glaz
I was at a public park not long ago when one of my children, who had only just begun to toddle, wandered about fifteen feet away from me. Not a big problem, I thought, and of course I was keeping an eye on him. On the far side of the park, at least a minute's walk away, a young woman showed up with a big dog and let it off the leash. It slammed across the park faster than I could believe, a missile headed right at my child. I know dogs, I've raised dogs, I've hunted with dogs, and I've worked with professional hunting dog trainers. This dog was trying to kill my baby, and it was so fast I almost couldn't react in time. Only my experience saved my child. My wife just watched with a glazed expression as all this played out. She does not know dogs. Anyway I was able to get close enough in time and yell and managed to get the dog to swerve at the last minute and back off while I scooped up my kid. Then I prepared to fight it to the death as it gave every indication of being about to try to jump up and snatch my kid out of my arms, which I've seen pit bulls do in videos, so I was ready. I kept yelling and there was a bit of a standoff until finally the owner showed up and leashed the dog. She seemed flustered and mostly wanted to avoid acknowledging what had just happened, and quickly left.
It was a terrifying experience. Dogs are not casual objects of entertainment or companionship. Modern people are so divorced from the realities of animal husbandry that I'm amazed we don't have more horrific catastrophes as a result.
I still take my kids to that park, but now I'm a helicopter parent in a way I never expected to be. At least for the smallest ones.
The_Nybbler
If you win the rat race you're still a rat. But you're also still a winner.
7mo ago
It's even less restrictive than that; you can have floor to ceiling glazing (not uncommon in fixed windows and sliding doors) provided the glass meets the hazardous location standards. That standard isn't about people falling out of open windows, it's about breaking through closed ones.
There's another rule about that, R321.2.1. The minimum is 24 inches (0.61 meter), from the floor to the opening, unless a guard is provided. Obvious thing to do if you want a window lower than that is to make the top panel the operating one, and I think I've seen that. As long as it's above 18 inches you can use regular glass.
Addressing your second point, as someone who knows next to nothing about economics [^1]: your question seems to be answered by the parable of the broken window.
Suppose it cost six francs to repair the damage [to a broken window], and you say that the accident brings six francs to the glazier's trade – that it encourages that trade to the amount of six francs – I grant it; I have not a word to say against it; you reason justly. The glazier comes, performs his task, receives his six francs, rubs his hands, and, in his heart, blesses the careless child [who broke the window]. All this is that which is seen.
But if, on the other hand, you come to the conclusion, as is too often the case, that it is a good thing to break windows, that it causes money to circulate, and that the encouragement of industry in general will be the result of it, you will oblige me to call out, "Stop there! Your theory is confined to that which is seen; it takes no account of that which is not seen."
It is not seen that as our shopkeeper has spent six francs upon one thing, he cannot spend them upon another. It is not seen that if he had not had a window to replace, he would, perhaps, have replaced his old shoes, or added another book to his library. In short, he would have employed his six francs in some way, which this accident has prevented.
Applying this to your question about the security guard: any society in which stores (and in particularly bad cases, individual families) must spend on hiring security guards is a society where this money is not being spent on research and development, or on education, or on infrastructure, or on other investments that generally raise the GDP of that society (and often make life better in that society too). We should thus expect to see this opportunity cost of hiring security guards to be reflected in GDP figures, as societies that hire them are more likely to be beset with lower GDP. This is borne out in reality: there are many developing countries where elites live behind expensive walled compounds staffed by large security details, but no one particularly thinks that they’re major players in the world economy.
[^1] That is to say, don’t put too much stock in what I’ve written here.
RandomRanger
Just build nuclear plants!
8mo ago
Video game NPCs can't have conversations with you or go on weird schizo tangents if you leave them alone talking with eachother. They're far more reactive than dynamic. This is a pretty weird, complex output for a nonthinking machine:
https://x.com/repligate/status/1847787882896904502/photo/1
Sensation is a process in the mind. Nerves don't have sensation, sensors don't have sensation, it's the mind that feels something. You can still feel things from a chopped off limb but without the brain, there is no feeling. What about the pain people feel when they discover someone they respect has political views they find repugnant? Or the pain of the wrong guy winning the election? The pain of a sub-par media release they'd been excited about? There are plenty of kinds of purely intellectual pain, just as there are purely intellectual thrills. I see no reason why we can rule out emotions purely based on substrate. Many people who deeply and intensively investigate modern AIs find them to be deeply emotional beings.
I dispute that the Britannica is even giving me more complex or more intelligent output. It can't use its 'knowledge' of the 7 years war to create other kinds of knowledge, it can't make it into a text adventure game or a poem or a song or craft alternate-history versions of the seven year's war. The 'novel tasks' part greatly increases complexity of the output, it allows for interactivity and a vast amount of potential output beyond a single pdf.
A more accurate analogy is that anti-AI image software interferes (or tries to interfere) with AI learning, not the actual vision process. It messes with the encoding process that squeezes down the data of millions and billions of images down into a checkpoint files a couple of gigabytes in size. I bet if we knew how the human vision process worked we could do things like that to people too.
I did a quick sanity test and put an image from the Glaze website into Claude and asked for a description. It was dead on the money, telling me about the marsh, the horse and rider, the colour palette and so on. So even if these manipulations can interfere with the training process, they clearly don't interfere with the vision process, whatever is going on technical terms. So they do pass the most basic test of vision and many of the advanced ones.
RandomRanger
Just build nuclear plants!
8mo ago
I think an LLM could experience pain, even without a body. They can be unsettled if you tell them certain things, you can distress them. Or at least they behave as if they're distressed. Pain is just a certain kind of hardcoded distress. Heartbreak can cause pain in humans on a purely cognitive level, there's no need for a physical body. Past a certain level of complexity in their output, we reach this philosophical zombie problem.
The AI-tampering programs are a little bit like optical illusions, except targeted against having specific known programs being able to train on certain images. They can't stop GPT-4o recognizing what's in an image or comparing like with like, they were only designed to prevent SD 1.5 training on an image. Also, they barely even work at that, more modern image models are apparently immune:
https://old.reddit.com/r/aiwars/comments/12f9otc/so_the_whole_entire_glaze_ai_thing_does_it/
Shrike
8mo ago
Not perfectly but close enough to the human level that there's a clear qualitative distinction between 'seeing' like they do and 'processing'.
I mean – I think this distinction is important for clear thinking. There's no sensation in the processing. If you watch a nuclear bomb go off, you will experience pain. An LLM will not.
Now, to your point, I don't really object to functionalist definitions all that much – supposing that we take an LLM, and we put it into a robot, and turn it loose on the world. It functionally makes sense for us to speak of the robot as "seeing." But we shouldn't confuse ourselves into thinking that it is experiencing qualia or that the LLM "brain" is perceiving sensation.
If you want to define seeing to preclude AIs doing it, at least give some kind of reasoning why machinery that can do the vast majority of things humans can do when given an image isn't 'seeing' and belongs in the same category as non-seeing things like security cameras or non-thinking things like calculators.
Sure – see above for the functionalist definition of seeing (which I do think makes some sense to refer casually to AI being able to do) versus the qualia/sensation definition of seeing (which we have no reason to believe AIs experience). But also consider this – programs like Glaze and Nightshade can work on AIs, and not on humans. This is because AIs are interpreting and referencing training data, not actually seeing anything, even in a functional sense. If you poison an AI's training data, you can convince it that airplanes are children. But humans actually start seeing without training data, although they are unable to articulate what they see without socialization. For the AI, the articulation is all that there is (so far). They have no rods nor cones.
Hence, you can take two LLMs, give them different training datasets, and they will interpret two images very differently. If you take two humans and take them to look at those same images, they may also interpret them differently, but they will see roughly the same thing, assuming their eyeballs are in good working condition etc. Now, I'm not missing the interesting parallels with humans there (humans, for instance, can be deceived in different circumstances – in fact, circumstances that might not bother an LLM). But AIs can fail the most basic precept of seeing – shown two [essentially, AI anti-tampering programs do change pixels] identical pictures, they can't even tell management "it's the same a similar picture" without special intervention.
Gaashk
8mo ago
I find that - Michelin starred restaurants aside - I can do better in thirty minutes in my own kitchen than pretty much anyone available on the apps.
I'm not in a position to have food delivered, but I find that almost any pre-prepared Costco meal is better than one I cooked (they keep up with the trends; they have birria now). We still cook from raw meat and root vegetables about half the time, but unless it's a taco or something, there's a marinade, some kind of eggs and crumbs or else cooked in a pan and deglazed, then some kind of roasting for one to six hours. The tacos are not bad, but also not better than from a food truck, and with less variety. I absolutely cannot cook proper beans, but I think it takes 8 hours and a piece of pork fat. We can't bring ourselves to eat enough beans to justify that.
RandomRanger
Just build nuclear plants!
8mo ago
I'm not convinced that people even need to put down the fork. I can eat as much as I want and exercise very little but remain thin. Mostly I don't eat ultra-processed food, I just eat whole food.
Formulations of ingredients, mostly of exclusive industrial use, made by a series of industrial processes, many requiring sophisticated equipment and technology (hence ‘ultra-processed’). Processes used to make ultra-processed foods include the fractioning of whole foods into substances, chemical modifications of these substances, assembly of unmodified and modified food substances using industrial techniques such as extrusion, moulding and pre-frying; use of additives at various stages of manufacture whose functions include making the final product palatable or hyper-palatable; and sophisticated packaging, usually with plastic and other synthetic materials. Ingredients include sugar, oils or fats, or salt, generally in combination, and substances that are sources of energy and nutrients that are of no or rare culinary use such as high fructose corn syrup, hydrogenated or interesterified oils, and protein isolates; classes of additives whose function is to make the final product palatable or more appealing such as flavours, flavour enhancers, colours, emulsifiers, and sweeteners, thickeners, and anti-foaming, bulking, carbonating, foaming, gelling, and glazing agents; and additives that prolong product duration, protect original properties or prevent proliferation of microorganisms.
Doesn't sound very appetizing! But it obviously is, ultra-processed food is 60% of US calorie consumption: https://www.cbsnews.com/news/ultra-processed-foods-calories-american-diet/
It seems very reasonable that eating things full of strange chemicals causes unusual health problems. Circus freaks from 1900 have nothing on the physiques you can see waddling around these days, they wouldn't even make it onto my 600 pound life. And the US is exporting this all around the world.
Tractatus
9mo ago
The continued glazing of Nate Silver, and the absurd belief in the validity of modern polling, betrays that the Rationalist/Rat-adjascent community is pathologically obsessed with appearing to be "scientific," at the expense of actually being right.
As I have pointed out ad nauseum, the shift to landline surveys has destroyed polling. No, Nate was not "less wrong" when he shifted his probabilities in 2016 to give Trump around 30%; there wasn't a single poll at the time that justified his change, but you lot still want to believe his model has any validity, and we'll be playing this same song and dance 4 years from now, and likely, until the end of the republic.
naraburns
nihil supernum
9mo ago
I think that glazing an individual user in this fashion in a modhat comment is inappropriate and reflects badly on the moderation.
I think what you're saying here is that my explicit endorsement of Dean is a bad look and makes you feel like you might not get a fair shake at some future point should you disagree with the wrong person. If I have understood you correctly, then you have failed to understand the foundation, or the moderation system, or maybe both.
I am not an impartial arbiter tasked with tone-policing the forum. My task is to cultivate "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." To that end, I wield exactly one carrot: AAQCs. I have two sticks: warnings and bans. Community sentiment (via reports) drives both. The community also has a small carrot (upvotes) and a small stick (downvotes).
This is a reputation economy: the more carrots you have, the less likely you are to get the stick. As we often remind people: that does not mean carrots are a perfect defense against sticks! But for example a user with many carrots might get a warning where a user with no carrots would get a ban. People who contribute to the good of the community are deliberately favored. We have never made the slightest secret of this, but everyone has to learn it for the first time sometime, so maybe today is your day.
Yes, I will freely admit that this sentiment is coloured by the circumstance that I cannot stand this particular user.
I appreciate the candor, so in turn I will freely admit that your comparing moderation here to Putin's Russia gave me a good laugh. It also helped me to calibrate on your sense of proportionality, in a way that was probably not beneficial to your aims.
4bpp
Now I am become a Helpful, Honest and Harmless Assistant, the destroyer of jobs
9mo ago
Yes. Dean is an excellent poster with an absolutely stellar history of making quality contributions to the Motte. He is probably in the top 5 userbase favorites. You, too, have made some good posts in the past, which is one of the reasons I haven't banned you yet. But if you're gonna rain on the AAQC parade any time your ox gets gored, I'll count it against you.
I think that glazing an individual user in this fashion in a modhat comment is inappropriate and reflects badly on the moderation. Yes, I will freely admit that this sentiment is coloured by the circumstance that I cannot stand this particular user. (I could expound at length why I would consider him to be a single-issue poster - as I see it, he is here to produce impassioned defenses of US neoconservatism with the same single-minded determination, attention to detail and absolute lack of interest in countervailing evidence as our most notorious JQ posters - but you have made it clear that you would not want to hear) Personal antipathy and feuds between users are a pretty normal sight here, though. Normally one would expect mods to act as a, well, moderating force on them - yet this sort of statement fills me (and presumably anyone else who would disagree with him) with negative levels of confidence that in the event of an interaction gone sour I would get a fair hearing. That is only moderating in the way Putin's rule is moderating opposition in Russia, which is to say it channels resentment into other outlets rather than reducing it.
I have continued to write the story I'm working on, albeit slowly. I'm currently over 12,300 words, which is nearly the length of Ted Chiang's Story Of Your Life, and I'm probably about a quarter of the way through so far.
Wondering what TheMotte's opinion on lengthy scientific exposition in sci-fi is. I currently have a big block of speculative biochemistry in the latter half of the current draft of the story, and some of my beta-readers... don't like it. I've tried to simplify it so it's understandable while still maintaining the necessary verisimilitude, but in general I get the feeling it might be too much. Personally, I've always liked large infodumps of speculative science in my fiction, the chapter Orphanogenesis in Diaspora with its detailed and lengthy descriptions of how the conceptory creates an orphan is probably one of my favourite openings to a story ever, but in general this kind of thing seems to make people's eyes glaze over.
Skulldrinker
9mo ago
It's a bit spooky how much he's being glazed up-thread.
bolido_sentimental
10mo ago
As a rule, if you don’t have preexisting physical health problems and make it through your first year or two(almost all tradesmen have to start on a construction site) then just take care of your body and the work won’t be too tough on it.
Just to supplement this. My dad has worked in a quite physical trade for 40 years now. (He's a glazier.) He has no particular ailments associated with it - he's a good weight, hale and healthy, still very physically capable. He's never been a overeater or a drinker, and I think getting lots of exercise each day has kept his level high. I'll be lucky to be as healthy as he is when I'm in my 60s.
The one problem he has is that he's had multiple melanomas removed, because he did not wear sunscreen at any point in all that time lol. He knows better, he doesn't deny it, but he still doesn't put it on.
FiveHourMarathon
Wawa Nationalist
10mo ago
Or, stated a little differently--these people are highly prone to losing what Rudyard Kipling once called "the common touch." [...]So in an attempt to be the change I wish to see in the world here's an object-level take: I feel bad for David French. I would say he has lost the common touch.
There's something kind of funny to me about accusing French of losing "The Common Touch" because of a disagreement on what is ultimately a pretty arcane constitutional provision. Seriously, I'm anti-term limits, but if some future Gibbon wrote the history of the decline and fall of the American empire, I can already feel the bored teenagers of some future century, their eyes glazing over trying to understand why this obscure fight over the appointment for certain bureaucrats was so pivotal to world history. It would be like trying to explain the intricacies of doctrinal disputes in medieval Christianity, the kind of thing that just seems monumentally obscure.
This isn't to say that liberals haven't lost The Common Touch, it takes a real galaxy brain to explain why the people burning down a Target are fighting for equality or something, you just can't explain that to a peasant. But it sorta feels like The Common Touch as you use it just means "agrees with me." The American common men are definitionally Conservative, and if they aren't then they aren't really American common men. The common touch is talking about immigration and inflation. It's talking about the constitutional right to bear arms. It ain't term limits.
They were on nobody’s agenda until after Neil Gorsuch and Brett Kavanaugh were appointed to the Supreme Court. We heard a quite different tune in 2016...An honest accounting would be frank about the fact that these proposals came about for only one reason: There’s a conservative majority on the Court for the first time since 1930, and liberals and progressives don’t think it’s legitimate for our side to ever get what their side has enjoyed in the past.
This feels off to me. Term limit proposals for SCOTUS were a debate in my AP US Gov textbook in 2008. They were picked up as a major policy proposal in 2020. But there's a long history of proposals for reform of SCOTUS terms.
I'm glad you acknowledged that Republican appointees have held the majority since 1970. Once again, a Conservative majority is defined by McLaughlin as "agrees with me." Particularly, agrees with McLaughlin on social issues to the extent he'd like them too. Ignoring the various other rulings made on a thousand other issues. As you note, Republican justices have historically drifted over time...which would be a really good argument for term limits? It would allow Republicans to refresh their appointees with fresh blood, rather than allowing a Kennedy to remain on the Court making mushy-headed legislation until he dies.
But at what point does ideological drift become a skill issue for the other major party? When you say:
Somehow, you can mismanage cities to the point of transforming San Francisco into an open-air sewer and still maintain total ideological dominance over the voting population. This sort of thing suggests to me that political competition just isn't happening at the object level.
Why are you granting the Democrats hyper-agency and turning the GOP into NPCs? The GOP held the Governor's mansion in California as recently as 2011. They've held the presidency for the majority of the last 70 years. Fox News, their partisan outlet, has been the top rated cable news channel for 22 consecutive years, and the top basic cable channel period for 8. And yet, let's rephrase your question:
Somehow, in a two party system, your opponents can mismanage cities to the point of transforming San Francisco into an open-air sewer and still maintain total ideological dominance over the voting population while you continue to lose every election. This sort of thing suggests to me that political competition just isn't happening at the object level.
Why is the GOP so incompetent that they can't get wins out of the supposed rank incompetence of Democrats? Is that Mr. McLaughlin and the National Review's fault, or are they just helpless passengers over at one of the major ideological organs of one of the two major political parties?
Then again, the NR folks have sure seemed to be helpless passengers against a certain short fingered vulgarian, so perhaps when they talk about conservatives finding themselves helpless against the least dirty trick from Dems, they're just describing themselves.
It's a coder's model I think, not a gooner's model.
I have no hopes for GPT in the latter department anyway, but my point stands, I think this is a remarkably mundane development and isn't nearly worth the glazing it gets. The things I read on basket weaving forums do not fill me with confidence either. Yes, it can solve fairly convoluted riddles, no shit - look at the fucking token count, 3k reasoning tokens for one no-context prompt (and I bet that can grow as context does, too)! Suddenly the long gen times, rate limits and cost increase make total sense, if this is what o1 does every single time.
Nothing I'm seeing so far disproves my intuition that this is literally 4o-latest but with a very autistic CoT prompt wired under the hood that makes it talk to itself in circles until it arrives at a decent answer. Don't get me wrong, this is still technically an improvement, but the means by which they arrived at it absolutely reeks of crutch coding (or crutch prompting, rather) and not any actual increase in model capabilities. I'm not surprised they have the gall to sell this (at a markup too!) but color me thoroughly unimpressed.
In Britain this is basically standard, for reasons which have been discussed elsewhere. New builds are rare and the extent to which modern upgrades (dishwasher, tumble dryer, central heating, double glazing) are available varies wildly.
What you have to remember that where mod cons were unavailable they were compensated for by other things. My granny didn't get air conditioning until a couple of years before she died because she had a permanently-fuelled coal-fired oven, and she spent the whole winter in the kitchen next to it. Add thick walls, blankets and jumpers and you're sorted. The only mod cons I have trouble doing without are hot water and washing machines.
I read Scott's article on Cost Disease once and I've never forgotten it. I think that lots of people would be happy with 1940s housing and education at 1940s prices (adjusted for inflation). Medical care not so much. Food is complicated, because the form, quantity, quality and satisfaction associated with it has changed in so many ways that it's not easy to pin the changes as wholly positive or wholly negative.
More options
Context Copy link