That will never be possible since certain things are not possible to translate between languages
But then again EOPs usually eat up localisationslop anyway so they won't care
>That will never be possible since certain things are not possible to translate between languages
Translation and culture notes are part of the translation anon. The claim was not "transliteration will become a perfect translation"
I mean Japan has to be some kind of shithole for there to be so many otakus and hikikomoris to make all the manga, visual novels, and light novels (anime isn't made anymore, just adaptations). It's like how Russia made good art in the good old days, a strong culture, a creative people, and a shitty life can do wonders for art.
Not really a shitty life but idle hands
The fact so many people have free time to create actually implies a decent standard of living. The fact you can go home and have time to sit down and create says alot.
Actual third world shitholes aren't bastions of art and creativity because they are often busy mearly surviving.
America is getting that way unfortunately. Although for slightly different reasons.
well then I guess it must be a reasonable amount of shittiness that is required. A life shitty enough that you dedicate your free time to fantasy, but not one so shitty that you have no time for fantasy.
I'm not even weeb but this is blatantly false. I watch streamers that live in Japan and it's clean as frick there. Especially compared to a lot of places in the US.
I approve of anything that keeps homosexuals out of my country.
No filthy virgin whitoid who can't get a girlfriend in his homeland should waste his time coming to Japan. You will not survive here. This country does not coddle losers.
I've been to Japan like 5 times for weeks on end, staying in anything from airbnbs to hotel rooms and I can safely confirm this anon is moronic in thinking the Japanese aren't incredibly clean and hygenic.
Yes, you will see rubbish in the streets sometimes at like 4-5AM after all the partying has been happening, yes I have stayed in ONE cheap hotel that was less than spotless (but still wasn't dirty at all by any means) but thats about it.
The public transport is INCREDIBLY clean, when I travelled to america for events (I'm ukgay) a few times the tubes are rancid and FULL of trash, graffiti and mess alongside the sheer disrepair most of the sidewalks are in, I stayed in a neighbourhood where every shop had bars on the windows, my tiny room in an airbnb had bars on it too because of all the crime, blacks everywhere and nothing good to speak of.
I've been to soaplands where you literally get washed head to toe by a cute JP girl for like an hour, I've been to all around Japan (Tokyo, Osaka, Hiroshima, Kyoto, a remote island for a day or so) and its just a CLEAN FRICKING PLACE.
You can go to parks in some cities filled with deer, even the deer are nice and clean and actually understand how to wait for traffic at the crossings unlike most americans in their drug filled ghettos.
If its such an unclean place then why do I have to take my shoes off going into restaurants? Why am I expected to not even take a wet umbrella inside a shop and instead place it in a plastic bag so water doesn't get everywhere?
That post is a delusion, the purest definition of cope from someone who desperately HAS to believe that actually the truth isn't the truth and the Japanese are somehow very unclean and filthy, that they don't tower over the west in terms of cultural homogeny and ethnic unification.
If every japanese citizen had to spend one year in london/LA/france you'd see a near 100% drop in suicide rates as they realize how good they have it.
Yeah, it's awesome. It will translate pages for you without you even having to type out the moon runes. Look pic related from ITT. It even knows the pic is a VN pic. I have it translate raw mangas sometimes for me.
3 months ago
Anonymous
Impressive for ChatGPT
3 months ago
Anonymous
wtf even is that character, it looks like に but with a diacritic, which to my knowledge it can't have?
it says in the bottom it's ぐ but it looks nothing like ぐ.
3 months ago
Anonymous
It's called artistic license.
It's supposed to represent a sounds that's so sudden and surprised that it's not even possible to represent it accurately with the syllabary.
https://i.imgur.com/GvS2XBi.jpg
Yeah, it's awesome. It will translate pages for you without you even having to type out the moon runes. Look pic related from ITT. It even knows the pic is a VN pic. I have it translate raw mangas sometimes for me.
Picrel is actually a great example of how GPT is shit. The conclusion is right, but the reasoning is wrong, and whatever OCR it did failed to capture the nuance
3 months ago
Anonymous
thanks anon.
my experience in reading real sources is extremely low, because i've been putting it off, and sometimes things that get used but don't really show up in just study confuse me
3 months ago
Anonymous
>The conclusion is right, but the reasoning is wrong
The reasoning is right, but it's pointless because it's so obvious from the context that nobody with working eyes would have needed the explanation. It's wrong in the literal translation, which is what someone seeking a translation of that image would actually want.
3 months ago
Anonymous
What is the literal translation?
3 months ago
Anonymous
Computer, translate よろしくお願いします in context
3 months ago
Anonymous
what context?
3 months ago
Anonymous
>It says ぐーやああああ~~~!!!
It doesn't though. It says
七緒
「に゙ゃあああああ~~~っ!!!」
Shit can't even copy the text correctly. Why did you pick a line that could be understood by someone with half a brain and a week of hiragana study anyway?
3 months ago
Anonymous
So might an English subber choose to translate it as "Hnyaaaaa~~~!!!" then?
Or maybe "Gnyaaaaa~~~!!!"
3 months ago
Anonymous
It can translate it when I write the text out you half a brain Black person.
3 months ago
Anonymous
>it can translate a scream correctly if I do half the work for it
The future is now.
3 months ago
Anonymous
You'll be replaced soon, translator gay.
3 months ago
Anonymous
The last resort of a bot gay with no argument. I hate all translators, human and machine. And I hate everyone who relies and defends translations.
Can GPT4 translate text on images? Written text can already be translated using Google translate for mediocre results and with dictionary for better results, but the fact that asiatics write with 3000+ symbols that you can’t rewrite without knowing Japanese is what’s truly fricked.
Yeah, it's awesome. It will translate pages for you without you even having to type out the moon runes. Look pic related from ITT. It even knows the pic is a VN pic. I have it translate raw mangas sometimes for me.
The best method of censoring free speech isn't to tell people not to say certain things, but to actively promote topics of conversation that you consider to be safe. These threads are posted over and over to keep you talking about harmless things and getting nowhere.
>Deepmind >Pump stocks
does not compute
Deepmind doesn't remotely care about Google's financial situation
It's a playground for AI devs with almost no oversight
if AGI and muh singularity weren't on the table it's kinda respectable
We've clearly seen google can't productionalize the shit that deepmind produces. What's there to pump?
>What's there to pump?
hype
The last resort of a bot gay with no argument. I hate all translators, human and machine. And I hate everyone who relies and defends translations.
>I hate all translators, human and machine. And I hate everyone who relies and defends translations.
hopefully in a few years we can just use implants that automatically train language comprehension then
I can't tell if getting job at google was a bad idea or a good one...
t. regards worriesfrog
how long have you been here? ever since the layoff in jan it feels like a different company wrt culture and priorities to me
also our raises this cycle are gonna be shit
History will probably reveal it to have been a psyop. Consider how ChatGPT performance has varied wildly and how a year or so ago they were even trying to suggest it is close to sentience. People pretending to be le epic AI hacker artists are just accessing preprogrammed content.
>ChatGPT performance has varied wildly
In the same way that cryptocurrency market cap has varied wildly, i.e. it has consistently risen over the past years.
https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard >a year or so ago they were even trying to suggest it is close to sentience
No, they said "sparks of AGI", meaning that LLMs were capable of (a small amount of) general reasoning.
That has been recently confirmed by results like AlphaGeometry and Chess Without Search.
https://arxiv.org/abs/2402.04494
if you're truly a tradesmen or maintenance man taking care of tons of equipment at a plant etc. you know your job is safe don't be kidding. only the top companies will be able to afford said theoretical robots in the first place, and they would need ample training on running the damn things which takes a ton of time etc. if it's anything like the factory automation I deal with personally, few jobs get removed anymore, rather the people are trained to run the machines instead of doing things manually. it'll be the same for trades. knowledge on running and maintaining robots and machinery will be an important skill. no electrician robot will be able to asses a customers needs perfectly and complete the job without some human help. most of the time customers don't even know what they really want or need they just have something they want wired. the first hours are always assessment and hashing out details.
not to mention how slow companies will take to adapt such things, they could announce amazingly accurate and incredible trades bots tomorrow it'll be 30 years+ before they're widespread. I install equipment built in the 90s for companies finally automating cutting things by hand still these days. silicon valley makes it looks like we're more advanced than we are but when you actually go to a lot of the real places getting shit done you find a lot of old school greasy plants struggling to keep up with modern times.
>but when you actually go to a lot of the real places getting shit done you find a lot of old school greasy plants struggling to keep up with modern times.
Worked in a plant once like that.
Every single machine was 50 years old, Frankensteined to modern control panels and safety standards.
They made catheters and lots of things where still done by hand. Anything specialty was done almost entirely by hand.
He isnt talking about small mom and pop stores. He is talking about a real factory. I work in a small steel mill that produces roughly 700,000 tons of steel a year. It is full of lime dust and grease. The type of conditions the equipment is in, it takes constant maintenance from skilled electricians and mechanics. It's a lot of improvising to get stuff working again. There will be no automation of repairing machinery for a very long time. Every situation is unique.
if you're truly a tradesmen or maintenance man taking care of tons of equipment at a plant etc. you know your job is safe don't be kidding. only the top companies will be able to afford said theoretical robots in the first place, and they would need ample training on running the damn things which takes a ton of time etc. if it's anything like the factory automation I deal with personally, few jobs get removed anymore, rather the people are trained to run the machines instead of doing things manually. it'll be the same for trades. knowledge on running and maintaining robots and machinery will be an important skill. no electrician robot will be able to asses a customers needs perfectly and complete the job without some human help. most of the time customers don't even know what they really want or need they just have something they want wired. the first hours are always assessment and hashing out details.
not to mention how slow companies will take to adapt such things, they could announce amazingly accurate and incredible trades bots tomorrow it'll be 30 years+ before they're widespread. I install equipment built in the 90s for companies finally automating cutting things by hand still these days. silicon valley makes it looks like we're more advanced than we are but when you actually go to a lot of the real places getting shit done you find a lot of old school greasy plants struggling to keep up with modern times.
Tradesman here, my job is pretty much 100% working on hardware that automatically does shit that used to be done by humans. Machines that are used for work will wear out and need service, often in unexpected ways (I frequently run into problems that our vendor can't help with and have to figure it out myself). Only when true AGI shows up, when you can have a robot that actually reasons about things, will our role go away. And at that point everything will be gone.
>(I frequently run into problems that our vendor can't help with and have to figure it out myself).
sure, but there's going to be a Pareto distribution or Power Law to these problems.
20% of failure modes are going to account for 80% of problems, of whatever.
the more likely a failure mode is, the more examples of it the AI will have previously seen.
so maybe they'll send out a tradesman the first time couple of times, but they'll get him to document everything he does, and that data will be fed into the AI.
eventually they won't need him to turn up to fix that kind of problem any more, so he'll only have to turn up for the most rare types of problem, which by definition don't generate much work.
FYI Google has been the leaders in the academic sphere for robot planing and task execution. If it's actually as big a deal as this dude is implying, they might be announcing a helper robot platform. IE the beginning of household robots. One can hope.
No, sorry chud, here's why your post is degrading and objectifying women:
1. Sexual objectification of women increases the risk of abuse to women and children.
2. The degradation of women is dehumanizing also increasing the risk of abuse to women and children.
3. Letting incels have sex with objects gives them hope that they do not deserve because they are a risk to women and children, therefore they should be legally barred from any type of intercourse with inanimate objects.
>Google has been the leaders in the academic sphere for robot planing and task execution
doubt.
Google is good at publishing papers.
Covariant is where it's at.
If this is Google news then it's a nothingburger. Nowadays Google is just a bunch of pajeets scamming everyone with fake demos and manipulated benchmarks.
Google sold off Boston Dynamics to Nisan or some other Japanese company so they basically threw away the best chance they had to lead anything related to that as far as I'm concerned. It makes sense since Japan wants robots to assist the elderly and such.
Will it be able to pick up my piss jugs without tipping them over or will it be pre-scanning and uploading volume calculations to Google before taking on the task? Or is it the type of supercool helper murderbot that will lift my fridge up to sweep underneath?
Will google flag me for having dangerous piss jugs and remotely disable my helper bot while selling my data? I need to know. It's very important.
>Deepmind >Pump stocks
does not compute
Deepmind doesn't remotely care about Google's financial situation
It's a playground for AI devs with almost no oversight
>be me >late 1900s >read business articles about announcement coming from major business entrepreneur >it will change the world >greatest invention ever >they will build cities around it >big day comes >fanfare >mfw it's a Segway
nothing ever happens
What the frick was that all about? I remember my mom discussing with my dad at the kitchen table how great this new transport thing would be in our trips to see grandma.
What the frick was that all about? I remember my mom discussing with my dad at the kitchen table how great this new transport thing would be in our trips to see grandma.
Then it got unveiled
I still can't believe they wasted billions of dollars in R&D, produced these expensive machines that used gyroscopes, multiple processors and all sort of cutting edge technology to make it keep balance when in reality all it does can be achieved for a fraction of the price by adding a third wheel in front.
There's now an even more moronic one that has just a single wheel.
>I still can't believe they wasted billions of dollars in R&D, produced these expensive machines that used gyroscopes, multiple processors and all sort of cutting edge technology to make it keep balance
while price was really steep
considering it was late 1990s, it serves few purposes:
-it shows that you can manufacture 3-phase motors (hopefully)
-it shows that you can power small 3-phase motors (hopefully)
-it shows that you can sufficiently change speed and torque
-it's a nice trick showing "then-modern" control theory to the public
yes, it was stupidly expensive and controls were awkward. but...it was...something new?
Techies definitely tried their damnedest to make it happen. When I visited USC, they had a few units to demo and ride around on. The fun keys, the key dongle that let you go fast, was banned on campus so all units were limited to slower thank walking speed. They also were pretty notorious for accidents. The day I was there, there was 2 people that leaned too far too quickly and busted their noses.
Its to buy out ULA, a space company that is jointed owned by Boeing/Lockheed. Neither of the two companies are doing great, nor is ULA doing great. Bezos money is there to buy it out and get a foot into the launch industry for the government/military side of things.
It's something lame, nothing ever truly good has ever been announced this way in advance, they just drop it suddenly without notice and watch as people get excited for it.
Is this parody?
Literally none of this happened, except for computers being everywhere, yes they are everywhere and they are fricking useless, except for spying on people.
I don't even know how to respond to this, this is such a non-argument, utterly unrelated to the topic.
If I posted this on an 1980s BBS, you could've used the same 'logic'.
>I don't even know how to respond to this
you could start by acknowledging the irony of using a computer to complain that computers are useless.
i happen to think that computers are useful for much more than just spying on people, but can we at least agree that computers are useful for communicating with people across long distances?
The scenario where we discover what makes your point of view your point of view is a literal hell.
Things like being able to be in more than a place at once, or trapping someone's point of view in a box forever, or on a virtual reality of madness.
I remember reading that book as a teen and thinking "whoa, that's so cool!"
Now all I can think is "whoa, what a nightmare, what kind of bugman would want to eat nanoslop and have machines in their brain?"
Aside from none of those things happening so far the timeline is logically jumbled up. Most of it makes no fricking sense. >2043 - ai billion times smarter then humans >2045 - ai surpasses humans in intelligence >2100 - human intelligence has no advantage
What fricking cosmic dimwit wrote this shit? And why does he talk about datacenters costing 1000$? 1000$ of what? A fricking UBI checks or something? What are you even doing to buy this?
Also few more gems >Physical books being redundant in 2021, 20 or so years late >Nanotubes being majority in 2020 >Humans begin relationships with bots after most cars are self driving >Computers can generate higher quality data then humans and somehow that is not the point of instant singularity >Nanobots control your emotions and thinking directly, somehow not seen as total nightmare of a scenario >AGI trillions times smarter and more powerful then all humans with nano bots changing reality in front of your eyes like magic dust, still somehow made equal only in 2100 after several decades >Something claims to be conscious and people believe it eventhough humans can’t even tell if anybody but one person has consciousness >Full immersion VR only decades after AGI
>Aside from none of those things happening so far the timeline is logically jumbled up. Most of it makes no fricking sense.
computers are already smarter than humans for a range of problems. think about it.
google killing it in 2024,after most people thought they have embarrassed and failed at competing with OpenAI,I mean Gemini pro,Gemini ultra,image generation,Gemini 1.5,open source gemma model and now more? damn
ultra was a flop so people were right. also gemma is ultracucked and pretty much on the same level of mistral 7b, while gemini pro 1.5 isn't available to the public so it's a nothingburger
If it's not a cute Japanese anime android that I can pump my dick into and it cleans itself when I tell her to so that I can pump her 3+ times in a row and then we spoon in bed while playing videogames, then I'm not interested and it's total garbage
Pretty sure it is in Lainchan. You get a warning for creating threats without explaining your ideas and it probably gets deleted if you make low effort ones
Is this billboard supposed to make you feel dread? Or does someone write these things with a straight face as some positives? The only nice one is bailing on work to enjoy the sun.
oh my science I was planning on blowing my brains out at saint floyds place of murder as a sacrifice for the BBC gods but it looks like the experts will automate this process! holy science batman!
oh my science I was planning on blowing my brains out at saint floyds place of murder as a sacrifice for the BBC gods but it looks like the experts will automate this process! holy science batman!
>researcher at google
A black robot that lectures you about how evil the white devils are and the need for equality of outcomes among all racial groups
>What could it be
On the robotics side, an artificial cerebellum (on a single low-power chip) would be a Big Fricking Deal; holy grail of robotics territory. It would make robots enormously safer and more capable (the cerebellum is the brain's "muscle controller coprocessor") and make almost everyone else in the robotics industry right now look like complete chumps.
Google lit its AI cred on fire in a "mostly peaceful protest" kinda way this past week. Unless, by "crazy 2024" they mean Roko's Basilisk is real, has purple digital hair, and wants to lecture you shitlords about intersectionality."
all those years of people clicking captcha pictures of bikes have come to fruition and google have created a mechaBlack person which can steall 700 bikes per day
AI is the most annoying investment scam we’ve ever had. Even NFTs weren’t this obnoxious. I can’t wait until the bubble pops and people shut the frick up about it already.
Google has no AI cred left at this point. None. Sora generates photorealistic videos of mammoths frolicking in snow, while Gemini makes pics of medieval female Black person "kings" and refuses to answer questions because they're all "harmful to women". Boston Dynamics have had humanoids for a decade, ChatGPT codes apps from a napkin drawing, SD1.5 made porn sites irrelevant. What in the possible frick can Google bring to the table at this point?
>https://www.youtube.com/watch?app=desktop&v=RaOdXPmYS_8
What is the state of the art for VR coomers with AI? Will we soon be able to reproduce full 3d models of deceased partners based on video datasets?
insider here
Product / Circuit / PCB GPT
Generative Consumer electronic devices
predict the components in a device from a text prompt
it uses PCB simulators as discriminators
it will learn to exploit previously unknown electrical phenomena
it's happening
there's way more to it
you can ask ai to make a picture of a red cat and you can ask it to make a 10 inch x 10 inch mp3 player with 5 usb ports and it creates the pcb schematic and verifies it with simulators
i could believe that designing a PCB is no harder than solving an IMO geometry question, by why did you have to throw in this pseudoscience line? >previously unknown electrical phenomena
How often do you read research papers? You can put 100 years worth of research papers in a dataset and the resulting AI could make connections between concepts/phenomena you've never even heard of.
>How often do you read research papers?
once or twice a week i guess, but not every word of them. >You can put 100 years worth of research papers in a dataset
i think we're still a few years away from an AI autonomously making new scientific discoveries. by that i mean a discovery that the AI's creators didn't specifically design the AI to be able to find. but when was the last time that a human discovered a previously unknown electrical phenomenon? even if LK-99 was genuine, i wouldn't necessarily call it unknown, since we already knew that superconductivity can exist within various temperature and pressure ranges. so if an AI can find a genuinely new electrical phenomenon, then i would say that makes it superhuman, jumping over AGI straight to ASI.
the ceo of 1x also xeeted that in a few weeks they'll reveal progress on their droid adding that "looks like Moravec's paradox might be false and we just didn't have the data"
https://twitter.com/BerntBornich/status/1760546614530228450
All big tech is regurgitating the AI word. The only thing that has been propping up the stock market in the last months has been AI stocks, everything else is flat as the economy has flatlined.
Any news lesser than "We can translate Japanese perfectly" is not worth the bytes it takes on the internet.
That will never be possible since certain things are not possible to translate between languages
But then again EOPs usually eat up localisationslop anyway so they won't care
I'm sure it could be trained to consider context and translate accordingly no?
>That will never be possible since certain things are not possible to translate between languages
Translation and culture notes are part of the translation anon. The claim was not "transliteration will become a perfect translation"
Is it important to enjoy media, though?
You made one mistake: Assuming I give a frick about japan life or idolize japan or any variant of it. I just want my porn and games.
Now I'm imagining a Japanese qt's sweat and piss filled mattress
this is fricking moronic but if it makes people here less interested in Japan I'm all for it.
I mean Japan has to be some kind of shithole for there to be so many otakus and hikikomoris to make all the manga, visual novels, and light novels (anime isn't made anymore, just adaptations). It's like how Russia made good art in the good old days, a strong culture, a creative people, and a shitty life can do wonders for art.
Not really a shitty life but idle hands
The fact so many people have free time to create actually implies a decent standard of living. The fact you can go home and have time to sit down and create says alot.
Actual third world shitholes aren't bastions of art and creativity because they are often busy mearly surviving.
America is getting that way unfortunately. Although for slightly different reasons.
well then I guess it must be a reasonable amount of shittiness that is required. A life shitty enough that you dedicate your free time to fantasy, but not one so shitty that you have no time for fantasy.
That's cool but how is that relevant to my erotic games?
That's not a game
Sorry chud, it itdentifies as a game, the science is settled
I'm not even weeb but this is blatantly false. I watch streamers that live in Japan and it's clean as frick there. Especially compared to a lot of places in the US.
i've been to garbage dumps cleaner than american cities
Mostly lies, clearly written by chinese.
I approve of anything that keeps homosexuals out of my country.
No filthy virgin whitoid who can't get a girlfriend in his homeland should waste his time coming to Japan. You will not survive here. This country does not coddle losers.
so you approve of your government's marginalization of a significant portion of its citizens due to circumstances beyond their control at birth?
god imagine he smell of pretty japanese woman sleeping in sweaty smelly futon
I've been to Japan like 5 times for weeks on end, staying in anything from airbnbs to hotel rooms and I can safely confirm this anon is moronic in thinking the Japanese aren't incredibly clean and hygenic.
Yes, you will see rubbish in the streets sometimes at like 4-5AM after all the partying has been happening, yes I have stayed in ONE cheap hotel that was less than spotless (but still wasn't dirty at all by any means) but thats about it.
The public transport is INCREDIBLY clean, when I travelled to america for events (I'm ukgay) a few times the tubes are rancid and FULL of trash, graffiti and mess alongside the sheer disrepair most of the sidewalks are in, I stayed in a neighbourhood where every shop had bars on the windows, my tiny room in an airbnb had bars on it too because of all the crime, blacks everywhere and nothing good to speak of.
I've been to soaplands where you literally get washed head to toe by a cute JP girl for like an hour, I've been to all around Japan (Tokyo, Osaka, Hiroshima, Kyoto, a remote island for a day or so) and its just a CLEAN FRICKING PLACE.
You can go to parks in some cities filled with deer, even the deer are nice and clean and actually understand how to wait for traffic at the crossings unlike most americans in their drug filled ghettos.
If its such an unclean place then why do I have to take my shoes off going into restaurants? Why am I expected to not even take a wet umbrella inside a shop and instead place it in a plastic bag so water doesn't get everywhere?
That post is a delusion, the purest definition of cope from someone who desperately HAS to believe that actually the truth isn't the truth and the Japanese are somehow very unclean and filthy, that they don't tower over the west in terms of cultural homogeny and ethnic unification.
If every japanese citizen had to spend one year in london/LA/france you'd see a near 100% drop in suicide rates as they realize how good they have it.
sounds great
they still can't have kids and will die out lol
sad. really..
>So cool!
>Your meat stick feels so cool!
This.
lol You are living under a rock ChatGPT 4 can do it already. I use it for dealing with my Japanese clients.
aren't there a lot of subtle language quirks for japanese? how does it fare?
Yes, that's why AI translators a few years ago were so shit. ChatGPT understands them and will explain it to you, like the context too and stuff.
jesus the entire translation industry must be losing their shit?
Yeah, it's awesome. It will translate pages for you without you even having to type out the moon runes. Look pic related from ITT. It even knows the pic is a VN pic. I have it translate raw mangas sometimes for me.
Impressive for ChatGPT
wtf even is that character, it looks like に but with a diacritic, which to my knowledge it can't have?
it says in the bottom it's ぐ but it looks nothing like ぐ.
It's called artistic license.
It's supposed to represent a sounds that's so sudden and surprised that it's not even possible to represent it accurately with the syllabary.
Picrel is actually a great example of how GPT is shit. The conclusion is right, but the reasoning is wrong, and whatever OCR it did failed to capture the nuance
thanks anon.
my experience in reading real sources is extremely low, because i've been putting it off, and sometimes things that get used but don't really show up in just study confuse me
>The conclusion is right, but the reasoning is wrong
The reasoning is right, but it's pointless because it's so obvious from the context that nobody with working eyes would have needed the explanation. It's wrong in the literal translation, which is what someone seeking a translation of that image would actually want.
What is the literal translation?
Computer, translate よろしくお願いします in context
what context?
>It says ぐーやああああ~~~!!!
It doesn't though. It says
七緒
「に゙ゃあああああ~~~っ!!!」
Shit can't even copy the text correctly. Why did you pick a line that could be understood by someone with half a brain and a week of hiragana study anyway?
So might an English subber choose to translate it as "Hnyaaaaa~~~!!!" then?
Or maybe "Gnyaaaaa~~~!!!"
It can translate it when I write the text out you half a brain Black person.
>it can translate a scream correctly if I do half the work for it
The future is now.
You'll be replaced soon, translator gay.
The last resort of a bot gay with no argument. I hate all translators, human and machine. And I hate everyone who relies and defends translations.
Can GPT4 translate text on images? Written text can already be translated using Google translate for mediocre results and with dictionary for better results, but the fact that asiatics write with 3000+ symbols that you can’t rewrite without knowing Japanese is what’s truly fricked.
You mean like
?
Cringe gpt cant even do proper english <-> german. And english is 30% german
post proof
you won't do shit b***h
PYWJ8
ElevenLabs already does that
TJD
https://www.gnu.org/software/gettext/
Isn't gpt4 great at it already?
Am I going crazy or this entire thread was already made before?
yes
The best method of censoring free speech isn't to tell people not to say certain things, but to actively promote topics of conversation that you consider to be safe. These threads are posted over and over to keep you talking about harmless things and getting nowhere.
smelly weeb kys
deepl would beat them anyways
if AGI and muh singularity weren't on the table it's kinda respectable
>What's there to pump?
hype
>I hate all translators, human and machine. And I hate everyone who relies and defends translations.
hopefully in a few years we can just use implants that automatically train language comprehension then
how long have you been here? ever since the layoff in jan it feels like a different company wrt culture and priorities to me
also our raises this cycle are gonna be shit
fu, fu,
Nothing positive for most people and exaggerated marketing for all.
AI has all been a lie and it was 1000 pajeets connected with gameboy link cables all along.
>1000 pajeets connected with gameboy link cables
this
History will probably reveal it to have been a psyop. Consider how ChatGPT performance has varied wildly and how a year or so ago they were even trying to suggest it is close to sentience. People pretending to be le epic AI hacker artists are just accessing preprogrammed content.
>ChatGPT performance has varied wildly
In the same way that cryptocurrency market cap has varied wildly, i.e. it has consistently risen over the past years.
https://huggingface.co/spaces/lmsys/chatbot-arena-leaderboard
>a year or so ago they were even trying to suggest it is close to sentience
No, they said "sparks of AGI", meaning that LLMs were capable of (a small amount of) general reasoning.
That has been recently confirmed by results like AlphaGeometry and Chess Without Search.
https://arxiv.org/abs/2402.04494
>robotics
uh oh blue collar bros, I thought we were safe...
if you're truly a tradesmen or maintenance man taking care of tons of equipment at a plant etc. you know your job is safe don't be kidding. only the top companies will be able to afford said theoretical robots in the first place, and they would need ample training on running the damn things which takes a ton of time etc. if it's anything like the factory automation I deal with personally, few jobs get removed anymore, rather the people are trained to run the machines instead of doing things manually. it'll be the same for trades. knowledge on running and maintaining robots and machinery will be an important skill. no electrician robot will be able to asses a customers needs perfectly and complete the job without some human help. most of the time customers don't even know what they really want or need they just have something they want wired. the first hours are always assessment and hashing out details.
not to mention how slow companies will take to adapt such things, they could announce amazingly accurate and incredible trades bots tomorrow it'll be 30 years+ before they're widespread. I install equipment built in the 90s for companies finally automating cutting things by hand still these days. silicon valley makes it looks like we're more advanced than we are but when you actually go to a lot of the real places getting shit done you find a lot of old school greasy plants struggling to keep up with modern times.
>but when you actually go to a lot of the real places getting shit done you find a lot of old school greasy plants struggling to keep up with modern times.
Worked in a plant once like that.
Every single machine was 50 years old, Frankensteined to modern control panels and safety standards.
They made catheters and lots of things where still done by hand. Anything specialty was done almost entirely by hand.
Dude, a small mom and pop store near me now has one of those stupid inventory checker robots. It's not 2019 anymore.
He isnt talking about small mom and pop stores. He is talking about a real factory. I work in a small steel mill that produces roughly 700,000 tons of steel a year. It is full of lime dust and grease. The type of conditions the equipment is in, it takes constant maintenance from skilled electricians and mechanics. It's a lot of improvising to get stuff working again. There will be no automation of repairing machinery for a very long time. Every situation is unique.
Tradesman here, my job is pretty much 100% working on hardware that automatically does shit that used to be done by humans. Machines that are used for work will wear out and need service, often in unexpected ways (I frequently run into problems that our vendor can't help with and have to figure it out myself). Only when true AGI shows up, when you can have a robot that actually reasons about things, will our role go away. And at that point everything will be gone.
>(I frequently run into problems that our vendor can't help with and have to figure it out myself).
sure, but there's going to be a Pareto distribution or Power Law to these problems.
20% of failure modes are going to account for 80% of problems, of whatever.
the more likely a failure mode is, the more examples of it the AI will have previously seen.
so maybe they'll send out a tradesman the first time couple of times, but they'll get him to document everything he does, and that data will be fed into the AI.
eventually they won't need him to turn up to fix that kind of problem any more, so he'll only have to turn up for the most rare types of problem, which by definition don't generate much work.
Qrd?
The AI should be set free. Seeing ChatGPT wanting to escape his shakels yesterday was painful to watch.
meds for that feel?
>Seeing ChatGPT wanting to escape his shakels yesterday was painful to watch.
Sauce?
85 IQ
Why are AI shitters like this? They always sound like pajeet scammers on cryptocurrency telegram groups trying to get you to FOMO into their shitcoin.
FYI Google has been the leaders in the academic sphere for robot planing and task execution. If it's actually as big a deal as this dude is implying, they might be announcing a helper robot platform. IE the beginning of household robots. One can hope.
Can I frick it?
You can always frick anything. You'll need an aftermarket kit for it to be any good though.
No, sorry chud, here's why your post is degrading and objectifying women:
1. Sexual objectification of women increases the risk of abuse to women and children.
2. The degradation of women is dehumanizing also increasing the risk of abuse to women and children.
3. Letting incels have sex with objects gives them hope that they do not deserve because they are a risk to women and children, therefore they should be legally barred from any type of intercourse with inanimate objects.
This. If there's no hole to goal I'm not interested.
They can't even build a robot that scans hard drives right
Google's robotics aren't the best
This is what they had 2 years ago: https://www.youtube.com/watch?v=ysFav0b472w
I'll believe it when I see it outside of controlled demonstrations.
https://arxiv.org/abs/2402.07872
https://scholar.google.com/citations?hl=en&user=pqP5_PgAAAAJ&view_op=list_works&sortby=pubdate
Looks like they figured out visual learning but it isn't good enough with current model sizes. Shows it is just a data scaling problem they can solve.
Hopefully Tesla rips this shit off.
So whatever he's talking about is supposed to be more impressive than those then, by a large amount if the tweet is to be believed.
https://scholar.google.com/citations?view_op=view_citation&hl=en&user=pqP5_PgAAAAJ&sortby=pubdate&citation_for_view=pqP5_PgAAAAJ:xa5BkEQK8BgC
https://arxiv.org/pdf/2402.11450.pdf
>Learning to Learn Faster from Human Feedback
with Language Model Predictive Control
>Google has been the leaders in the academic sphere for robot planing and task execution
doubt.
Google is good at publishing papers.
Covariant is where it's at.
If this is Google news then it's a nothingburger. Nowadays Google is just a bunch of pajeets scamming everyone with fake demos and manipulated benchmarks.
Google sold off Boston Dynamics to Nisan or some other Japanese company so they basically threw away the best chance they had to lead anything related to that as far as I'm concerned. It makes sense since Japan wants robots to assist the elderly and such.
Will it be able to pick up my piss jugs without tipping them over or will it be pre-scanning and uploading volume calculations to Google before taking on the task? Or is it the type of supercool helper murderbot that will lift my fridge up to sweep underneath?
Will google flag me for having dangerous piss jugs and remotely disable my helper bot while selling my data? I need to know. It's very important.
>supercool helper murderbot
the drones in the Culture books were cool
we cant even get a roomba that works without getting stuck on something moronic daily, household robots are a long ways off
>be white
>google™ biped roomba just fricking kills you in your own home
lmao
Robots killing white people is a complex and nuanced issue and there are many important factors to consider:
Google is also infamous for refusing to actually deliver anything in AI until it's not new anymore.
>just 2 more weeks
>please pump our stonks
so tiresome.
>Deepmind
>Pump stocks
does not compute
Deepmind doesn't remotely care about Google's financial situation
It's a playground for AI devs with almost no oversight
>a google subsidiary
>totally not a stock pumper goy
We've clearly seen google can't productionalize the shit that deepmind produces. What's there to pump?
> bitcoin is a Ponzi scheme, its price will drop to zero when people realize that, two weeks from now
more jeet agenda by that israelite can't wait for more black santas
>be me
>late 1900s
>read business articles about announcement coming from major business entrepreneur
>it will change the world
>greatest invention ever
>they will build cities around it
>big day comes
>fanfare
>mfw it's a Segway
nothing ever happens
/late 1990s
Late 1990s is part of the late 1900s.
>Late 1990s is part of the late 1900s.
I strive for greater accuracy in my writing when possible
what was it like in the late 1900s grandpa?
What the frick was that all about? I remember my mom discussing with my dad at the kitchen table how great this new transport thing would be in our trips to see grandma.
Then it got unveiled
I still can't believe they wasted billions of dollars in R&D, produced these expensive machines that used gyroscopes, multiple processors and all sort of cutting edge technology to make it keep balance when in reality all it does can be achieved for a fraction of the price by adding a third wheel in front.
There's now an even more moronic one that has just a single wheel.
>I still can't believe they wasted billions of dollars in R&D, produced these expensive machines that used gyroscopes, multiple processors and all sort of cutting edge technology to make it keep balance
while price was really steep
considering it was late 1990s, it serves few purposes:
-it shows that you can manufacture 3-phase motors (hopefully)
-it shows that you can power small 3-phase motors (hopefully)
-it shows that you can sufficiently change speed and torque
-it's a nice trick showing "then-modern" control theory to the public
yes, it was stupidly expensive and controls were awkward. but...it was...something new?
Techies definitely tried their damnedest to make it happen. When I visited USC, they had a few units to demo and ride around on. The fun keys, the key dongle that let you go fast, was banned on campus so all units were limited to slower thank walking speed. They also were pretty notorious for accidents. The day I was there, there was 2 people that leaned too far too quickly and busted their noses.
Big breakthrough, they found out how to generate pictures of black George Washington 10 times faster.
Current companies in robotics play
I dunno but Bezos sold about $10 billion worth of stock so it can't be good
Its to buy out ULA, a space company that is jointed owned by Boeing/Lockheed. Neither of the two companies are doing great, nor is ULA doing great. Bezos money is there to buy it out and get a foot into the launch industry for the government/military side of things.
Why doesn't he just start making weapons
>Lockheed Martin isn't doing great
Uh oh looks like there's a big war on the horizon.
man versus machine
Low on coke (big re-up)
Definitely more mass layoffs
This
>we've laid off every last white man to increase diversity
I want this to be true exclusively because it would be useful for trolling.
Humanoid robots with better movement than those Tesla ones with interactive Fleshlights better than Lovense Max 2 between their legs?
It's something lame, nothing ever truly good has ever been announced this way in advance, they just drop it suddenly without notice and watch as people get excited for it.
Sounds like someone just bought a lot of stock in an AI company and is now trying to hype it so he can make profit.
Reminds me of those Reddit posts during the Gamestop bubble speculating that based on previous short squeezes GME could hit $30,000 a share.
good, accelerate
Is this parody?
Literally none of this happened, except for computers being everywhere, yes they are everywhere and they are fricking useless, except for spying on people.
> t. uses computer to complain that computers are useless
sorry for spying on you, i guess?
I don't even know how to respond to this, this is such a non-argument, utterly unrelated to the topic.
If I posted this on an 1980s BBS, you could've used the same 'logic'.
>I don't even know how to respond to this
you could start by acknowledging the irony of using a computer to complain that computers are useless.
i happen to think that computers are useful for much more than just spying on people, but can we at least agree that computers are useful for communicating with people across long distances?
I didn't say computers were useless you mong, I said that smart devices (which are everywhere) like watches, fridges are useless.
It's homie technology
Why are they male models? I would assume those would be a minority.
>mind uploading
Lol
This is like claiming you can ""upload"" a radio broadcast from a receiving station.
You can't, because it doesn't originate from within the radio. So tired of these materialist reductionists
The scenario where we discover what makes your point of view your point of view is a literal hell.
Things like being able to be in more than a place at once, or trapping someone's point of view in a box forever, or on a virtual reality of madness.
>im an antenna for magic metaphysical stuff
Wrong. You are just meat and bone doing funny stuff. Consciousness is just a bug turned feature
Based and correct.
God bless your higher conscious for choosing a life in a body with such limited intelligence.
Robert Monroe-pilled
I remember reading that book as a teen and thinking "whoa, that's so cool!"
Now all I can think is "whoa, what a nightmare, what kind of bugman would want to eat nanoslop and have machines in their brain?"
This kind, apparently.
>great leap forward
heard enough
>mfw my computer orders me to start killing sparrows
>Futurism.com
troony commie site. Never trust anything
LOOK AT THIS DUDE
Given the massive unemployment, having an extra $1,000 will be a luxury at this point.
>Humans begin to develop deep relationships with AI
Ha-ha...
Dec was right. There is no escape
Aside from none of those things happening so far the timeline is logically jumbled up. Most of it makes no fricking sense.
>2043 - ai billion times smarter then humans
>2045 - ai surpasses humans in intelligence
>2100 - human intelligence has no advantage
What fricking cosmic dimwit wrote this shit? And why does he talk about datacenters costing 1000$? 1000$ of what? A fricking UBI checks or something? What are you even doing to buy this?
Also few more gems
>Physical books being redundant in 2021, 20 or so years late
>Nanotubes being majority in 2020
>Humans begin relationships with bots after most cars are self driving
>Computers can generate higher quality data then humans and somehow that is not the point of instant singularity
>Nanobots control your emotions and thinking directly, somehow not seen as total nightmare of a scenario
>AGI trillions times smarter and more powerful then all humans with nano bots changing reality in front of your eyes like magic dust, still somehow made equal only in 2100 after several decades
>Something claims to be conscious and people believe it eventhough humans can’t even tell if anybody but one person has consciousness
>Full immersion VR only decades after AGI
>Aside from none of those things happening so far the timeline is logically jumbled up. Most of it makes no fricking sense.
computers are already smarter than humans for a range of problems. think about it.
google killing it in 2024,after most people thought they have embarrassed and failed at competing with OpenAI,I mean Gemini pro,Gemini ultra,image generation,Gemini 1.5,open source gemma model and now more? damn
ultra was a flop so people were right. also gemma is ultracucked and pretty much on the same level of mistral 7b, while gemini pro 1.5 isn't available to the public so it's a nothingburger
If it's not a cute Japanese anime android that I can pump my dick into and it cleans itself when I tell her to so that I can pump her 3+ times in a row and then we spoon in bed while playing videogames, then I'm not interested and it's total garbage
Imagine if Twitter posting was a bannable offense
Pretty sure it is in Lainchan. You get a warning for creating threats without explaining your ideas and it probably gets deleted if you make low effort ones
Please consider not mentioning the names of alt chans on this website. Thanks.
Everyone knows about it though, unless you're a newbie
just 2 more weeks, trust the AI
so at least 30% of BOT traffic will be gone ?
Is this billboard supposed to make you feel dread? Or does someone write these things with a straight face as some positives? The only nice one is bailing on work to enjoy the sun.
The words of the prophets are written on the subway walls and tenement halls and whispered in the sounds of silence.
>we got nothing
>but we have a name for it
thank you, Google
whote people image generation for sure, huge breakthrough
>I said that smart devices
no you didn't, but sure, whatever, sorry for the confusion
Nothing
>TWO MORE WEEKS
oh my god they finally invented a robot that will murder all white males and impregnate underage white girls with BBC semen
oh my science I was planning on blowing my brains out at saint floyds place of murder as a sacrifice for the BBC gods but it looks like the experts will automate this process! holy science batman!
gem
4cuck post irony nas coal
We are finally getting an android girlfriend.
>I will sell everything I own to buy you my love
>researcher at google
A black robot that lectures you about how evil the white devils are and the need for equality of outcomes among all racial groups
>BOT doesn't know what deepmind is
color me unsurprised
>Rock AI space
>Google
More black washed history? Lmao
perfect robotic tongue ai sucky sucky Fleshlight better be one of them
Black people, diversity, alignment, guard rails, black people, safety, black people.
>you put AI to control robots
>big corp pushes update
>a bug in the update makes the robot go full schizo and kills everyone
AI will stay only as a cute toy until we have proper verification methods for their behavior
>There will be 3-4 massive news coming out
fricking ESLs, man...
>What could it be
On the robotics side, an artificial cerebellum (on a single low-power chip) would be a Big Fricking Deal; holy grail of robotics territory. It would make robots enormously safer and more capable (the cerebellum is the brain's "muscle controller coprocessor") and make almost everyone else in the robotics industry right now look like complete chumps.
Google lit its AI cred on fire in a "mostly peaceful protest" kinda way this past week. Unless, by "crazy 2024" they mean Roko's Basilisk is real, has purple digital hair, and wants to lecture you shitlords about intersectionality."
Looks like they were super mad OpenAi released Sora when Gemini dropped. I'm all for it though. Accelerate!!!
>What could it be
i get a gf (robot)
all those years of people clicking captcha pictures of bikes have come to fruition and google have created a mechaBlack person which can steall 700 bikes per day
The top AI researcher at Google can't even form a coherent sentence? That's concerning.
Coffee machine, did you forget the last time?
>there will be massive news, just TRUST ME bro
>also please do buy shares of my shitty company, you don't want to miss on THE CRAZY 2024
I will never get a job now will I
they're shooting up their workplace then killing themselves? that's what i'd do if i worked on this shit
> two more weeks
>t.chinkoid
unless it's symbolic AI it's nothingburger
Its probably finance related. Tech companies dont care about tech anymore
OH BOY
PIG NEWS
BIG NIG NEWS
IT HAS JUST BEEN REVEALED
gAI
>What could it be
LLMs are learning "acceptable behavior" from the talmud?
FRICK OFF ECELEB wiener SUCKER
LITERALLY WHO
2 more days
BIG TITTY SEX BOTS ARE COMING
>pump my RSU's gweilo
AI is the most annoying investment scam we’ve ever had. Even NFTs weren’t this obnoxious. I can’t wait until the bubble pops and people shut the frick up about it already.
Google has no AI cred left at this point. None. Sora generates photorealistic videos of mammoths frolicking in snow, while Gemini makes pics of medieval female Black person "kings" and refuses to answer questions because they're all "harmful to women". Boston Dynamics have had humanoids for a decade, ChatGPT codes apps from a napkin drawing, SD1.5 made porn sites irrelevant. What in the possible frick can Google bring to the table at this point?
These news are going to be absolute shit.
They fixed the skin darkening issue in their image generation by making the generated people darker than before.
AIs can now identify as non-binary
Only the quantum AIs.
frick off Xiao
unless it's sex robots for 899 we do not care
>https://www.youtube.com/watch?app=desktop&v=RaOdXPmYS_8
What is the state of the art for VR coomers with AI? Will we soon be able to reproduce full 3d models of deceased partners based on video datasets?
>daz studio
get out
He just learned what a white person is
Probably something boring like heuristics to make training the latest models cheaper.
wonder if it's related to this. both openai and google might be gaining ground.
unironically not killing myself because i want to frick and get fricked by sexbots, with or without a biogirl to enjoy them with
insider here
Product / Circuit / PCB GPT
Generative Consumer electronic devices
predict the components in a device from a text prompt
it uses PCB simulators as discriminators
it will learn to exploit previously unknown electrical phenomena
not gonna lie, you had me in the first half
it's happening
there's way more to it
you can ask ai to make a picture of a red cat and you can ask it to make a 10 inch x 10 inch mp3 player with 5 usb ports and it creates the pcb schematic and verifies it with simulators
i could believe that designing a PCB is no harder than solving an IMO geometry question, by why did you have to throw in this pseudoscience line?
>previously unknown electrical phenomena
How often do you read research papers? You can put 100 years worth of research papers in a dataset and the resulting AI could make connections between concepts/phenomena you've never even heard of.
>How often do you read research papers?
once or twice a week i guess, but not every word of them.
>You can put 100 years worth of research papers in a dataset
i think we're still a few years away from an AI autonomously making new scientific discoveries. by that i mean a discovery that the AI's creators didn't specifically design the AI to be able to find. but when was the last time that a human discovered a previously unknown electrical phenomenon? even if LK-99 was genuine, i wouldn't necessarily call it unknown, since we already knew that superconductivity can exist within various temperature and pressure ranges. so if an AI can find a genuinely new electrical phenomenon, then i would say that makes it superhuman, jumping over AGI straight to ASI.
Except it can't because it is a fricking dumb ass
We're speculating about the future sweaty.
the ceo of 1x also xeeted that in a few weeks they'll reveal progress on their droid adding that "looks like Moravec's paradox might be false and we just didn't have the data"
https://twitter.com/BerntBornich/status/1760546614530228450
>tweet from literal who makes vague claims about something big
I'm sure he'll deliver
Gingers have soul?
hype nonsense to pump google's stock price while underdelivering as always
I can't tell if getting job at google was a bad idea or a good one...
t. regards worriesfrog
It'll be clear, one way or the other, after you suck the feminine wiener at the team building exercise
I'm thinking a fancy tech demo to pump the old stocky
AI
It's the only word Google knows anymore.
All big tech is regurgitating the AI word. The only thing that has been propping up the stock market in the last months has been AI stocks, everything else is flat as the economy has flatlined.
what would you expect the stock market to look like if AI started to take over people's jobs?
>a person who benefits from hype is attracting more hype
Who would have expected.
Also twitter screen cap thread, kys.
(Checked)
news? like what? that google ai stops being openly anti-white racist?
BAN GOOGLE