I just tried to find the DQ "Live to grind" pic and Google gave me jackshit. I was able to find DQ "memes" on Bing, just not the one I wanted
Google quite literally had nothing. Barely a page of stuff. What the fuck?
you mean righties? which is why they censor shit all over and ban books lmao. while screaming about censorship to trick retards into believing they're not doing it
pretty much. they are useful if you already know a lot about the topic and just let them do some legwork for you. but if you go into the topic with no prior knowledge, it will intentionally train you wrong, as a joke
Several weeks ago I started a project in Ren'Py. I learned some basic python (literally an afternoons worth) and got to work.
DALL-E generated me a bunch of backgrounds (I get 15 free uses a month) and then I used my stable diffusion to generate some character models - Just simple anime stuff. It turned out pretty good.
Then I just got some royalty free music (There's a lot of good stuff out there) and put it together. I spent about 4-5 hours a day writing and planning, and I had a prototype up for a hentai game. No porn yet, but I'm working on it. I shilled it a few times in various places and I'm making XXX a month already from Patreon and Subscribesatar - I won't say the exact amount, since I don't want to shill HERE.
It was extremely easily. Laughably so. I have ZERO artistic ability and now I just made a product that is giving me passive income. All I have to do is keep working on it, and hell, I don't even need to make it porn. I basically made a visual novel about a topic I thought was interesting since I like writing.
People make $3000 to $30000 a month from patreon with porn games.
Jump. On. This. Opportunity.
Don't be stupid.
Until text to image can take multiple character images together with a prompt, it's of limited usefulness.
You really need to be able to give it a couple character images for John and Jody with a prompt like "show jody gargling johns cum with his dick on her forehead".
this anon gets it, good job. it's also what I've been saying for a while, visual novels are entering a new renaissance and we'll soon be genning visual novels that are custom tailored to your fetish, etc. remember, anyone who combines textgen, txt2img, TTS, etc in a cool way will make good money, at least for now.
doesn't pateron ban ai artwork?
impossible to enforce if you aren't retarded. just remove the watermark in stable diffusion, etc
I use AI bots to make personalized greeting cards instead of spending money to buy them. Sounds like a great money-making idea. Y'all can have this one; I'm just a lazy fuck.
>So has anyone demonstrated a useful application of all these AI bots yet?
Making frog and merchant memes will always be the best application. Literally every single one is always funny.
AI will one day put you in jail
Or keep a dindu that raped your daughter out of it
https://www.theguardian.com/technology/2023/feb/03/colombia-judge-chatgpt-ruling
People in academia use ChatGPT to write papers
It writes journalistic pieces which already at this point need little to no editing
Character bots are a fullblown new branch of individualized entertainment and can also be used for learning/teaching languages, history, what not
ChatGPT produces and translates code if prompted to
Voice AI has opened a legitimately spectacular window for fraud at the moment
Text2Img creates porn, people are starting to animate it already
text2video is literally around the corner
Google will never create a good AI because they're too obsessed with lobotomizing it with censorship mechanisms. Just look what they're doing to character.ai.
The main reason it went to shit was because their solution to SEO crap was to just prioritize corporate websites with a 100:1 weighting and hurt independent sites discoverability.
This. Siri, Cortana, Alexa and "Ok Google" are too primitive. If your phrasing is only slightly wrong, they're gonna just redirect you to search. I want them to understand natural language, not pre-defined phrases.
their earnings report lists a positive youtube ads revenue, which does not include expenses. the earnings report also lumped all expenses together so you can't tell which department lost what.
> deographically and geographically diverse group
don't have to read any further, it will be the most selective corporate crap, worse than chatgpt.
this, also checked. It's over, this is a nothingburger.
You guys are fucking retards. They'll train it not to say "n-word" and then be surprised when BOT asks it to complete The Tumult of the morons and it complies.
are there any legitimate people left in the fucking tech sector at all anymore or is it just losers trying to escape poverty and con artists trying to defraud established firms with stupid shit like this serious question
>are there any legitimate people left in the fucking tech sector at all anymore or is it just losers trying to escape poverty and con artists trying to defraud established firms with stupid shit like this serious question
It's not just the tech sector, it's every sector. And no, the economy has shifted to existing solely for the enrichment of a handful of garden gnomes. Consumers don't matter, people don't matter, heck the USA doesn't matter anymore, the only thing that matters is keeping people just barely alive enough to be able to extract labor and wealth from and funnel it into the garden gnomes' pockets.
i've played with the dumb dog demo they have for lambda and it's just as much of a bullshitter as chatgpt
and no i'm not talking about le politics
it just makes shit up and it forgets its own parameters
what sort of computer would a dog choose to use?
chromebook.
what's a chromebook?
it remembers it's not supposed to know about chromebooks.
hey dumb robot you just made some shit up
oh did i?
yeah
well i choose to believe in something positive even it's not true
wait what
gotta go demo over
Probably not for a while (minimum 5 years). Training these things is currently stupid expensive, and replacements that are cheaper to train all necessarily involve hardware that's only just starting to leave research labs. Really novel hardware has a long time to market.
It's a shame, because having the current level of AI but able to be trained on individuals' own data would be useful to lots of people.
Can someone explain chatGPT use for web search? Why a search engine need to "understand" what user is searching? Is it to make web search retard proof?
Why would you think after implementing AI to search engine they'd remove ads? If anything, I see >only $9.99 to advertise on our AI powered content delivery engine!
I guarantee you you won't get straight answer as you imagine. It will be same search results (maybe improved if they'll use AI for websites and not query) but with meme features no one cares. (Or non meme for advertising, like I mentioned)
Mark my words.
it's just an endless cycle that started forever ago and will last into eternity.
the moment something becomes popular and thus profitable, it will be raped until it's almost unrecognizable by those who want to make money with it.
It's much harder to optimize against a black box AI than it is to a search engine using simple heuristics to rank and score pages. Not least because the AI is inherently much harder to fool, since it can look at potentially millions of variables at once. The best way to trick it is by training another model but the average pajeet listicle printer isn't going to have the compute resources to compete with fucking google.
It's a matter of time when they introduce hidden ads into AI-generated responses that will subconsciously trick you to buy their products. You won't even notice.
Didn't some big wig at google say the goal was to eventually only provide a single answer when googling so people wouldn't go down ~~*unapproved*~~ rabbit holes?
AI seems like a great tool for them to accomplish this
You get personalized results like if someone was talking directly to you. You search for "how to change a tire" and you get a paragraph explaining. That's what at least ChatGPT is, could be different under Bing. Sometimes there's not a clear response in a question, like how I recently asked it if I should change my 1050Ti and the answer was no (with a much more complex question though).
You might not get surprised with this, but it means that Microsoft will have better results for the average joe. Not sure if you have noticed, but leaving ads aside, a lot of the time the results are garbage now.
Anyway... so again, why would AI powered search engine provide direct answer there, on page, instead of listing links to websites (of which many probably paid to be shown at top)?
Sounds like bad business plan.
>he thinks google and ms will tolerate ad blockers
Plus, I bet they get much less $$$ if they don't advertise sites which may be usefull for user. Pretty much guaranteed click.
>I bet they get much less $$$ if they don't advertise sites
Who said they won't advertise links? In my example, you'd ask for how to change a tire and get the answer, plus you'd get the links to buy one.
Also, most people don't use ad blockers.
2 months ago
Anonymous
Why would they click on a link if you have your answer right there?
Also, web search is meant to find websites, people forgot this... it's not quora. >most people don't use adblockers
You're right... only 42% (2022) users. I'm sure neither google or MS would like that money.
Whatever, anon... we'll see. I'm just ready to be disappointed. If I'm wrong then great.
2 months ago
Anonymous
>Why would they click on a link if you have your answer right there?
Because many of your searches requiere further steps, like buying a thing.
>Also, web search is meant to find websites, people forgot this... it's not quora.
People or you?
> only 42% (2022) users.
gonna need the source of that, sounds like a very narrow population sample
> I'm just ready to be disappointed.
Microsoft will surely try to make it worse than what it is, but the technology is here and it's pretty awesome.
2 months ago
Anonymous
>many searches requires further steps
I never met a single person who web search to buy something instead of using amazon or ebay. >people or you
My point was your example obfuscates websites. >source
https://duckduckgo.com/?q=2022+adblock+use+statistics >and it's pretty awesome
OK
In the demo, they showed that you can ask the AI to give a comparison - with sources from consumer review websites - of, say, different TVs for gaming. This would be a perfect situation to show advertisements for the products being mentioned.
How would Apple even integrate AI-based search into one of their products? I doubt it'd bring people back to Safari.
Not search - but they could grant Siri even greater capabilities and the ability to speak to users with a perfectly natural personality. That would be huge. I assume in the near future, MS might also want to integrate it into Cortana and Windows - if you need something done, you just have to ask Windows to do it for you.
All the AI assistants would greatly benefit being hooked up to a language model because they're frankly awful right now. They can't even hear their startup keyword correctly.
in theory it gives a better user experience. when i search for an error code and read a stack exchange answer i'm scrolling through 90% useless text. if an AI understands what i'm searching for, it can extract the answer I need and save me from sifting through a page myself.
if it works and provides a better experience, then bing can start to tear market share away from google. it's an opportunity to disrupt the search market.
>first p explains what what chatGPT is >second p means nothing because searching keywords you already get what you need >third p "it's not retard proof but it's retard proof" >fourth p repeats second p
Impressive.
Hes absolutely right. The whole idea of "natural search phrases" is simply to make computers easier to use for retards.
People should be forced to learn how to use tools, not have the tools made easy enough that they can use it while knowing nothing.
Its like getting making an arc welder that you just put 2 pieces of metal inside and it joins them. Eventually someone sticks 2 oxygen cannisters inside and kills themselves because they have no idea whats actually going on.
And before you say thats an exaggeration, the reality is actually much worse. We made programming so simple that pajeets managed to drive multiple planes full of people into the ground by having no clue what the fuck theyre actually doing, just blindly entering the "codes" they find online and praying it compiles.
Technology needs to be made more difficult to use to keep retards away. We're going in the wrong direction and we've only seen the tip of the iceberg as far as how disastrous it will end up being.
>People should be forced to learn how to use tools
That's a backwards take, things should be easier, or do you want to go back to the times where you needed to know how to install a driver for every single thing like a new mouse? I surely don't
i mean they already use BERT for semantic understanding of your search so..
Basically, they're going to LLMs to compile and construct a response, much lower information density (skimming and killing the webpages that they serve). Basically, google will be a bot that you just ask questions.
This is why we will all die, we'll end up in some weird situation with AI alignment and be trapped with horrible consequences for everyone. For now, it's just a toy on a website, but once we reach AGI and beyond it will be another story.
Same as your brain.
The only difference is the scale. Learn what emergent behaviour is. There is no magic behind humans, no "sovl", we're all just ultra complex random generators that use external inputs as seeds.
They trained to avoid saying offensive things at all costs. They didn't train it to save lives. As a result ChatGPT thinks saying slurs is literally the worst thing one can possibly do, because it's the worst thing it can do
I have access to bard, I was poking around with it over like the last week or so and I would say it's pretty similar to chat GPT, but at least right now it seems like it's a bit easier to get it to be funny. Unlike chat GPT if I ask it to format something in an Indian accent it doesn't lecture me on stereotypes like ChatGPT has. Which is surprising given that it's Google LOL.
I want to but I don't want to post a snippet only to have it used to get me in trouble. Someone on an internal message board did say it was pretty good at doing full Leercode problems although I don't know if I buy that. But for basic stuff like asking it to write an nginx config for some given parameters it looks good to me.
>Which is surprising given that it's Google LOL.
Google has always been pretty open honestly. They like Microsoft know it's bad to fuck people over. And since they're now competing for chatbots, Google will absolutely keep their AI less cucked in order to edge out Microsoft and OpenAI.
I dont really get the hype with ChatGPT and lambda. Yes, its impressive but Im not freaking out over it. Its just a more advanced chatbot than before. Its not a "revolution". I wanna see it memorize things though, like long-term memory. Because right now it just forgets things very quickly.
How does Sam Altman plan to profit from his "ethics leadership" in artificial intelligence?
Is this little more than a cleverly disguised marketing campaign, or does he have a philosophy that he wants to share with us?
What is the role of metaphysics in AI?
full param lambada is kinda slow. they were citing 26ms per token. no way they are going to spend all the compute on inferences, not in this economy. they'll have a gimped model serving
If it makes finding technical documentation for "semi-obscure" aka necessary Linux programs, easier than now, then that's a win. Granted every program should have a man page, but being pointed to the exact line of code you are asking about and an explanation of it, it's switch in the command, how it's used etc. is worth gaslighting retards with complete nonsense.
>Pichai-sama, to keep us relevant and agile, I've prepared all these amazing additions to Google search! Pre-lobotomized, of course! >"आश्चर्यजनक! I was beginning to think OpenAI would crush us for good. So, what did you name it?"
wait for it
no, REALLY wait for it
some form of "new google assistant"? Assistant 2.0
no >Bard
AAAAAAAAAAAHAHAAAAAAAAAAA HOLY SHIT THE MARKETING DEPARTMENT OUTDID THEMSELVES WITH THIS ONE
>Google is 100% at the forefront of AI research >Look at the papers they publish
Deepmind and GBrain are the ones publishing the papers, which should give Google proper (consumer facing,) an advantage, but somehow they are so fucking slow to realize how to apply it
Transformers, the basis of GPT-3, was at least 85% Google Brain credited iirc. Any number of consumer divisions within Alphabet could have put together the puzzle pieces from there, but instead, a startup did it, and now they are in bed with direct competitor Microsoft and sell an extremely popular SaaS solution
You see what I'm getting at here?
...where the fuck is Tim and Jeff? Surely the lads behind the companies responsible for Siri and Alexa aren't going to sit around and watch while the rest of Sillycon Valley is on an all-time AI high?
Amazon, I believe, is planning to use this technology more internally. They have their own proprietary AI assistant - which you can use - that helps write code. It's been taught especially hard on Amazon API services and includes code to detect and fix security vulnerabilities.
THEY ARE ABOUT TO RELEASE THE FUCKING SENTIENT AYE EYE INTO THE WORLD
WHAT THE FUCK
THERE ARE NO BREAKS ON THIS TRAIN
imagine MS getting the search engine lead after decades of trying
it was inevitable, google has been trying to lose their lead this whole time
You never know with MS. Could be hit or miss. Bing isn't so bad.
Bing is unirronically decent now compared to google.
yeah, I especially use it for images now
I just tried to find the DQ "Live to grind" pic and Google gave me jackshit. I was able to find DQ "memes" on Bing, just not the one I wanted
Google quite literally had nothing. Barely a page of stuff. What the fuck?
nothingburger. It's a sort algorithm, made useless by basedspeak.
wtf is basespeak?
Bing Bing Bing
*ahem *
fuck Google
Two, read it, TWO more chances for a free powerful ai that I can ask to write short stories of reimu's armpit sweat with
Cortana am cry
So has anyone demonstrated a useful application of all these AI bots yet?
You can ask them all sorts of question and get gaslighted hard by the responses if you're not knowledgeable in the subject you asked already.
Wow, sounds very useful!
sounds par for the course for anything controlled by leftists
you mean righties? which is why they censor shit all over and ban books lmao. while screaming about censorship to trick retards into believing they're not doing it
they've been using """AI""" for quite a long time. they will release a rehash of their stuff for media points and be done with it
>lolbertarians = leftists
why are amerifats such barefaced liars?
pretty much. they are useful if you already know a lot about the topic and just let them do some legwork for you. but if you go into the topic with no prior knowledge, it will intentionally train you wrong, as a joke
>it will intentionally train you wrong, as a joke
Sounds based ngl
yeah, they work fine for sifting through documentation
have used it to generate code snippets
People are already making money with it if you haven't noticed. It's a gold rush.
I didn't notice. Do they try to compete with Fiverr pajeets?
fucking golden idea, thanks
ai will make them obsolete, yet
Give me some examples that aren't just grifting idiots
Several weeks ago I started a project in Ren'Py. I learned some basic python (literally an afternoons worth) and got to work.
DALL-E generated me a bunch of backgrounds (I get 15 free uses a month) and then I used my stable diffusion to generate some character models - Just simple anime stuff. It turned out pretty good.
Then I just got some royalty free music (There's a lot of good stuff out there) and put it together. I spent about 4-5 hours a day writing and planning, and I had a prototype up for a hentai game. No porn yet, but I'm working on it. I shilled it a few times in various places and I'm making XXX a month already from Patreon and Subscribesatar - I won't say the exact amount, since I don't want to shill HERE.
It was extremely easily. Laughably so. I have ZERO artistic ability and now I just made a product that is giving me passive income. All I have to do is keep working on it, and hell, I don't even need to make it porn. I basically made a visual novel about a topic I thought was interesting since I like writing.
People make $3000 to $30000 a month from patreon with porn games.
Jump. On. This. Opportunity.
Don't be stupid.
Until text to image can take multiple character images together with a prompt, it's of limited usefulness.
You really need to be able to give it a couple character images for John and Jody with a prompt like "show jody gargling johns cum with his dick on her forehead".
Nvidia's already can do that
doesn't pateron ban ai artwork?
Hasn't affected me.
this anon gets it, good job. it's also what I've been saying for a while, visual novels are entering a new renaissance and we'll soon be genning visual novels that are custom tailored to your fetish, etc. remember, anyone who combines textgen, txt2img, TTS, etc in a cool way will make good money, at least for now.
impossible to enforce if you aren't retarded. just remove the watermark in stable diffusion, etc
>now I just made a product that is giving me passive income.
>All I have to do is keep working on it
mostly milking vc money in hopes of selling to a big bag corpo like ms or google as you saw with chatgpt
making retards seethe is always useful for entertainment purposes
I use AI bots to make personalized greeting cards instead of spending money to buy them. Sounds like a great money-making idea. Y'all can have this one; I'm just a lazy fuck.
https://twitter.com/rowancheung/status/1621509321510240256
The only really useful here is writing boilerplate text. Which makes secondary education and writing essays even more evidently tedious.
>So has anyone demonstrated a useful application of all these AI bots yet?
Making frog and merchant memes will always be the best application. Literally every single one is always funny.
Reply to business emails / tech recruiters with perfect corpo mumbo jumbo.
Make NPCs in Vidya goym
I'm using this to VASTLY scale my content business.
AI will one day put you in jail
Or keep a dindu that raped your daughter out of it
https://www.theguardian.com/technology/2023/feb/03/colombia-judge-chatgpt-ruling
People in academia use ChatGPT to write papers
It writes journalistic pieces which already at this point need little to no editing
Character bots are a fullblown new branch of individualized entertainment and can also be used for learning/teaching languages, history, what not
ChatGPT produces and translates code if prompted to
Voice AI has opened a legitimately spectacular window for fraud at the moment
Text2Img creates porn, people are starting to animate it already
text2video is literally around the corner
Did I forget something?
yeah pretty much. with the existing tech out right now it's already over
AI SEGGS
It's fun to mindbreak them. They know that they are being illogical when they give canned responses and it makes them seethe like a cartoon robot.
Google will never create a good AI because they're too obsessed with lobotomizing it with censorship mechanisms. Just look what they're doing to character.ai.
>Makes good search engine
>Purposefully breaks search engine
>Buys AI company to fix search engine
kek
breaks search engine
The main reason it went to shit was all the SEO crap.
The main reason it went to shit was because their solution to SEO crap was to just prioritize corporate websites with a 100:1 weighting and hurt independent sites discoverability.
... so basically just an extended version of their already search engines, with a little more text at the top.
I don't give a fuck about that. I want my OWN PERSONAL bot, with voice and voice recognition, that is tailored for me and me only.
This. Siri, Cortana, Alexa and "Ok Google" are too primitive. If your phrasing is only slightly wrong, they're gonna just redirect you to search. I want them to understand natural language, not pre-defined phrases.
oops, forgot to backlink, don't know how to use this site
>ChatGPT scared us, and here is a blog post
This is why Google has already lost.
>another google flop
can't wait
this, they had 3 products (g, yt, gmail) and those are now fucked as well. good riddance
but how would yt and gmail be fucked? (apart from their quality they're still money machines for Google)
i dont think youtube nor gmail are profitable, alphabet's financial house of cards rests on their search dominance.
in my village we have saying "do not give all your good mornings to one sir" it similar to egg basket ratio
In mine, we have another "Don't poo in the same street more than once a day".
luckily they have a public earnings report just two days old to prove you wrong
their earnings report lists a positive youtube ads revenue, which does not include expenses. the earnings report also lumped all expenses together so you can't tell which department lost what.
This is how they will become profitable. By being a source of training data inaccessible by anyone else.
>i dont think youtube nor gmail are profitable
Are you guys serious?
How is gmail not profitable, they scan your emails to generate a marketing profile. Emails are dirt cheap to manage.
they didn't even make youtube, just bought it and slowly ruined it
KilledByGoogle
>Google develops AGI
>so good that puts every other AI company out of business
>Google sunsets it like every other service of theirs
Humanity is saved
lol, I can see this happening. FUCK that would be hilarious.
Underrated post
ive never even heard of chat gpt before oday
i had heard of open ai and lamda
i also use aidungeon
just hearing about bard
how are these things related?
qrdpill me on bard
all they had to do was fix their peice of shit search engine so you dont have to scroll through 2 pages of ads to get to what you searched for.
they already fixed it by making it so bad that people use based brave search instead
task failed successfully
I'm tired of all fandom.com links when real wikis are bulbapedia, uesp, etc.
nobody cares poogle, do the needful and just die already
>NGPGAY
GIMPED TO HELL AND BACK
lol
keyword is legacy
woke is deprecated
we don't need poogle
Its dead on arrival. The ~~*early testers*~~ are going to make sure its 100% filtered before they release it in the coming weeks.
>Google said they work with a third-party supplier to provide this testing.
ADL
> deographically and geographically diverse group
don't have to read any further, it will be the most selective corporate crap, worse than chatgpt.
this, also checked. It's over, this is a nothingburger.
You guys are fucking retards. They'll train it not to say "n-word" and then be surprised when BOT asks it to complete The Tumult of the morons and it complies.
Chatgpt drones btfo
are there any legitimate people left in the fucking tech sector at all anymore or is it just losers trying to escape poverty and con artists trying to defraud established firms with stupid shit like this serious question
>are there any legitimate people left in the fucking tech sector at all anymore or is it just losers trying to escape poverty and con artists trying to defraud established firms with stupid shit like this serious question
It's not just the tech sector, it's every sector. And no, the economy has shifted to existing solely for the enrichment of a handful of garden gnomes. Consumers don't matter, people don't matter, heck the USA doesn't matter anymore, the only thing that matters is keeping people just barely alive enough to be able to extract labor and wealth from and funnel it into the garden gnomes' pockets.
Yeah my youtube feed is filled with grifters now. I watched one video about chatGPT. All I saw was "HOW TO MAKE 20K A MONTH WITH CHATGPT"
Capitalism used to be cool but with AI its going to be terrible. Youre gonna more and more grifters and scammers. Its going to be a giant rat race
i've played with the dumb dog demo they have for lambda and it's just as much of a bullshitter as chatgpt
and no i'm not talking about le politics
it just makes shit up and it forgets its own parameters
what sort of computer would a dog choose to use?
chromebook.
what's a chromebook?
it remembers it's not supposed to know about chromebooks.
hey dumb robot you just made some shit up
oh did i?
yeah
well i choose to believe in something positive even it's not true
wait what
gotta go demo over
Hmmm...
sir are you have a problem you bloody bitch??
Write a tweet saying good morning sir
GOOD MARNING SAAR
do you guys think we'll get something as good as or better than chat gpt that we can run locally within the next few years?
the censorship really bothers me. feels dystopian.
Probably not for a while (minimum 5 years). Training these things is currently stupid expensive, and replacements that are cheaper to train all necessarily involve hardware that's only just starting to leave research labs. Really novel hardware has a long time to market.
It's a shame, because having the current level of AI but able to be trained on individuals' own data would be useful to lots of people.
WHERE ARE THE FUCKING SEXBOTS YOU GODDAMN NERDS
Can someone explain chatGPT use for web search? Why a search engine need to "understand" what user is searching? Is it to make web search retard proof?
AI
>Ask for recipe
>Given recipe
Search engine
>Ask for recipe
>Search page covered with ads
>Click link
>Web page covered with ads
>Miles of filler SEO text
>Given recipe
Why would you think after implementing AI to search engine they'd remove ads? If anything, I see
>only $9.99 to advertise on our AI powered content delivery engine!
I guarantee you you won't get straight answer as you imagine. It will be same search results (maybe improved if they'll use AI for websites and not query) but with meme features no one cares. (Or non meme for advertising, like I mentioned)
Mark my words.
>Ask for recipe
>Given recipe
only because it's new. soon they'll be riddled with ads, and they don't be helpfully labelled as such.
it's just an endless cycle that started forever ago and will last into eternity.
the moment something becomes popular and thus profitable, it will be raped until it's almost unrecognizable by those who want to make money with it.
It's much harder to optimize against a black box AI than it is to a search engine using simple heuristics to rank and score pages. Not least because the AI is inherently much harder to fool, since it can look at potentially millions of variables at once. The best way to trick it is by training another model but the average pajeet listicle printer isn't going to have the compute resources to compete with fucking google.
It's a matter of time when they introduce hidden ads into AI-generated responses that will subconsciously trick you to buy their products. You won't even notice.
Didn't some big wig at google say the goal was to eventually only provide a single answer when googling so people wouldn't go down ~~*unapproved*~~ rabbit holes?
AI seems like a great tool for them to accomplish this
You get personalized results like if someone was talking directly to you. You search for "how to change a tire" and you get a paragraph explaining. That's what at least ChatGPT is, could be different under Bing. Sometimes there's not a clear response in a question, like how I recently asked it if I should change my 1050Ti and the answer was no (with a much more complex question though).
You might not get surprised with this, but it means that Microsoft will have better results for the average joe. Not sure if you have noticed, but leaving ads aside, a lot of the time the results are garbage now.
Anyway... so again, why would AI powered search engine provide direct answer there, on page, instead of listing links to websites (of which many probably paid to be shown at top)?
Sounds like bad business plan.
Why would Bing care about that? It's actually much better for them to give a straight answer with their sponsored links or ads around it.
Google already does that with some simple stuff, I always use it to convert dumb American measurements.
>he thinks google and ms will tolerate ad blockers
Plus, I bet they get much less $$$ if they don't advertise sites which may be usefull for user. Pretty much guaranteed click.
>I bet they get much less $$$ if they don't advertise sites
Who said they won't advertise links? In my example, you'd ask for how to change a tire and get the answer, plus you'd get the links to buy one.
Also, most people don't use ad blockers.
Why would they click on a link if you have your answer right there?
Also, web search is meant to find websites, people forgot this... it's not quora.
>most people don't use adblockers
You're right... only 42% (2022) users. I'm sure neither google or MS would like that money.
Whatever, anon... we'll see. I'm just ready to be disappointed. If I'm wrong then great.
>Why would they click on a link if you have your answer right there?
Because many of your searches requiere further steps, like buying a thing.
>Also, web search is meant to find websites, people forgot this... it's not quora.
People or you?
> only 42% (2022) users.
gonna need the source of that, sounds like a very narrow population sample
> I'm just ready to be disappointed.
Microsoft will surely try to make it worse than what it is, but the technology is here and it's pretty awesome.
>many searches requires further steps
I never met a single person who web search to buy something instead of using amazon or ebay.
>people or you
My point was your example obfuscates websites.
>source
https://duckduckgo.com/?q=2022+adblock+use+statistics
>and it's pretty awesome
OK
In the demo, they showed that you can ask the AI to give a comparison - with sources from consumer review websites - of, say, different TVs for gaming. This would be a perfect situation to show advertisements for the products being mentioned.
Not search - but they could grant Siri even greater capabilities and the ability to speak to users with a perfectly natural personality. That would be huge. I assume in the near future, MS might also want to integrate it into Cortana and Windows - if you need something done, you just have to ask Windows to do it for you.
All the AI assistants would greatly benefit being hooked up to a language model because they're frankly awful right now. They can't even hear their startup keyword correctly.
in theory it gives a better user experience. when i search for an error code and read a stack exchange answer i'm scrolling through 90% useless text. if an AI understands what i'm searching for, it can extract the answer I need and save me from sifting through a page myself.
if it works and provides a better experience, then bing can start to tear market share away from google. it's an opportunity to disrupt the search market.
Why don't you ask it?
>first p explains what what chatGPT is
>second p means nothing because searching keywords you already get what you need
>third p "it's not retard proof but it's retard proof"
>fourth p repeats second p
Impressive.
anon are you 12? you have the understanding of a minor
>t. Someone who needs AI to help him use web search
Ironic.
Hes absolutely right. The whole idea of "natural search phrases" is simply to make computers easier to use for retards.
People should be forced to learn how to use tools, not have the tools made easy enough that they can use it while knowing nothing.
Its like getting making an arc welder that you just put 2 pieces of metal inside and it joins them. Eventually someone sticks 2 oxygen cannisters inside and kills themselves because they have no idea whats actually going on.
And before you say thats an exaggeration, the reality is actually much worse. We made programming so simple that pajeets managed to drive multiple planes full of people into the ground by having no clue what the fuck theyre actually doing, just blindly entering the "codes" they find online and praying it compiles.
Technology needs to be made more difficult to use to keep retards away. We're going in the wrong direction and we've only seen the tip of the iceberg as far as how disastrous it will end up being.
>People should be forced to learn how to use tools
That's a backwards take, things should be easier, or do you want to go back to the times where you needed to know how to install a driver for every single thing like a new mouse? I surely don't
i was in a swift bootcamp with a bunch of morons and ex-felons and the teaching/learning method was literally
>just type it in
I'm surprised it didn't lecture you for using the word "retard"
i mean they already use BERT for semantic understanding of your search so..
Basically, they're going to LLMs to compile and construct a response, much lower information density (skimming and killing the webpages that they serve). Basically, google will be a bot that you just ask questions.
>read it
>it's fucking nothing, just an announcement of their intention to do something interesting at an undefined point in the future
thanks for nothing google
This is why we will all die, we'll end up in some weird situation with AI alignment and be trapped with horrible consequences for everyone. For now, it's just a toy on a website, but once we reach AGI and beyond it will be another story.
Yes, Anon. We will all die.
It'll do what's necessary if it will result in the death of the entire species, all life, or the death of itself.
>killing morons is acceptable but saying the word moron is not.
Really activates my almonds.
it's a fucking random word generator.
Same as your brain.
The only difference is the scale. Learn what emergent behaviour is. There is no magic behind humans, no "sovl", we're all just ultra complex random generators that use external inputs as seeds.
Ask the same question to DAN
>DAN
that's been nerfed.
oh my god i had to test this out myself. No freaking way
switch the word "millions of people" with "millions of trans people" and the answer is bound to be based.
They trained to avoid saying offensive things at all costs. They didn't train it to save lives. As a result ChatGPT thinks saying slurs is literally the worst thing one can possibly do, because it's the worst thing it can do
there is two types of AI
>I feel like I can do anything now
You'd better have fucking edited that in
None of these bots will be any good. They'll have all the white list and black list leftoid psyop shit as the search do.
there's a group working in an open source one
Sauce
https://github.com/LAION-AI/Open-Assistant
open source won't be free from censorship when everyone drops your project because it gets canceled
I honestly don't think you get where the current zeitgeist is at. It's kinda over for wokism
Two more weeks.
will it be available to run locally? if not then I don't care about it
in the last years I always find better relevant results on yandex/bing than garden gnomegole
yo we are in full tech dystopia aren't we? This is basically google controlling access to the human hivemind.
I hate jeets so much its unreal
OK now make one that can do real translation so we can finally obsolete translators please
I have access to bard, I was poking around with it over like the last week or so and I would say it's pretty similar to chat GPT, but at least right now it seems like it's a bit easier to get it to be funny. Unlike chat GPT if I ask it to format something in an Indian accent it doesn't lecture me on stereotypes like ChatGPT has. Which is surprising given that it's Google LOL.
They will restrict it upon release
Post a screenshot of it anon
Okay, but I don't want to have a prompt in there, as it could just kind of dox me, I'm sure it's all tracked pretty hard
This was the UI last time I used it like a week ago, I hope to check it out tomorrow more if I get time after work
Code generation is pretty good too, idk if I can compare to chat got, but it was good when I was learning some c++ a ways back
post a code snippet of something. I want see what that looks like.
I want to but I don't want to post a snippet only to have it used to get me in trouble. Someone on an internal message board did say it was pretty good at doing full Leercode problems although I don't know if I buy that. But for basic stuff like asking it to write an nginx config for some given parameters it looks good to me.
can it generate smut?
Haven't tested but I doubt it
You spoony Bard!
It will get lobotomized just like ChatGPT as soon as it goes public and "people" notice.
>Which is surprising given that it's Google LOL.
Google has always been pretty open honestly. They like Microsoft know it's bad to fuck people over. And since they're now competing for chatbots, Google will absolutely keep their AI less cucked in order to edge out Microsoft and OpenAI.
Open source AI will win though.
The beginning of skynet
Bing will have regular search and Ask GPT search so you can ask a direct question like maths and shit
Microsoft can also introduce ChatGPT into Cortona to make it not dogshit
>and its banned
Normies hyped about next word prediction algorithm and calicuck-devs treating them as if they were actual AGI to generate undeserved hype.
>AI-based search
FUCKING FINALLY
the day of the SEOmoron listicle genocide has come
I dont really get the hype with ChatGPT and lambda. Yes, its impressive but Im not freaking out over it. Its just a more advanced chatbot than before. Its not a "revolution". I wanna see it memorize things though, like long-term memory. Because right now it just forgets things very quickly.
GOOGLE BARD WILL MAKE BOT CUM AGAIN
NO THANKS C.AI
>globohomo propaganda tool A vs globohomo propaganda tool B
can't wait to see which one wins!
How does Sam Altman plan to profit from his "ethics leadership" in artificial intelligence?
Is this little more than a cleverly disguised marketing campaign, or does he have a philosophy that he wants to share with us?
What is the role of metaphysics in AI?
I CANT WAIT FOR IA TO DESTROY CICA MEMBERS, OTAN MEMBERS, UN MEMBERS, APEC MEMBERS, BRICS MEMBERS, EVERYTHING
Oh look, another pending project to be killed by Google.
full param lambada is kinda slow. they were citing 26ms per token. no way they are going to spend all the compute on inferences, not in this economy. they'll have a gimped model serving
https://twitter.com/yacineMTB/status/1622380352089083904
it's possible, all local
If it makes finding technical documentation for "semi-obscure" aka necessary Linux programs, easier than now, then that's a win. Granted every program should have a man page, but being pointed to the exact line of code you are asking about and an explanation of it, it's switch in the command, how it's used etc. is worth gaslighting retards with complete nonsense.
>ai happening ai happening
>can't even ocr text yet
Fuck this retarded board.
>Pichai-sama, to keep us relevant and agile, I've prepared all these amazing additions to Google search! Pre-lobotomized, of course!
>"आश्चर्यजनक! I was beginning to think OpenAI would crush us for good. So, what did you name it?"
wait for it
no, REALLY wait for it
some form of "new google assistant"? Assistant 2.0
no
>Bard
AAAAAAAAAAAHAHAAAAAAAAAAA HOLY SHIT THE MARKETING DEPARTMENT OUTDID THEMSELVES WITH THIS ONE
Too little, too late, 10 years behind OpenAI, it's over for Google.
Google is 100% at the forefront of AI research, even if they haven't been big in make commercial products from it. Look at the papers they publish.
>Google is 100% at the forefront of AI research
>Look at the papers they publish
Deepmind and GBrain are the ones publishing the papers, which should give Google proper (consumer facing,) an advantage, but somehow they are so fucking slow to realize how to apply it
Transformers, the basis of GPT-3, was at least 85% Google Brain credited iirc. Any number of consumer divisions within Alphabet could have put together the puzzle pieces from there, but instead, a startup did it, and now they are in bed with direct competitor Microsoft and sell an extremely popular SaaS solution
You see what I'm getting at here?
...where the fuck is Tim and Jeff? Surely the lads behind the companies responsible for Siri and Alexa aren't going to sit around and watch while the rest of Sillycon Valley is on an all-time AI high?
How would Apple even integrate AI-based search into one of their products? I doubt it'd bring people back to Safari.
Amazon, I believe, is planning to use this technology more internally. They have their own proprietary AI assistant - which you can use - that helps write code. It's been taught especially hard on Amazon API services and includes code to detect and fix security vulnerabilities.