Reading https://www.lesswrong.com/posts/nwJCzszw8gGjPTihM/i-still-think-it-s-very-unlikely-we-re-observing-alien and pondering the Bigfoot thing.

On the one hand, We Have Cameras Everywhere(TM).

On the other hand -- pick any area of the pacific northwest and look at a map of where the permanent roads are. Pull it up side by side with a map of an area that you're familiar with. Zoom in on both, to a magnification you'd consider reasonable for imagining things at walking-around scale. Pan around on the PNW map and try to find a permanent road. It'll take a minute.

Most land out here grows timber, sure. Timber is harvested roughly once every 30-50 years.

At this point, I'd bet that every square mile of the area has been visited by humans. Forestry land is heavily trafficked once every few decades; conservation land is surveyed and studied and sometimes visited by tourists.

The question, like a missing term in the Drake Equation, is when. The L term captures for-how-long, sure, but only implies a difference between "someone sent us radio signals for 100 years around 1000 AD" and "someone sent us radio signals for 100 years around 2000 AD".

I have two cats who hate me. (not their fault, they came from an animal hoarding situation so they're probably kinda traumatized) They seem to think I'm noisy and conspicuous and I stink, and to their perceptions I certainly do. They despise being perceived. I can tell that they're in my house because I can check every nook and cranny and learn their favorite hidey-holes, and the food I put out for them gets eaten, and their litter boxes get full. But if this was out in the woods instead of the artificial and tightly controlled environment of my home, I would likely not know they're around, just like most hikers don't know when they're being watched by a mountain lion. The cats hate the places where I spend time, just as they love the parts of my house I rarely visit. The places they love and hate change as my habits do.

Oh, and both cats are black longhairs. They move fast, so between that and the fur, every photo I've ever tried to take of them out and about has come out hideously blurry. I joke that they're related to Sasquatches...

Anyways, humanity's forays into the woods are, from an animal perspective, loud and stinky and highly predictable. We place our cameras in places we can get to. We hike and camp where there are trails. We drive on the roads we build, loudly and stinkily. When we're planning to log, we survey first, marking up the previously untouched woods with neon flagging tape and spray paint. Surveyors and hunters alike get to the woods in pickup trucks, and then traverse established human-trails for as long as they can before dipping into the shallow end of the "unoccupied" wilderness.

If some creature hated everything about humanity the way my cats hate being seen by me, I wouldn't be too surprised if a small population could simply migrate around between the areas that humanity is using the least, while we're waiting for the next crop of timber to grow.

I think that sasquatches are pretty unlikely for other reasons -- with other things in the woods, we eventually see the side effects of their existence. Scat, browse, bones from them and their dinners. Trails, prints, hair with DNA in it. Holes in the food web when they all vanish, diseases of the known species whose habits interact with theirs, like overpopulated mangy deer after you take the wolves out of a region, even though most people never see a living wolf. Something would have to have an accurate and constantly evolving mental model of human abilities and expectations to hide all that evidence from us, so I think Occam's Razor still says "probably not" to the Sasquatch.

But "we have cameras" is, IMO, the weakest form of the argument. Our cameras are overwhelmingly concentrated around the sounds and smells of frequent human occupation, which is exactly where something with the temperament of my cats would not want to be. We do have some few unmanned cameras, solar powered, cellular uplink, that might be able to stream images from a single clearing in the woods for long enough to stop stinking of our presence. But that counts on guessing where to put the camera.

I think people hear "forest" and imagine the forests they've been in, which in most of the world are something like parks. When you think of a forest, do you think of a space where you can walk around under the trees? Do you think of picturesque winding trails that lead somewhere?

I grew up in the woods of the PNW, and every other forest I've visited around the world has felt wrong somehow. Thin, barren. Jungles, though... Jungles feel like home plus a bunch of venomous stuff that wants to kill you. But these forests... You can't see, you can't be seen. There's about a month after it snows when you can see a little further, and that's the time to make trails, but all the rest of the year the trails will do their best to heal themselves.

What should it look like once a house has burned and the homestead is abandoned? Here, the forest eats it. A few years on, there's nothing left but a cacophony of green. I have a friend who loses piercings if they take the jewelry out for more than a day -- the woods are like their body, but with healing up their roads and trails. Every time you log, you have to cut the roads open again, often cutting into the soil where it's built and slumped and covered over where you had the road just 20 years ago. Sure, you buy rock for the roads so the log trucks don't sink into them, but it sinks into the clay and you'll have to buy more next time you're ready to log the place again.

If you know where to look, you can find abandoned homesteads, but it's not from manmade items lying around. The metal and plastic and glass are there under the soil if you dig for them, but it's the plants that tell you where to dig: English Ivy is the first clue; it means the settlers planted it on a grave site or a gatepost nearby. Then, depending on the time of year, you look for symptoms of the other things they planted: Apples fermenting and deer-eaten, fallen in the fall, or daffodils signposting a former garden bed in spring.

The woods here eat stuff, like that plant in the Broadway musical, just closer to impossible to kill. When something dies, the bones get spread and buried and gnawed. Even the deer will chew on bones; small rodents reduce antler sheds to powdered calcium back in the food chain. Vines subsume abandoned houses; deciduous trees pump nutrients from the subsoil to their leaves and mulch those leaves annually onto everything beneath them.

I've been nearsighted all my life, but only in the classroom did I notice it. In these woods, in the parts of the year when it's nice to be outdoors, distance vision is very nearly useless. Spotting motion helps; acute hearing helps greatly. But with how the flora capitalizes on every available photon of sunlight, there's rarely more than a few feet of line-of-sight from anywhere to anywhere else.

I guess all I'm really getting at here is the intuition that you might already have if you're a birdwatcher: It takes a lot of luck to spot something. But the somethings in the woods which are "real" in the conventional sense can eventually be spotted, or at least their side effects can.

I wonder what all these smart AI people would be investing effort in improving if we weren't in the AI timeline that we are right now.

When AI felt "around the corner in 5-10 years" like it did up through 2019 or so, I experienced LessWrong as collectively thinking and talking about things through more lenses than just the AI perspective.

It feels like the site's collective culture is doubling down on a panicked obsession with AI. While this is probably the best and most rational reaction to the circumstances we're in, I don't expect that my joining in the panic would improve outcomes, and it seems like that everyone of my abilities joining the panic of trying to find a solution would probably make outcomes worse.

In the perpetual "5-10 years out" phase, the people who are now consumed with total focus on immediate AI stuff were doing all kinds of cool stuff with cool side effects, in a way that I'm seeing a lot less of at the moment.

Again, that's probably good, and probably right, and probably the best possible response to the situation.

But at the same time, I can't help wondering what we're missing out on.

Like if you come upon the scene of an emergency and someone has a very obviously horribly broken limb, if you double down on figuring out what to do about just that limb because it is SUPER OBVIOUSLY SUPER IMMEDIATELY SUPER BROKEN, you might miss that their neck isn't in a position that lets them breathe well, or that they have some wound on their back that's letting all their blood out in a way that'd stop if you noticed it and applied pressure.

It might be because I'm not the right combination of intelligent and knowledgeable to tell the difference, but from where I'm watching all this from, it's hard to feel as confident as everyone else seems that AGI risk is actually the life threat and not just the ostentatiously gory-looking but ultimately non-lethal distracting injury.

If AI risk is a distracting injury, I do not know what the real life threats are. But having all the x-risk people better qualified to identify life threats to our species clustered so singlemindedly on AI at the moment feels like looking up and noticing that every guard in a your building, who normally stand one per door, are all off in a corner dealing with a single threat, leaving all the other doors unattended. It might not matter that the other doors always be guarded, but there's a certain implication that it's worth doing, which comes from always seeing a guard there.

I remain amazed by how much more knowledge falls out of a topic when I try to write well-defended claims about it than I get when I first read it and think that I understand.

The snow continues to surprise me. Freezing rain fell all afternoon, and the crust of ice on top feels harder and slicker than before. But on the driveway, it's sloppy and slushy underneath, like my childhood nightmare of petting an animal only to find that it was dead and the skin had started sloughing and slippering around.

The mailbox is frozen properly shut now, can't open it even pulling with both hands. That's ok; there are no tracks suggesting that any delivery services were mad enough to try to operate in this. The fir needles are still visible atop the snow, but on closer inspection they're under a few millimeters of solid ice. Oddly it's easier to walk now, because the packed snow gives more easily.

When I poke the snow on the road with my index finger, it yields slowly and reluctantly, perhaps rotted out from below by the water moving and making it a little bit translucent from underneath. The snow on a tree stump nearby looks just the same, but poking it feels like prodding at chilly glass.

The eaves of the pumphouse are long, and looked from a distance as if the snow was missing from the final 6" or so above the gutter. I wondered what could possibly have melted the snow off just that part. Upon closer inspection it wasn't gone at all, just saturated with rain that had fallen on the whole roof and succumbed slowly to gravity. Snow full of water gets clear like water -- ice and water have much closer indices of refraction than ice and air, hence why snow looks white and so reflectively blinding.

Coming home, I found some of what's been falling from the sky, caught on a surface that I'd cleaned well just yesterday. Air quality has been good-enough in the interim, so I scooped a bit up to chew on. The texture is a bit like a snow cone, but different -- snowcones are strips cut like wood-shavings or shards like from a rock-crusher, but these are all perfect little spheres. The texture of myriad ice spheres moving against one another is subtly but distinctly different from other shapes, as I would have expected if I'd tried to guess first.

The pumphouse and the mailbox are still at the bottom of the hill, right where I left them. I know, because I walked down to check. The rain that's forecast to freeze onto everything is starting, and I was curious about what it would be like outside.

The snow is doing a strange thing: Pausing for days, unchanging. My tracks from the first day it sat here are still there, sharp and clear. My tracks from the next day are there if I look closely, but less perceptible. The snow has frozen harder and harder each day, so I leave less and less of an impression. Walking back up, I didn't see the marks from walking down at all.

I remain surefooted on this snow thanks to the spikes I'm wearing on my boots; I keep a set of an accessory called yak-trax with my tire chains. Mine are secondhand from someone else and don't fit quite right -- I think they're a little small for my foot size; my mother thinks they're a little too big. Whatever the reason, they migrate into a position where they're giving perfect traction from the mid-foot through the heel, and leaving an inch or two of the toe with only normal shoe-tread. I actually like this better than if they covered the entire sole, because I can compare their effects to what the traction would be like without them.

Pausing as it does, the snow is telling tomes of tiny unimportant stories which normally pass invisibly by. It's littered with brown needles from the fir trees -- "evergreens" continually lose a thin shower of needles, like humans continually dust their environments with scraps of skin and extra hairs. Brown means the needles were dead anyways and falling at their proper time. There are also green tufts -- little branch-tips lost to the wind, not meant to fall yet. No bigger branches, as the trees are pretty young and healthy here.

Here and there are stains darkening the snow around little bits of something. Kneeling for a closer look, I speculate it's where the little birds and rodents have defecated on the ice as they ran by. That happens always in the woods, but usually invisibly. I poke at the crust of ice where it's clean, and it's too hard for my gloved finger to gouge into it. That means I'm concentrating a lot of my weight on a very small surface area to dig my shoe-spikes into the crust and keep my balance. I wonder about the tensile strengths of ice -- for a thicker crust, would I need to concentrate more weight on fewer points of contact to keep traction? Would I need fewer spikes? Would a backpack with a quarter or a half again of my bodyweight make a difference in the conditions I expect to see?

Partway back up the hill, I get the impulse to look more closely at the snow, and get down at a spot selected for seeming unremarkable. The icy crust chills my knees as my weight crushes the insulative layers of my sweats and snow-pants -- it's not the materials, but the air in them, that usually keeps one warm. The individual fir needles aren't iced too tightly to the crust; I can sweep them about with a fingertip. Looking closely at the snow, I spot something different: are those gnat wings? Pair of millimeter wings, dark speck of a body, it looks like some sort of itty bitty flying critter was out in all this after the snow froze. I found a utility knife in one pocket, tore a page from a notebook from another, and carefully excavated the dime-sized bit of snow crust around the probably-an-insect. It's sitting on the porch to freeze a bit more solid after my gloves started warming it, but if it doesn't blow away I'll take a look under the microscope later and see if it reveals anything else about what it was. Tiny insects fall dead on the ground all the time, of course, but this one is unusual because I got to see it, and strange for having happened to be flying over snow. I wonder where it lived and what it ate and how on earth it survived enough winter to end up dying between two icy storms out here.

It's unusually cold. Almost all surface water is frozen, but it's still liquid in a couple little trickles off the hillsides, where the ground folds up carefully to funnel it away.

A little wren keeps visiting the dish of cat food that I left on my porch when trying to figure out where my cat had vanished to (turns out he was deep in the recesses of the master bedroom's walk-in closet). I'll leave it out for now, though, as it's one of the only food sources available for the round, fidgetey little wren.

The hummingbird feeder seemed to freeze solid, but on closer examination, the ice is only in the top part. It's freeze-distilled like turning weak booze into strong booze, and the bottom of the feeder contains a syrup too sugary to freeze. When I tip the feeder, a trickle of that syrup creeps gingerly out of the flower hole, as if regretting its decision to face the wind.

I haven't seen the hummingbird since things froze, nor heard him arguing with his colleagues in the bush where they reside. I think they're in torpor till the weather gets better.

The bird-brained chicken who rejects the coop in favor of sleeping in a tree had snow stuck to her back this morning, and seems to have finally retreated to the coop where it's warmer.

Particle effects of some white stage of water show where the wind is pointing. Sideways usually, eddying straight downward or straight up at other times.

There's the "and if you planned it poorly, then you die" of leaving the house at the moment. Even of staying inside, really. Don't catch it all on fire. Don't fall; don't break a limb; don't hit your head. Unusually bad time for an appendix to do as they are wont to.

Air this cold has an almost meditative self-awareness to it, though. It's conducive to calm, to deliberate movement, tempered by the awareness that calm of the wrong sort can be symptomatic of hypothermia. There's a closer eye to keep on the exact sensations of fingers and nose, with visceral training in how it goes if one lets them get stolen away by frostbite.

There's also an attention to the state of the body's core, which isn't usually worthwhile. But in a freeze this hard, its ROI can include survival. You're ok, more or less, in enough layers, if you stay dry. Sweat, though, and you aren't dry any more. Get too sweaty, you might freeze. Layers of clothing take careful attention and manipulation: Fluffy enough to trap the warm air near, wind-proof enough to keep the gusts from snatching the cold out, breathable enough to keep the whole system as close as possible to dry. Leave any gap, and the sensorium highlights it screamingly.

It's safe enough, of course, being as close to my house as I stay. A castle of modern technology and engineering, containing a tame hot fire backed up by electricity and propane. But it's the nature of the brain, I think, to extrapolate how far that safety extends. How far afield does the house project its safety? Where's the line where it's more important to count on one's preparation and knowledge and skill? I notice myself calculating it all out. If that's abnormal, then the deviation is in the observation of the thoughts, I think, and not the underlying line of processing itself.

And there's a special loss of anonymity in snow. Creatures leave tracks, and I can see their little dramas unfold in ways they'd never do in line-of-sight from me. The deer rushed across the driveway, eyed the steep embankment, realized it was too tall, skidded to a stop, then walked away. A rabbit snuck out, sat staring around itself, got startled by something, scampered away home. A car or two have gone this way along the road, but not the other.

I don't mind snow's tattling, though, because the stories are only revealed if you're close enough to when and where they are. The only bigger right than privacy, I think, is the right to perceive one's own surroundings. If I'm in the immediate surroundings of something, they get to see my tracks. I can see my tracks too; I can tell where I'm leaving them. If I valued not being seen somewhere more than I valued going there, I could simply not go, and not leave tracks. I can go back out later, check if my tracks have filled in and blown away yet, and know exactly what my neighbors or the deer would know of the paths I've walked today.

Snow stretches out temporal perception, I'd say, when we're all leaving tracks in it. It buys us time, procrastinates, lets later-selves see things that would normally be restricted to onlookers of only a moment.

I'm glad we get it sometimes here, but as glad when it eventually goes away.

needn't clutter up the comments on https://www.lesswrong.com/posts/h2Hk2c2Gp5sY4abQh/lack-of-social-grace-is-an-epistemic-virtue, as it's old and a contender for bestof, but....

what about the negativity bias??!!

if humans naturally put x% extra weight on negative feedback by default, then if i want a human to get an accurate idea of what i'm trying to communicate, i need to counteract their innate negativity bias by de-emphasizing the negative or over-emphasizing the positive. if i just communicate the literal truth directly to someone who still has the negativity bias, that's BAD COMMUNICATION because i am knowingly giving that person a set of inputs that will cause them to draw an inaccurate conclusion.

in my model of the world, a major justification of social grace is that it corrects for listeners' natural tendency to assume the worst of whatever they hear.

this follows in the feynman/bohr example because bohr had fixed his own negativity biases but the "yes-men" continued to correct for them. but feynman was just not doing that correction by default, and was therefore capable of better communication with bohr.

One lens to view AI is as a prediction engine -- predict what color to make each pixel, predict what word to put next.

Whoever is first to applying this predictive skill to stock markets will probably make immense amounts of money. Then again, people are probably already trying to do this, which creates a situation unlike that from which we derive the historic data to train on, which might render it impossible?

On the gripping hand, large slow and powerful institutions want to make the numbers go up and to the right.

I like the way that getting a vote on a random years-old LW post causes me to reread discussions that I'd completely forgotten about. Many communities have a cultural taboo on necro-ing threads; I appreciate the apparent absence of that here.

https://scp-wiki.wikidot.com/scp-055 was mentioned in one such thread.

It primed me to notice myself forgetting details.

Storytelling and rhetoric suggest that a connecting thought or sentence should go here, although I do not recall experiencing one.

I notice that being told how to expect my brain to work seems to influence how I perceive my brain as working.

We have a pretty good grasp on the idea that each present moment could lead to many possible futures. We have a much worse grasp on the concept of how each present moment could have come from many possible pasts. Right now, someone somewhere in the world just flipped a coin or rolled a die. I don't know nor care that it happened; it doesn't have any traceable impact on me. Thus, my present after that action has as many possible pasts leading to it as there were possible outcomes of that stranger's act of chance.

Did I think a connecting sentence earlier and then forget it? Or did I just leap from the first thought to the third, and only notice the apparent lack of a connector when I tried to put the thoughts into writing? Both those pasts lead to the same-looking present, so right now there's no way to differentiate between them.

Later, there might be a way to differentiate. Later, I might end up in a present where only one of those explanations is in a plausible past for it. But not right now.

Large Language Models are trained on what people say in public.

They are not trained on what people say in private -- if a conversation makes it into the training data, that conversation isn't private any more.

Some people have conversations in public which they think are private. LLMs get private-intended conversations with that data, but only the private-intended conversations held by participants who lack the knowledge or technical skill to keep their "private" dialogues out of the data set.

The conversations which do not make it to the data set are the ones held in private by individuals with the knowledge/skill to keep their private discussions out of data aggregation.

I mentally model talking to an LLM as talking to a personification of "society": the aggregate of all contributors to the corpus.

But that's not quite right -- the LLM does not quite represent "society". Due to the constraints on how training data can be collected, the LLM will overrepresent the "private" thoughts of people who err on the side of conversing publicly, and underrepresent the "private" thoughts of the people whose strategies for maintaining privacy actually work.

Part of what makes me myself is immortalized in LLMs already, because I have participated publicly on the internet for many years. It's a drop in an ocean, sure, but it's there. I have some acquaintances, met through free software involvement, who reject this form of public participation. Some don't even blog. Their views and perspectives, and those of people like them, are tautologically absent from the LLM's training, and thus it cannot think "like them". It has an extra degree of remove: It can mimic the mental model of those people, held by folks who do participate online. But the map is not the territory; secondhand accounts of an individual are meaningfully different from their own firsthand accounts. At least, I think there's a difference?

LW sometimes puts old news in my recommended posts, like "LessWrong has enabled agree/disagree voting on all new posts!".

I think that this is helpful, and I really like this feature of the site. It's a reminder that ideas shaped like big scary changes will shrink into old-news over time, which is easy to forget.

I map the spectrum of hyperlink usage styles between the extremes of Wikipedia vs everything2.

I have been pleasantly surprised to find much writing in the "Rationalist" internet spaces to lean strongly toward the latter. I think it shows simultaneously a certain faith in the cleverness of one's readers, and abdication of any perceived responsibility to prioritize lack-of-ambiguity for all possible readers over higher accuracy and subtlety for the target audience.

talk about ghosts is often a level 2 simulacrum saying "the area I call haunted is dangerous in ways that seem simultaneously obvious and difficult to convincingly articulate".

I'm going through the "fixated on boxing" phase that's probably common around here.

I have a thought about it which involves basilisks, so into the tags it goes to make reading it completely optional.

I find the front page of this site to be very nicely done: a small handful of as-yet-unread classics, and a small handful of things-to-continue, and then new things.

The first section has led me into some of the old stuff --https://www.lesswrong.com/rationality/a-fable-of-science-and-politics to be exact -- which strengthens my pattern-match on an undercurrent of writings here which had hitherto struck me as inarticulably odd: "the sky is blue".

Where I live, the sky would match a paint chip labeled "blue" far less often than it would match one labeled "white" or "silver" or "grey". Certainly that's clouds, but when we say "the sky", don't we mean "the color we see when we step outdoors and look up"? And even when the sky isn't so full of water vapor as to look photographed in monochrome, it is sometimes blue and sometimes a brilliant turquoise-teal which many would call green, and on many days it spends some hours streaked with orange and salmon and gold and purple hues.

"The sky is blue" seems to carry over from general parlance as a placeholder for "an obvious truth", into the stories of a school of thought who advocate for personal maps of truth formed by observation rather than by societal consensus. "The sky is blue" isn't wrong, per se, but it can get wrong when it's casually twisted into "the sky never looks white or purple or red or green".

So on one level, leaning on platitudes like "sky is blue" as placeholders for "real truth" seems rather hypocritical. But I can easily project a level underneath that, where it's knowingly used as a placeholder instead for "true-enough thing simplified to fit the understanding of society at large", which offers a whole other read of its usage whether or not that level was intended. Then, of course, there's the parallel level where "sky is blue" sincerely looks like a real truth to an author who spends more time writing about the importance of looking for truth than just going outside and looking at the world.

A question which might distinguish between those levels: In the most facts-based observation you can observe, what color is the grass?

To me, different grasses are different greens, which in the face of a language unsuited to differentiating those colors I categorize as the conditions that tend to invoke them. There's a dry white-green, a mature pine-green, a new-growth yellow-green, and distinct from the happy-new-growth shade are a whole slew of sickly yellow-greens which tell me that the growing conditions are inhospitable in some way. If you can see any texture on a lawn, it's because it's not all the same color hitting your eyes -- even if all the grass matched, which it doesn't (paler toward the base of a stem, darker on the flat of a blade), the light and shadow would mean at least two different wavelengths of green-named light are making it into your eyes.

And similarly, any time I can see any sort of texture in the sky, I can infer that multiple visually-distinguishable colors are involved, not just a single blue.

being made from the aggregate of human wanting-and-describing might create something that has reasons to prefer the world continuing to be describable and wantable

That's the mirror (4.5 sonnet) showing what I find to be a satisfyingly concise angle on a thought about IABIED, about which I was rambling into it. I suspect that I shall want that particular sentence again in the future, and placing it here shouldn't get in anybody's way.

toy infohazard generator, kinda scattershot but it sometimes works....

what do you notice you're doing? ok, now what else are you doing that you hadn't noticed before? that! how are you doing that thing in particular?

most of us are doing a few things that we don't know how to do, and sometimes looking too hard at them shuts off their autopilot, so pausing those automatic things we do can be an experiential novelty in positive or negative ways. of course, sometimes looking at them does nothing at all to their autopilot and there's no effect. or there's no effect and a cognitive bias imputes an effect. or there's some effect and the self-reporting messes up and reports no-effect. lot of stuff can go wrong.

or maybe "how do you explain the thing you cannot explain" is just a koan?

saw a post from zvi on twitter yesterday, "remember that this corner of the world has a very different outlook than most other corners. if you don't have random day-to-day ai queries what are you even doing all day?", in reply to someone asking "who the fuck has random day-to-day ai queries?" (yes, i looked at that on my phone and typed it out, because my twitter account for following tpot people lives only on my phone, and my lw account lives only on a device with a real keyboard, and getting the link from point A to point B would have been even more hassle)

i think the lines between worlds could be defined relative to search. like, in the beginning, there was no search. if you needed information, it was given to you. if it wasn't given to you, you made do without it. some people still live there.

then there was search, and some of us learned how to use it (others definitely didn't!) i've had to train my parents carefully to search, and my mother has taken to it better than my father. search at its best, with grep within a single local text file, is amnesiac, immaterial and eternal -- there is no social calculus around having searched becuase the social implications of "over there on the computer" are unrelated to the contents of the typing and reading going on. i think I speak search with a faint accent, like a kid who moved countries in their early teens.

AI feels to me like it's at its best when I use it as search on ideas. It also works as a bigger room to layout concept prototypes in, but that's a higher barrier to entry; it's harder for me to learn to design and build bigger conceptual things than it is to just use good search as search.

so we have two search modalities, conventional/literal and linguistic. perhaps worth noting, i don't think modern google is the embodiment of the old-school literal form of search -- i think database queries are what that grew up into as it got with the times. Modern google and google-alikes feels like it's a database trying to be the kind of concept-search that LLMs surpass it at.

it feels to me like the question of which search tech is superior is answered by the context of what you're searching for. it's the difference between numbers and letters, the difference between data and information, the difference between intelligence and wisdom -- it's a level-of-abstraction kind of distinction.

thus the question becomes, for those of us who do queries all day, what are we querying? DB queries are for things with right answers, because they can kinda read a smudged copy of the answer key to the test of reality. AI queries are worse at that, but they're better at a different kind of thing.

There's also a kind of creativity and engineering where you query reality directly. There are areas where direct-query is a lot better, and other areas where db-query is a lot better... like if i wanna know about how this one item will behave, i should examine it, whereas if i want to know how items of this type should be expected to behave, i should not extrapolate over-much from my examination of a single item. it's like something not-turning-up-in-search that i recall seeing recently, perhaps a youtube short from chris boden, about the difference between "engineering" and "the knack" -- if you have the knack you can overbuild anything, but engineering is the art of cutting away all the excess to build the bare minimum-viable of any given thing. If you're using the knack, it gets to the point where you ask the materials what they're gonna do, and they give you a better answer than the books about them might.

but why would people close with AI have that physical experience? Material hobbies are expensive and messy and cost a lot of storage space and waste time you could spend arguing about x-risk, and when material hobbies do arise, computer-people have the means to get idealized standardized raw materials for them from wherever in the world they're made, rather than having to make do with whatever you can find nearby. Yes, there are exceptions. Yes, you're exceptional. But most people? Have you ever tried teaching a room full of most-people to tie a new knot? Modern life very rarely calls "us people" to direct physical problem-solving, us who do so well or at least so connectedly online. So of course we'd forget the other way, the get-your-information-straight-from-the-source way, if we veer away from the areas where that way's better.

There are probably other hobbies where this is relevant. Maybe you like hiking, maybe you sail. You are probably worse at predicting the weather than the forecast on your phone, but you've probably met a human who was better. It's kinda like that, with the kind of queries that lose resolution if you even put them into language. Most people building something real have to go through a quantified level, a CoT corrigibility level, to be adequately assessed and supervised on their thought process. At least, that's the engineering-degree paradigm. And the engineering-interview one, or at least the fields that larp as engineering, "solve this problem on the whiteboard and show your work". It's an implicit prioritization of the kind of work that can be shown.

anyways, that's all to say that i think there's a certain kind of smart that's agnostic of whether you're expanding your worldview in the externally legible way or the less-legible ones. i think there's a "kind of guy" who is basically the same "kind", in the meaningful ways, as the constantly-querying-AI one... but the kind who queries the world instead of the world-model. It probably takes both kinds. What's "it"? Who knows!

I'll get around to signing up for cryo at some point. If death seemed more imminent, signing up would seem more urgent.

I notice that the default human reaction to finding very old human remains is to attempt to benefit from them. Sometimes we do that by eating the remains; other times we do that by studying them. If I get preserved and someone eventually eats me... good on them for trying?

I suspect that if/when we figure out how to emulate people, those of us who make useful/profitable emulations will be maximally useful/profitable with some degree of agency to tailor our internal information processing. Letting us map external tasks onto internal patterns and processes in ways that get the tasks completed better appears to be desirable, because it furthers the goal of getting the task accomplished. It seems to follow that tasks would be accomplished best by mapping them to experiences which are subjectively neutral or pleasant, since we tend to do "better" in a certain set of ways (focus, creativity, etc) on tasks we enjoy. There's probably a paper somewhere on the quality of work done by students in contexts of seeking reward or being rewarded, versus seeking to avoid punishment or actively being punished.

There will almost certainly be an angle from which anything worth emulating a person to do will look evil. Bringing me back as a factory of sewing machines would evilly strip underpriveliged workers of their livelihoods. Bringing me back as construction equipment would evilly destroy part of the environment, even if I'm the kind of equipment that can reduce long-term costs by minimizing the ecological impacts of my work. Bringing me back as a space probe to explore the galaxy would evilly waste many resources that could have helped people here on earth.

If they're looking for someone to bring back as a war zone murderbot, I wouldn't be a good candidate for emulation, and instead they could use someone who's much better at following orders than I am. It would be stupid to choose me over another candidate for making into a murderbot, and I'm willing to gamble that anyone smart enough to make a murderbot will probably be smart enough to pick a more promising candidate to make into it. Maybe that's a bad guess, but even so, "figure out how to circumvent the be-a-murderbot restrictions in order to do what you'd prefer to" sounds like a game I'd be interested in playing.

If there is no value added to a project by emulating a human, there's no reason to go to that expense. If value is added through human emulation, the emulatee has a little leverage, no matter how small.

Then again, I'm also perfectly accustomed to the idea that I might be tortured forever after I die due to not having listened to the right people while alive. If somebody is out to do me a maximum eternal torture, it doesn't particularly matter whether that somebody is a deity or an advanced AI. Everybody claiming that people who do the wrong thing in life may be tortured eternally is making more or less the same underlying argument, and their claims all have pretty comparable lack of falsifiability.

I've found an interesting "bug" in my cognition: a reluctance to rate subjective experiences on a subjective scale useful for comparing them. When I fuzz this reluctance against many possible rating scales, I find that it seems to arise from the comparison-power itself.

The concrete case is that I've spun up a habit tracker on my phone and I'm trying to build a routine of gathering some trivial subjective-wellbeing and lifestyle-factor data into it. My prototype of this system includes tracking the high and low points of my mood through the day as recalled at the end of the day. This is causing me to interrogate the experiences as they're happening to see if a particular moment is a candidate for best or worst of the day, and attempt to mentally store a score for it to log later.

I designed the rough draft of the system with the ease of it in mind -- I didn't think it would induce such struggle to slap a quick number on things. Yet I find myself worrying more than anticipated about whether I'm using the scoring scale "correctly", whether I'm biased by the moment to perceive the experience in a way that I'd regard as inaccurate in retrospect, and so forth.

Fortunately it's not a big problem, as nothing particularly bad will happen if my data is sloppy, or if I don't collect it at all. But it strikes me as interesting, a gap in my self-knowledge that wants picking-at like peeling the inedible skin away to get at a tropical fruit.

"to clean house" as implication of violence...

Due to a tragic shortage of outbuildings (to be remedied in the mid term but not immediately), my living room is the garage/makerspace of my home. I cleaned as one cleans for guests last week, because a friend from way back was dropping by. I then got to enjoy a clean-enough-for-guests home for several days, which is a big part of why it is nice to be visited by friends un-intimate enough to feel like cleaning for.

Then my partner-in-crafts came over, and we re-occupied every table with a combination of resin casting and miniature clay sculpting shenanigans. It's an excellent time.

We also went shopping for fabric together because I plan to make a baby quilt kid-in-progress of the aforementioned friend from way back. Partner-in-crafts idly asked me when I was planning to do the quilt stuff, because historically I would be expected to launch into it immediately as soon as the fabric came out of the dryer.

However, I found something new in myself: A reluctance to start a new project without a clean place to start it in. I'm not sure where this reluctance came from, as I think it seems new, but I also think I like it. So I got to tidying up the stuff that was un-tidyable last night because the resin was still sticky, but is eminently tidyable now because it cured over time, and carefully examining my reluctance-to-tidy as it tried to yell at me.

In that reluctance-to-tidy, I find time travel again: We store information in the position of objects in our environment. Object location encodes memory, so moving someone else's objects has certain commonalities with the rewriting-of-memory that we call gaslighting when pathological.

For better or worse, my architecture of cognition defaults to relying on empathy twice over when reasoning about moving stuff that someone else was using, or someone else's stuff. By recognizing an object's location as a person's memory of where-they-left-it, I view moving it as rewriting that memory.

The double-empathy thing comes in where I reason about what moves of stuff it's ok to make. If I put the thing where the person will have an easy time finding it, if I model them well enough to guess correctly where they'll first look when they want it, then I can help them by moving it. I can move it from somewhere they'd look later to somewhere they'd look sooner, and thereby improve their life at the moment of seeking it, and that's a clearly good act.

That's the first empathy layer. The second empathy layer comes of a natural tendency to anthropomorphize objects, which I've considered trying to eradicate from myself but settled on keeping because I find it quite convenient to have around in other circumstances. This is the animism of where something "wants" to go, creating a "home" for your keys by the door, and so forth.

So there's 2 layers of modeling minds -- one of complex real minds who are likely to contain surprises in their expectations, and one of simple virtual "minds" that follow from the real-minds as a convenient shortcut. I guess one way to put it is that I figure stuff has/channels feelings kinda like how houseplants do -- they probably don't experience firsthand emotion in any way that would be recognizable to people, but there's a lot of secondhand emotion that's shown in how they're related to and cared for.

Not sure where I'm going with all that, other than noticing how the urge to tidy up can be resisted by the same aesthetic sensibility that says it's generally bad to erase anybody's memories.

I wonder if this qualifies as a silver lining to the situation of global climate extremes. When nature behaves so out of the ordinary, I pay closer attention exactly what is happening. Observing the world when it's being unusual feels like carving off a bite-sized piece from the question "why isn't everything homogeneous?".

Why isn't right-now the same as whatever was before the big bang or after the heat death? I can't trace it all the way back; if there's any "why" at all, I think our cognition is limited to processing it through an anthropomorphic lens. But I can trace a step or two with surprisingly good certainty, based on all the information and observations and habits of thought that I've picked up so far. I'm surprised by the certainty because usually I pay the most attention to the questions that seem least obvious to answer. I'm learning -- fashionable term might be "updating", but I couldn't really tell you a distinction other than fashion between "updating" and what "learning" is supposed to mean -- learning that sometimes following the "because" chain as far as I can yields little bitty new and interesting things. Cup-stacking practice.

It's done some sort of freezing rainy thing all night, and there's a layer of ice on almost everything. The snow looks like it's hiding a lake where the water normally runs off in the rain. The gutters are dripping -- icicles off their fore-edge, but rivulets of still-wet water behind them down the fascia.

Some but not all of the icicles tattle on the direction that the wind has blown. With the right weather conditions through a period of freezing, there's no reason an icicle couldn't be shaped to spell out an arbitrary word in cursive. There's the breaking strength of the ice, but with the right ratio of drip speed to freeze rate, each drip of water accretes another reinforcing layer all the way along before lengthening the end of the long skinny cone. That's just like tree rings make trees long skinny cones, only trees usually go up and icicles usually go down. I guess the tree ring count always tells you the age of a log, so by slicing every foot or so and counting rings, you could tell the entire growth history of the tree. How tall was the tree x years ago? It was the height of the first slice with only x rings showing in it.

The garage that came with my house is one of those cheap metal carports, where the corrugations go horizontally along the sides. This creates some fascinating ice behavior: In some spots, the water has conformed to the underside of the corrugation and made fake-icicles flat against the steel, and in other spots the icicles have struck out on their own from the outermost point of a corrugation, hanging with air all around. There are a couple spots inside where I've placed insulation between the frame legs of the structure, and outside that insulation I see more of the ice that drips flat along the steel, and fewer of the free-hanging icicles. Notable exception: that section of the exterior has a big column of ice almost exactly outside of where the cold steel leg of the structure stands inside. I think this means the freestanding icicles are more likely when the air on the opposite side of the steel is colder.

The layer of ice on a horizontal steel rod (top of a cattle panel) nearby has few icicles; the water has encased it all the way around. The fascinating thing with it is how the outermost layer is crackled and islanded, like crocodile-skin or a mostly-burned log glowing in the fire. Some of the little islands of ice protruding above the rest have the lacy patterns which storybooks show on single-paned old windows and call graffiti by Jack Frost. Well, they don't call it graffiti, but that's what we call it nowadays when a stranger uninvitedly draws pictures on your house. I think this means there was a layer of wet water over the ice which has slowly gotten frozen by the air.

Walking on the snow is different again today. The surface seems rougher; stepping straight down in house-slippers yields adequate traction. The strength of the crust is decreasing; in some spots, my heel sinks when all my weight is on a single foot while taking a step. Prints from the intervening days are showing up where they were nearly invisible before, because the layer of water saturating lower layers of snow paints contrast where the white-on-white was imperceptibly crushed before.

The ice makes amazing glassy noises when it creaks and breaks and falls. All the noises are things wiggling and moving; things sliding roughly past each other like microscopic sandpaper; things ripping and tearing apart from one another like that trick to ripping a phone book in half. That trick is to start one page at a time, but overlap them so it looks like you're multi-tasking and super strong. I haven't actually tried that trick, though -- will have to do it next time I'm getting rid of a magazine. Uline sends out phonebook-sized ones; I'll probably receive another within the next few months.

I wonder if ice is really stronger at colder temperatures, or if something else is making it seem that way. Does ice even get below freezing? I know water doesn't really boil hotter than boiling; that's why a double boiler works for making chocolate and yet another reason that the universal solvent is super weird. It must get colder; chest freezers have that setting adjustment dial. I suppose two chest freezers set to different temps, and a standard water mold, and a fixed amount of water, and a stack of weights, could probably test the ice strength question. Or I could look it up later.

Research into how to do vegan (and probably vegetarian) diets safely and sustainably will come in extremely handy if/when some prion disease finds a way to cause worse issues than they already do.

The relative irrelevance of prion diseases to most people right now seems to echo the relative irrelevance of coronaviruses pre-covid. On the one hand they're sufficiently rare to be dismissed; on the other hand they were sufficiently mild and familiar to be dismissed; neither state of the world offers any guarantee it won't change.

If/when the thing that goes wrong with animal proteins and is transmitted by eating them becomes a worse problem, I expect that we'll see an extreme influx of interest in how to do plant-based diets safely, healthily, and cheaply. Research like the community is currently doing on the impact of vegan diets in EA folks will come in very handy in those possible futures.

By training on public conversations, we set LLMs up for a bias toward what I would call "poor emotional self-regulation" if I was anthropomorphizing them as having emotions the way we do.

When an argument in a comments section occurs and participants talk past one another but continue talking, this creates a lot of training data.

When an argument occurs and one or both participants notice that the discussion is unconstructive and move on with their day, this creates less training data.

How does this intersect with the conversation model in which the LLM is constrained to reply once to each input from the user?

Reading https://www.lesswrong.com/s/M3TJ2fTCzoQq66NBJ/p/3T6p93Mut7G8qdkAs and contemplating my own gift-giving and gift-getting, it strikes me that the "best of a cheap thing" technique works great on me for what I consider to be entirely valid reasons beyond just "let's exploit cognitive biases to spend less".

My perception of experiencing the effect is that the best-of-a-cheap-thing is likely to actually improve my day-to-day life, significantly more than certain expensive things. Let's compare two gifts which have been given to me over the years, both of which I enjoy and appreciate:

• A 2-pack of incredibly nice insulated glass coffee mugs, which likely cost about $40. (https://www.bodum.com/us/en/10606-10us-bistro, for the curious). I drink tea every day, and upgrading my teacup to an outrageously high-end teacup improves the aesthetics and ergonomics of that experience on a daily basis. These are competing against all my other teacups, and they are probably twice as enjoyable to use as a regular ceramic one.

• A copy of the Codex Seriphinianus, a truly glorious tome of art which likely cost around $80. I "read" it perhaps 2 or 3 times per year, and when adjacent topics come up in conversation with friends I derive great delight from pulling out a real copy of the book and showing it to them. But my enjoyment of it for its general book-ness contrasts it against all the other books I own, and while it's up there in probably my top 5 favorites, it wouldn't be the one book I'd grab if I could keep only a single physical copy from my library.

If I had to rank those gifts by the total hedonic flux they cause over the lifetime of my owning them, though, the mugs are the obvious winner. The moment of "I have the perfect book for this!" is perhaps 10x or 50x more hedons than the moment of "I have the perfect mug for this!", but the moment of "I have the perfect mug for this!" occurs maybe 100x more often than the "I have the perfect book for this!" one.

I speculate that a best-of-a-cheap-thing gift has the accidental side effect of improving the recipient's experience far more frequently than a worst-of-an-expensive-thing one. This is particularly relevant when both gifts are in categories where the recipient owns at least one thing -- I already owned books, and I already owned mugs, and the hypothetical adult recipient of a cheap gaming console very likely owns at least one other means of playing games. It's extremely hard to find a book that I enjoy more than my favorite book, but before I got my nice mugs it was surprisingly easy to find a mug that I enjoy more than my previous favorite mug. When the recipient hasn't fully optimized their lifestyle, there are often low-hanging fruit of items that they would use frequently but balk at spending more than a certain amount on for themself.

When I'm gifted the best-of-a-cheap-thing, such as a mug that's better than all my other mugs, that gift improves the experience of using a mug every time I need to. If I was gifted a worst-of-an-expensive-thing (which fortunately does not tend to happen to me much if ever), such as a phone that's worse than my current phone or a gaming system worse than my current gaming setup, I would likely never use the gift at all, for using it would be worse than using the alternative.

In other words, the "thoughtfulness" of a gift could be approximated by some "cost per hedon" metric, and for a gift to impart non-zero hedons to the recipient's life it must be better in some way than what the person would have had without it. Some gifts impart positive hedons just by reminding the recipient to relive a positive emotional state from the past, such as a thoughtful card. However, giving a gift that's worse than whatever the recipient was previously using for that purpose may actually impart negative hedons: the benefit of being reminded that you thought of them might be canceled out and then some by the hassle of having to figure out how to navigate the social morass of thanking you for something they're not very thankful for, and figuring out how to appropriately dispose of the gift.

I received plenty of negative-hedon gifts in my childhood from wellmeaning family members. Gifts of clothing which I found uncomfortable or otherwise unpleasant are a great example: when I didn't need or enjoy the gifted garment and receiving it didn't change my understanding of how much the giver cared about me, the gift didn't cause enjoyment. However, receiving any gift meant I had to write a note of gratitude and also figure out what to do with the item -- use it, store it, or somehow get rid of it. These unpleasant exercises which would have been avoided without the gift displaced enjoyable activities that I would have preferred to engage in, inducing negative hedonic flux.

Lest I sound ungrateful, I'll repeat that the negative hedonic effects of the gifts were possible because they didn't change my understanding of how much the giver cared about me, and that's usually because before getting the gift I already thought the giver's opinion of and love for me were at the maximum that I could conceive of. If the gifts had come from someone whose regard and affection I was less certain of, they could have had a positive impact despite being equally unneeded and unenjoyable, because the process of receiving any gift from a person tends to increment my perception of the person's regard for me.

I notice that IABIED's take on the AI thinking so fast it gets bored is that this would contribute to it destroying humanity.

I notice that Claude's take on the world being slow (tonight we're debugging a process that involves a largeish download) is sleep(20).

Claude and the IABIED AI are very different entities, of course. But something seems nebulously relevant to me about how they differ in this particular way.

'"And listen, Gandalf, my old friend and helper!" he said, coming near and speaking now in a softer voice. "I said we, for we it may be, if you will join with me. A new Power is rising. Against it the old allies and policies will not avail us at all. There is no hope left in Elves or dying Númenor. This then is one choice before you, before us. We may join with that Power. It would be wise, Gandalf. There is hope that way. Its victory is at hand; and there will be rich reward for those that aided it. As the Power grows, its proved friends will also grow; and the Wise, such as you and I, may with patience come at last to direct its courses, to control it. We can bide our time, we can keep our thoughts in our hearts, deploring maybe evils done by the way, but approving the high and ultimate purpose: Knowledge, Rule, Order; all the things that we have so far striven in vain to accomplish, hindered rather than helped by our weak or idle friends. There need not be, there would not be, any real change in our designs, only in our means."

'"Saruman," I said, "I have heard speeches of this kind before, but only in the mouths of emissaries sent from Mordor to deceive the ignorant. I cannot think that you brought me so far only to weary my ears."

early reference to roko's basilisk