its AC or speech situation.
I am also not convinced on the correlation, but I will note that I think civilizations are very different and a causal chain that exists in some cultures may not exist in other cultures at all. Sometimes just the idea that something is true makes it so.
I also suspect Europe's free-speech situation is, at least in some respects and specifically in some places, about as bad or perhaps even worse than Russia's – it looks like England might be in some ways worse than Russia, arresting 12,000 people in 2023 while Russia detained about 20,000 people since 2022 as per this 2024 article as part of crackdowns on anti-war speech (note that these don't measure convictions, and of course note also that Russia has nearly three times the population, but also that the article I pulled was focused on the Russian anti-war crackdown and might not measure people taken in for other views.)
Either way, the heat death figures you refer to always seemed fairly cooked to me - Eurocrats have an incentive to inflate them to support the climate change narrative, while the US figure seems pretty inappropriately small for its burgeoning homeless population.
I think it's pretty rare for ~healthy adults to die from heat stroke (some of these numbers might be due to aging European demographics) and a lot of the American homeless are in pretty temperate places like California. I believe US cities generally have lots of places for homeless people to get out of the cold, or ways for them to travel to more temperate regions. If I had to guess, most exposure deaths among the homeless involve drugs of some kind. But that's a guess.
netstack
Texas is freedom land
1hr ago
Concealed carry isn’t about freaking people out.
It got popular in the 70s and 80s as part of the broader political struggle. Liberal, urban states adopted stricter policies and rural ones looked for ways to signal their opposition. Revoking old concealed-carry bans was one option.
Arguments about “not marking yourself as a target” or “strategic ambiguity” are secondary.
roystgnr
2hr ago
Try to not typical-mind as much.
I do, but I also try to be charitable, and the idea that you want to win a knife fight with an unarmed tackle wouldn't have passed that test even if it had occurred to me.
VoxelVexillologist
Multidimensional Radical Centrist
2hr ago
One of the biggest sources of illegal guns is parked cars.
I'd be curious what fraction of guns left in parked cars are left their because their owners can't legally keep them on their person where they've parked --- bars (depending on state), private property that disallows carrying, etc. I'm not going to blindly use that to push for legalizing carry everywhere (honestly, probably not my preference), but the numbers would at least be interesting.
Corvos
2hr ago
He’s saying that someone with a gun could kill you at any moment without even trying. The same is not true of, say, a rubber duck. It’s natural to be a lot more nervous around one than the other, even if the owner has not yet demonstrated ill intent or stupidity.
this is distinct from intelligence
I disagree, it is not distinct, it is integeral. What is the value that a human equipment operator (or any other "intelligence" for that matter) adds over a machine carrying out a set of scripted movements if not the ability to react and adapt on the fly?
Tell me do you want your "self driving car" to plow into the back of a stopped vehicle because it was programmed to drive south on [route x] at [speed] for [distance], or would you prefer that it percieve and react to the obstacle by applying the brakes and/or going around? Which of those options do you think is the more "intelligent" of the two?
Supah_Schmendrick
2hr ago
Come on, be charitable. It's not a perfect analogy. The point I'm trying to make is that it's a dangerous thing to be carrying around in public. It does require volition, but volition may be influenced by rage, or alcohol, or psychosis, or mental illness, or one bad day.
But that's a significant difference! You've moved the goalposts from "that's something that can kill if you don't concentrate on it sufficiently" (untrue, but would strongly favor your position a la "ultrahazardous activities") to the true argument of "but people are sometimes idiots, impaired, or negligent" which is a major shift with significant consequences!
Im sure with perfect adherence to a special diet plan you are correct, but as someone in the medical field I’m sure you’re aware there’s a wide gap between recommended use and typical use of anything. Typical vegan diets are not healthy for kids, and typical vegan diets are what the modal vegan kid is eating. Studies show the typical vegan kid is stunted, and that’s a bad thing.
It’s not good for the elderly either, where veganism is associated with risk for bone fractures, sarcopenia, anemia, and depression.
Maybe all of this could be eliminated with the perfect vegan diet. Maybe Real Veganism Has Never been Tried. I don’t really care, I only care about what empirical works for most people.
It's a curiosity because without principles, what makes someone choose any particular side to begin with?
And I’m answering: familial loyalties, suspicions of whether their tribe will benefit, aesthetic preferences, etc.
Who is having their goods delivered? Whose fault is it that what is happening is happening? Macroeconomics and the like are so nebulous, unreadable and unproven that you will find people’s opinions on the effect of price controls is strongly determined by their loyalties, and not the reverse.
I am come from an upper-class family, I went to the appropriate schools in the UK, I read the Soectator, etc. You could pretty easily predict my views on the merits of taxation and on the usefulness of the Laffer curve, my voting affiliation, my views on fox-hunting, on globalisation, all from those pieces of information.
Then Brexit happened and there was a big alignment but it’s amazing how you can predict people’s carefully worked out opinions on the results of certain policies once you know their class, gender, age and job.
netstack
Texas is freedom land
2hr ago
I would naively expect it to help, if only by making charging and sentencing easier.
Oregon tried that decriminalization experiment with drug possession. But it was hideously confounded by fentanyl, and I didn’t find any studies from the recriminalization last year.
Maybe there’s something in gang violence stats? Police have a longstanding interest in disarming gangs. It should be possible to tell whether changes in general gun policy, or even in enforcement, actually reduce gang shootings.
wayfarer
2hr ago
define an axis along which we can work to evaluate both animals and algorithms
trying to evaluate intelligence on one axis is going to take you nowhere, as made clear by your claimed results
perceptivity + reactivity spectrum
this is distinct from intelligence
and if asked "which is smarter, Grok, Claude, Gemini, or an orangutan?" I am going to pick the orangutan every time.
that just proves your intelligence definition is bad and not worth using
wayfarer
2hr ago
It's almost as if there must be some intermediate layer in between.
Can you explain what claimed layer is there?
After all, if all the LLM is doing is predict the next most likely word
that is with so called temperature parameter set to 0, in practice typically is set to LLM predict next likely token
how do you get a chess engine or python script out of that?
by training it on large pile of text first
No. They’re awaiting the opportune moment. Kind of like a SIG.
In all seriousness, the purpose of a thing can be divorced from its usage statistics. The vast majority of nuclear weapons have never killed anyone. Instead they work via the threat of fulfilling their purpose.
Guns are an effective threat against almost anyone. That makes them useful whether or not they actual kill.
More options
Context Copy link