comfy threadpost cute/comfy images and chill
sorry let the last one die because i had to go out and get a new hob because my old one broke. got an induction one from ikea :D
>>24016837cute image - where is that art from?
>>24016902https://safebooru.org/index.php?page=post&s=view&id=6468558
I love non non biyori and its threads >>>/a/285960075
>>24016913is that u brazilbro?
>>24016916Yes it is :)Still feeling kinda tired though
>>24017187that's understandable when studying such intensive stuff. I hope you got some better rest over the weekend even if it wasnt quite enough to have you back at 100%
>>24017187i'm glad ur still here>>24016840>i had to go out and get a new hob because my old one broke. got an induction one from ikea :Dooh
>>24017773induction hobs are cool as fuck, I've not cooked on it yet but it boiled some water in like 20 seconds which is fucking crazy
>>24017773k0taaaaaaaaaaaaaaaaaa
>>24017825koX ID
>>24017820i've heard they're really good for that, that is cool>>24017208>sample_574da0cc17430e85c1(...).jpg (222 KB, 850x560)rip>>24017825aw FUCK i didn't noticei'm not used to checking for IDs in these threads specificallythere was not another murder of crows, sadly, at least not while i was awake today
>>24017829you might get more friends innafewID gods bless
>>24017833it's just living specifically here i read the ID before i read the poster's text.
>>24017833I didnt even notice when brock said, I thought he was just rolling.shame you didn't see the crows again, they probably left the area if youve not heard them because they are loud as fuck when in a big murder
>>24017839i always respect and check the blessings of the ID gods.your thread is blessed by a comfy anom with a really good ID.
>>24017903<3
>>24017904>PG8okay i'll stop the swearing and slurs just here.
>>24017909hehe
This image looks very comfy
>>24017998:D
>>24017998yes very :Dgotta do a bunch of shit tomorrow gonna try sleep.night everyone <3
>>24018034gngn <3sleep well
>>24017773THE KOT!!! °O°
>>24018379
>>24017773KOT!!!!!!
>>24017903
>>24018567>>24018576
>>24016837why have a door mat when half your house is dirt floor?
>>24018967so you don't get outside dirt on ur indoors dirt obviously
>>24017773someone get me the kot alarm.on top of that it's the allusive k0t. the most sacred if all the variations.
>>24018967looks like wood floors?>>24019577obviously im wrong
oh lawd not da mufuggin fake ass backwoodsdis blunt gon gimme feline aids fr
>>24019989he means the half with grass growing on it I think
>>24020776man the stress of being around haruhi must rlly be getting to her xD
head hurtstoo much computer todaygood night
g'day>>24020941sleep well <3>>24018967my mom always bought too many floormats even when they'd end up making things dirtier loli guess some people use them decorativelyu right, though>>24019577hehe
>>24020941hope u feel better tomorrow..night :)>>24021073g'day :Dhow's it goin?
>>24021082pretty goodit's very windy outside, i saw a guy trying to rake the lawn of one of the vacant houses but he gave upcan't think of much else
>>24021148is it windy enough that and u can hear the wind howling past ur window while ur all warm and comfy inside? :3this week is gonna be so expensive, had to buy a new cooker, new glasses and now I find out I need to buy new pots :'(
>>24021223yees it's fairly loudthe wind here almost never does that whistling thing, sadly, but it's still pretty cooli crack the window sometimes, the fresh air is niceit's calmed down a little, now i hear different birbs>this week is gonna be so expensive, had to buy a new cooker, new glasses and now I find out I need to buy new pots :'(aw manwhat happened to the pots?
>>24021258there needs to be some hills or tall buildings for it to whistle, it sometimes happens with trees but if ur surrounded by forest it may dampen it.>I hear different birbsmy mum got some app that can tell u what birds ur hearing, it sounds interesting but i don't want some fucking AI bloatware on my phone. its called merlin I think, maybe they have some online tool or we can find something similar.>what happened to the potsthey don't work on the induction stove xD i have no idea why they are all stainless steel and magnets stick to them.my only guess is that where the base is welded on there some copper or something underneath so they heat up quicker on a gas stove.
>>24021543>there needs to be some hills or tall buildings for it to whistle, it sometimes happens with trees but if ur surrounded by forest it may dampen it.ah tythere are some big trees but not too many, very flat terrain & the tallest buildings are maybe 5 or 6 storiesbut this is good to know>>24021543>my mum got some app that can tell u what birds ur hearing, it sounds interesting but i don't want some fucking AI bloatware on my phone. its called merlin I think, maybe they have some online tool or we can find something similar.i think my dad has used similar apps for birds, there are ones for identifying plants tooi remember i found some google platform in 2019 with open challenges in machine learning, one of them was exactly that problem of identifying one of a few bird calls in a noisy audio streami really wish i could remember the name of that platform, i wanna go back & investigate it further, even if it's not running anymore...>my only guess is that where the base is welded on there some copper or something underneath so they heat up quicker on a gas stove.ah, good guess maybe
>open challengesnot like unsolved problems or anythingit was more like a platform aimed at teams of students etc. to compete, but i think anyone could submit an entry to one of the challenges?iirc
>>24022700I think it only happens when the wind blows between 2 tall things, maybe the buildings are too far apart or something. I don't rlly know, fluid dynamics is confusing as fuck so I try not to think about it too much xD>I think my dad has used similar apps for birdsyeah they seem pretty handy desu. I might look into them and see if there's one that's not got privacy invading features or raped with ads.>open challenges in machine learning...sounds interesting, id imagine its probably not around anymore with how ubiquitous consumer AI has become, ive been trying to learn about boolean machine learning and decision trees but idk if I'm interested enough to actually learn much.I think that whole area of research died down quite a long time ago when fuzzy logic became more common.>>24022720that's pretty cool but also very fucking jewy that google was expecting ppl to just do that shit for free xD
safety bumop
I always laugh when i remember that good smile company invested 2.4 million dollars in this mongolian basket weaving forum.
>>24023787<3>>24024326loool, wtf happened to the money?
>>24024579they got a 30% share from here for some years, pretty sure they abandoned it after seeing the profit was 500 yen and a pack of cup noodles.it's actually pretty interesting that hiro probably thought he would be as influent as he was in the 2ch days with this place>>24025139i love her so much
>>24025139thats very kawaii desu ngl omg mhmm
>>24025193lool that's fucking funny. its pretty crazy that hiro is so well respected in japan compared to moots reputation in the states when he was still running this place. I wonder wtf he's up to these days and if he's ever gonna comment about the epstein shit.>>24025529nano is the cutest robot
>>24022738>that's pretty cool but also very fucking jewy that google was expecting ppl to just do that shit for free xDvery true>>24022738>ive been trying to learn about boolean machine learning and decision trees but idk if I'm interested enough to actually learn much.:oidk too much eitheri studied neural networks around 2020 & implemented several examples, i was interested natural language processing as welli was interested in these things for their own sake, but also in relation to algorithmic trading, never got too far into that thoughit seems like the field of "AI" was classically much more related to algorithm design for a long time, during the 20th century mainlythat stuff seems more based, but i know almost nothing about itmachine learning almost seems like an abstraction layer which can be used as a crutch, where people would previously have had to design algorithms, which might've been harder or more esoteric, but also more of a crafti wish i knew...>I think that whole area of research died down quite a long time ago when fuzzy logic became more common.ah ye, i had that impression toobut i think maybe it was always kind of a small field, and fuzzy logic/machine learning is almost a separate thing which became much biggerthat's my impression, but again idk much theory or history>yeah they seem pretty handy desu. I might look into them and see if there's one that's not got privacy invading features or raped with ads.hehewith a lot of those "apps", and applications of machine learning in general, the more interesting problem usually becomes how to source the right kind of datathe methods are more or less the same between a lot of them, speaking broadly
>>24025580>its pretty crazy that hiro is so well respected in japanit's also hilarious that it's mostly young people and some who know him from the old 2ch days that actually like him as a personfrom what I'm told, he's liked by young men since he says what they think. He also gives some advice on stuff that ranges from finding a job to maintaining its moneybut other groups such as nettouyo, traditional nationalists, conservatives, leftists, and even some weirdly specific groups such as okinawan activists who fight against US bases despise him lmaoi've seen some people pretty much describing him as a cunt who only wants social chaos>I wonder wtf he's up to these dayswouldn't doubt if he's in japan these days, now idk what would he be working with.i've seen people talking on how he has a wife and even a daughter there, i've also seen some pics being thrown around as evidence.>ever gonna comment about the epstein shit.probably never, maybe my wording is too strong but i really think he despises this place and wishes he never created it, he probably feels like that since the gamergate daysi know he recently privated his github and reddit accounts, two otakon videos were also set privated on 4chan's youtube account but they were so specific and harmless i don't doubt they were just shadowed by youtube and its weird copyright shit
>>24025580>nano is the cutest robotNext to mahoro btw :)Well, i guess i'm going to try to sleep since my head hurts a bit, good night.
>>24025737>very truexD>it seems like the field of "AI" was classically much more related to algorthm design for a long time.yeah that's pretty much the case, decision trees are "non-parametric supervised learning algorithms" (whatever the fuck that's supposed to mean lol) and AI research was basically was all just ppl trying to create algorithms that essentially mimicked learning. as far as i understand it got kinda abandoned becaue of fuzzy logic but i think now they are starting to be used a bit again in generative AI but i haven't learned enough yet to tell u what their application is or anything.>which mightve been harder or more esoteric, but also more of a craftyeah that's sorta the impression that I got which is why I became interested in this shit, I feel like there's a lot more charm to doing it the hard way and maybe im just stupid enough to think of some shit that was overlooked by people far smarter than myself that maybe could be the key to ridding ourselves of these stupid fucking "neural network" clankers that i hate so much.>fuzzy logic/machine learning.is almost a seperate thingive gotta admit my reaserch is pretty lacking in the specifics but i think ur on the right track and fuzzy logic systems were never intended to make shit like neural networks and i think they just wanted machines that werent limited to boolean logic.>the more interesting problem usually becomes how to source the datado u mean the like going out and recording millions of birds part or the making sure the data set is accurate? sorry my brain is on limp mode today I feel like I'm just pretending to understand what we are talking about in a way xD>>24025754yeah i can imagine a lot of groups fucking despise him lol, I was thinking about like how he hosted some political talks and shit so he's seen as like legitimate whereas if like moot had hosted some political debate in the states it would have caused a fucking riot or something xD(1/2)
(2/2)>>24025754>cunt who only wants social chaos/ourguy/>I've also seen some pics being thrown around as evidenceI bet he's thrilled to have photos of his daughter appearing on this website lol.thats pretty cool if he's moved to japan, id imagine it was prolly a big dream of his.>he recently provated his GitHub and reddit accountsI mean that's pretty understandable, he was probably getting harassed by /pol/ shcizos or something. I dont think he hates this site but id imagine he hates the cancer that its become full of.>>24026026I've not watched that show, ill have to check it out.sleep well brazilbro <3
>>24026038>do u mean the like going out and recording millions of birds part or the making sure the data set is accurate?both, but there are often many other steps before data is usable, and decisions made while preparing the data can have a huge effect on how effective the finished system isthere's often a lot of room for hacking & ingenuity in collecting/preparing data, depending on the case & what the system is supposed to do>going out and recording millions of birdsthe architecture of the neural network(s) can affect how big the training data needs to bemillions of samples are often required, but there are cases where you can get away with the order of thousandsi think it's typically very hard to get anything usable/effective with less than 5,000-10,0000 samples, outside of very special cases, but i could be wrongat the same time, if you use too BIG a training dataset, you can get what's called "overfitting" which results in the system performing very well against random samples from the training data, but too inaccurate against new samplesthere are probably different schools of thought when deciding the cost of collecting/preparing data vs the cost of a less accurate/perfect model, and what the sweet spot is for different cases>>24026038>ive gotta admit my reaserch is pretty lacking in the specifics but i think ur on the right track and fuzzy logic systems were never intended to make shit like neural networks and i think they just wanted machines that werent limited to boolean logic.icic>decision trees are "non-parametric supervised learning algorithms" (whatever the fuck that's supposed to mean lol) and AI research was basically was all just ppl trying to create algorithms that essentially mimicked learning. as far as i understand it got kinda abandoned becaue of fuzzy logic but i think now they are starting to be used a bit again in generative AI but i haven't learned enough yet to tell u what their application is or anything.:o
>>24026038>yeah that's sorta the impression that I got which is why I became interested in this shit, I feel like there's a lot more charm to doing it the hard way and maybe im just stupid enough to think of some shit that was overlooked by people far smarter than myself that maybe could be the key to ridding ourselves of these stupid fucking "neural network" clankers that i hate so much.it's a strange state of affairsi think economics has a lot to do with why things are the way they are right now, but i know that's kinda vague lolwith neural networks, people (often finance bro types without much interest in computers) can throw together a service very quickly AS IF they had taken the energy to solve the problem & create an algorithmthere are probably much better/smarter solutions to a lot of things "AI" is being used for, but people want to get something running very quickly without doing anything esotericAI gives laymen a way to make something run very inefficiently, where previously they wouldn't have been able to make it run AT ALLit removes the high entry barriers to producing software, while letting performance/efficiency eat that cost, basicallyunsustainable in the long run, and kind of an inversion of what computers were designed for, maybe
instead of paying with time/education/money to learn to write software, as a means to creating a product, you can create the product RIGHT NOW and make everyone else eat the cost in performancebut the cost doesn't go awayit's almost like loans, when you put it that way, and reminds me of "loan culture" in the US in the 20th century, which i don't know much about but it seems badi don't know how well this metaphor works but that makes me think about interest & whether that would/could exist or be imposed/taken advantage of in some way
>>24026116all loans are scams with contracts in which you pay significantly more than you would have up front with cashbut it's impossible to swing higher things without financial assistance debtor rigged economics
>>24026133i know there are some dirty secretsi just wonder when it was rigged, to what degree, how much was just incompetence/frailty/good intentions vs how much was uhhh insidioushttps://en.wikipedia.org/wiki/Jekyll_Island#The_Jekyll_Island_Clubi just wonder
fuck sorry, this is not the right thread for this
>>24026139you know the story in the bible with the money lenders? yeah loans have always been a scam
>>24026139probably the ponzi schemes and housing crisis when boomers started using houses as monopoly pieces
>>24026151>>24026180hmmm..
>>24026183i wonder what the venn diagram of ponzi/puramid schemes looks like next to a network of influential sick people who all have dirt on each other.
>>24025580nano is my go-to terminal text editor (plz no vim wars)...think im gonna make that my terminal background while im at it :)
>>24025139>>24025529reminds me of these creatures
>By the way, yes, I was the one who studied math at UBonn. i.e. I was the one who posted https://archive.palanq.win/bant/thread/23949917/#q23965735.>I do have a job. I am a mathematics researcher. I work in arithmetic geometry, and look forward to making progress on the Langlands program later in my career. You're a fucking bakery employee, runt. You can't touch me.>I'm definitely the smartest person here. I got my PhD earlier than my peers and now I'm working on some of the hardest problems in my field of expertise. That's why I'm called "Superweapon".>I plan to try and get a professorship at Harvard one day. It might work out, it might not. Who knows? I'm already on the way to becoming an IAS fellow.
>>24026232what do you want?
>>24026226people who divided themselves from everyone on money influence ect are trying to rob us of the fact that we are deeply social creatures and in denial of the fact that we'll immediately bounce back if not subjected to their propaganda
>>24026084>there's a lot of room for hacking & ingenuity inbcollecting preparing data I see, I had never really thought about this aspect of this shit, if makes sense thought that it'd be massively important.>overfittingyeah id heard about this before, i think I even saw someone saying that it was so common with LLMs that the ppl behind them are now overfitting them deliberately because customers expect/like the results it produces. not sure how much truth there was to that though and maybe it was just their cope for how fucking shit these things are at simple tasks.>I think economics has a lot to do with why things are the way they areyeah that's almost always a driving facto behind shit getting worse xD>but people want to get something.running quickly without doing anything esotericyeah totally, its really fucking annoying that "AI" enables these fucking pricks to churn out slop. even like 10 years ago these people would have been sorta stopped in their tracks because they would have eventually had to consult someone who actually gives a fuck about the quality of their work in order to advance their hairbrained project at which point they would be told to fuck off.>AI gives laymen a way to make something run very inefficiently, where previously they wouldn't have been able to make it run AT ALL....this is what i was trying to get at above but i hadn't read this far yet. yeah this is pretty much my thoughts exactly but put forward far more eloquently than I am currently capable of with this stupid fucking migrane.>>24026116yeah not to mention the cost to the planet and other industries and shit. they need to fucking regulate this shit its ridiculous
>>24026183two very differnt answers lool.i think the concept of usury is about as old as human civilization and that's why all abrahamic religions forbid interest based loans in either some form or entirely.>>24026216rice noodle bro? are you and rave anon the same person?>>24026226yeah theres a similar sorta vibe.have u seen Nichijou?>>24026235us to tell him what a clever boy he is ¯\_(ツ)_/¯
>>24026327>rice noodle bro? are you and rave anon the same person?loool yeah, been a min since i've visited the threadidk how you put those two together so quickly
>>24026278>i think I even saw someone saying that it was so common with LLMs that the ppl behind them are now overfitting them deliberately because customers expect/like the results it produces. not sure how much truth there was to that though and maybe it was just their cope for how fucking shit these things are at simple tasks.i wouldn't be surprisedmy guess is the companies were more concerned with effectiveness at the beginning, before these things went viral & before people were dependent on theman overfitted LLM will be much more predictablea smaller one (fewer neurons and/or smaller dataset) will tend to be more unpredictable/schizophrenic, but also more versatilea smaller model, like what LLM would generally have meant before these huge consumer ones, will be more likely to spit out some strange incoherent racist ramblings or other shit which could be bad for PR & disrupt the fire displayi think in order to stamp out some of the danger talk the earlier models produced, & continue with their controlled marketing campaign, overfitted (crippled) models would be in their interestit's a very poor solution, but it's a fast/simple onethe second reason might be a more internal/corporate push to have some idea of "progress"like, "this next version will have 200 GORILLION NEURONS" might be a good way to justify your job, an excuse to buy more hardware, etc. even if the figure is actually a bad thingsupreme retardation
>a smaller model, like what LLM would generally have meant before these huge consumer ones, will be more likely to spit out some strange incoherent racist ramblings or other shit which could be bad for PR & disrupt the fire displayto put it another way, the companies can afford to release a less effective/useful model & take some backlash for thatbut they CAN'T afford the things looking ridiculous & retarded & at the same time """dangerous"""/racist, from a PR perspectiveit would pop the bubble too early, prevent them from sucking more government money by pretending they're doing some sort of arms race, and change public perception from "ooh look at this disruptive tech, i may or may not be in favor of it, but i'm kinda anxious nonetheless" to "lol">>24026278>I see, I had never really thought about this aspect of this shit, if makes sense thought that it'd be massively important.ye>yeah that's almost always a driving facto behind shit getting worse xDi'm noticing it more as i get older lol>even like 10 years ago these people would have been sorta stopped in their tracks because they would have eventually had to consult someone who actually gives a fuck about the quality of their work in order to advance their hairbrained project at which point they would be told to fuck off.very truethis probably still happens, wish i could think of examples..i assume it will go back to being that way eventually>this is what i was trying to get at above but i hadn't read this far yet.that happened to me with one of your posts earlier lol, i can't remember where>migrane.sorry
>>24026327>two very differnt answers lool.ehehi guess they could both be true, just different scales>i think the concept of usury is about as old as human civilization and that's why all abrahamic religions forbid interest based loans in either some form or entirely.i just wonder whether the problem is so simple & ancient, or whether there's other weaponization going on, whether the weaponization is also ancienti don't take anything off the table for now, but i hope to know one day>have u seen Nichijou?i meant to watch it a few times, but noit good?>us to tell him what a clever boy he is ¯\_(ツ)_/¯hehehe do be clever>>24026255checked & true
>>24026380lol. i was 99.99% sure it was you I was just waiting for u to pop up in one of these threads to ask because it felt improper to ask in spaghetti tuesday for some reason.>>24026382ahhh I see, its to fucking nerf them xD>200 GORILLION NEURONS200 gorillion ricecookers does not an artificial intelligence make.yeah i think think thats probably a big part of it, they are constantly being pushed to make progress so because its a stagnant field they are just pushing token "advancements">>24026420yeah I think ur right and its mostly just a battle to control public perception rather than to make the best product.>I assume it will go back to being that way eventuallyI really fucking hope so. I mean It will 100% it just may take a full societal and technological collapse then return to a more agrarian lifestyle >that happened to me with one of your posts earlierlol, glad its not just me.>sorry<3>>24026432the weaponization is certainly ancient. sure they may have found new ways to weaponize it but the concept of enslaving people through debt is as old as currency.>it good?yeah very, its so all over the place that its pretty much fine to just watch it at random so you wouldn't have to binge it to keep track of shit, its not exactly a sketch show but it kinda has that feel where continuity isn't prioritized. >he do be cleverI mean yeah if he is telling the truth.
I gotta go to bed soonsomeone pls bump the thread if it hits page 10 while I'm asleep.night guys <3
>>24026752mom says i'm not allowed to bump threads anymore
>>24026479yeah it makes sense, im just playng a little dumb about it tb.hi know some of you are much more perceptive than others :)
>>24026761tell her I gave u permission>>24026769well I'm glad ur here <3
>>24026761did you ask miku and teto? cuz liek, i asked them and they said you can and should>>24026790thanks anon, idk how you do it but you have the best images haha :Dnite nite im off to bed too
Due to threads like these, /bant/ is hurdling towards the state of being a flyover board like /c/, /h/, /3/ or /pw/.
>>24026939you dont think its from the perpetually bumped negro threads?
>>24027846he dosent seem to understand this is what /bant/ was made for. if these threads aren't relaxing and having fun with people from all over the world idk wtf is.probably some kf refugee who has only been here a few months and dosent have even a basic understanding of site culture.
good afternoon, thinking of rewatching di gi charat today.>>24026038>I was thinking about like how he hosted some political talks and shit so he's seen as like legitimate it's actually pretty weird how he seems to be everywhere sometimes lmaoi have to respect of how well his transition went from 2ch to TV, books, youtube, twitter, etc. it was pretty smooth, 2ch is pretty much dead these days but his career and fame is still going strong>thats pretty cool if he's moved to japan, id imagine it was prolly a big dream of his.i'm sure it was :)>I dont think he hates this site but id imagine he hates the cancer that its become full of.I always wonder how he reacted to some /pol/ shit appearing on the news in the late 2010s and so on.>I've not watched that show, ill have to check it out.it's pretty nice, i prefer nano but i wanted to mention her anyway >sleep well brazilbro <3hope you slept well too :)
>>24028164good afternoon, sorry if It takes me a while to press post on this I'm doing several things at once.never watched di gi charat but it looks cute.>how well his transition went from 2ch to TV, books, youtube, twitter, ectyeah i guess if he was just the 2ch guy maybe he wouldn't be hosting political talks ect xD I still don't really understand what it is hes even famous for, kinda seems like he's just famous for being famous like an influencer or something.>I always wonder how he reacted to the /pol/ shit appearing on the news in the late 2010sprobably a mixture of amusement and dissapointment xD like I know he hated /pol/ shit but getting donald trump elected is by far the greatest 4chan prank of all time >it's pretty nice I prefer nano.but i wanted to mention her anywaythank you for mentioning her now I have another show to add to my list :)>I hope you slept well :)yes I did thanks <3 I slept during the day instead of at night but i dosent rlly matter as I'm a neet and I managed to get up before the birds bedtime so they weren't left alone all day :)
I'll dump a few
>>24028587tyvm.very comfy <3
brazilbro do u maybe have that image thats the white pride banner but instead of the celtic cross in the middle its one of the girls from hidamari sketch (can't remember which one) and it says "wide pride world wide"
