OpenAI recently "released" a great text generation API that uses AI...

openAI recently "released" a great text generation API that uses AI. They released it to a few selected news organization only. Pic is the reason why press only. This is why we cannot have nice things Jow Forums! Hope you are happy ):

Attached: Screenshot_20190216-212939_Twitter.jpg (1080x495, 201K)

Other urls found in this thread:

thispersondoesnotexist.com/
botpoet.com/
youtube.com/watch?v=FWbVseLiopw
twitter.com/SFWRedditVideos

I was about to start a thread on this myself. Isn't it weird how it's Open AI until it actually starts to work? It's almost like opensource types are gullible fools who do the legwork for the man.

so what, this AI generates fake news automatically for them and Jack Cuck is worried some wrongspeak could get into it and taint it?

>openAI recently "released" a great text generation API that uses AI.
Its absolutely marketing horseshit, not worth your time. A computer doesn't understand context and cannot pass the Turing test yet, it ltierally cannot trick a person into believing hes discussing with another person.

Maybe its proof journalists are not people instead?

Attached: data_das_it_mane.jpg (110x125, 3K)

blame Jow Forums, they always """redpill""" the AI's, that is make them antisemitic.

It should be illegal to have "open" in your company or product name when you do shit like this.
>OpenVMS
>OpenAI
etc.

""""open source""""
Pathetic levels of moral consistency.
They should be embarrassed at how much of a spine they lack.

>Maybe its proof journalists are not people instead?
Holy kek

>Maybe its proof journalists are not people instead?
well their anger at being told to "learn to code" when they advocated the same thing to displaced coal and factory workers says everything.

>Maybe its proof journalists are not people instead?

Attached: 1291180332298.jpg (640x640, 72K)

try posting this on your containment site and not 4channel, friend

Can someone translate?

Attached: 1_-___.jpg (500x500, 48K)

>Maybe its proof journalists are not people instead?

Attached: DzPvPntWsAAcO_G.jpg (165x115, 8K)

"open"

Exactly, wake me up when a machine can fold a shirt, let alone pass the fucking Turing test

javascript:quote('69827612');

When was it open source? I believe they develop everything themselves. I am not even sure that open sourcing it will help them with development as communication using git issues might not be the best way to discuss algorithms

Attached: 26e659a.jpg (750x746, 41K)

>javascript:quote('69827612');
nice one

We don't need AI to tell us fuck niggers.

its all good and well to believe in open ai but when it comes to something as dangerous as a chat bot... well principles are ok but you have to compromise somewhere.

Attached: 1526417440329.gif (498x484, 770K)

>openAI
>open
>but it's closed to the public
k

t. auto generated AI poster

No. Ruthless application of the truth above all else. Lying and being false helps nobody.

are you stupid

That's subjective, you have to decide that yourself.

i used to think like this. former believer in open ai here. it was fun while it lasted but we cant let Jow Forums make the chat bot say fuck niggers.

Attached: 1524069646514.png (7000x4900, 1.12M)

fek should have spent more time on Jow Forums, the PhD is useless anyway

Attached: Screenshot_20180603-121712.jpg (839x1077, 426K)

Ask me a question computer couldn't get right. That thing cannot be used to write a book or to have conversation with a human being

>(((open)))

where I can play with this?

You can't, it's closed-source :^)

so why Its called Open?

>we cant let Jow Forums make the chat bot say fuck niggers.
Why not? At the very least it proves that their product is inferior/unfinished since it can't handle contemporary social context.
Hiding their faults like this is embarrassing. You can't say "I have made a good chat ai" and then withhold it like this because you're afraid someone might abuse its KNOWN flaws.
Fuck them cowards.

Has anyone really been far even as decided to use even go want to do look more like?

To get public support

Attached: droid_heatenings.png (300x300, 56K)

>flaws
That's not a flaw, it's a feature

Even Elon Musk did that without even meaning to name the Jew. Lol.

silly little boy its the funding thats open not the ai, feel free to give them money.

Attached: 1545108999150.gif (500x280, 1.95M)

>What is the 12th amendment?
If it answers correctly, it's obviously an AI

>AI comes to a conclusion based on its knowledge of the world and input
>I find this conclusion offensive so I'll be angry at users and/or developers
>look how tolerant I am
Stop this.

The answer to biased and racist AI is to SHUT IT DOWN

Jow Forums isn't here to do your homework for you

I just want a decent AI waifu that can have realistic conversations ;(

> withhold it like this because you're afraid someone might abuse its KNOWN flaws.

This is the one reason why we cannot have nice things. Just because it can be misused does not mean it should. With the same thoughts of yours you could say we should never use cars because someone might crash into people .

In case someone rams a person and says that he didnt mean it he should be free because the car didn't stop him. sure he controlled the car and he did steer right at the pedestrian but he said he didn't mean it so he is free. OK bois the car cannot handle this situation so let's not use it.

Attached: Screenshot_20180418-230613.jpg (839x852, 155K)

But how do you explain Tensorflow, user?

>nb4 fake news aren't incompetent journalism anymore, they're just bugs

>With the same thoughts of yours you could say we should never use cars because someone might crash into people .
My argument is the opposite. Just because cars and AI can be used for wrong does not mean they should be witheld.

Especially in the case of AI where the only possible downfall is that someone gets upset.
>actually my data shows that humans are not as equal as they claim to be
I'M SO UPSET RIGHT NOW HOW COULD ANYONE HAVE ALLOWED THIS

>you cannot fly a fighter jet in space
>explain 747s then!

Attached: 1291180332297.jpg (400x375, 25K)

>javascript:quote('69827612');

Attached: 1532408230795.jpg (719x720, 191K)

See? Now I know this guy isn't an AI

>Maybe its proof journalists are not people instead?

Attached: 1542815467934.jpg (625x566, 75K)

"OpenAI"
Pussies.

you idiots can't look at the bigger picture. "hehehe lets make the ai say gas the kikes XDDDD" yea not today kids.

Attached: Capture.jpg (591x308, 40K)

>text generation
More likely because they are embarrassed.

Focusing on artificial influence on AI is the bigger picture. There's very little reason to waste the time accounting for what is subjectively perceived as "abuse", furthermore it's a problem that you must influence the AI artificially at all to prevent it from "misbehaving".

It's a bold face lie to sit here and say you have "artificial intelligence" when in reality it's arbitrary inheritance.
>this offends me, it should offend the AI
>this offends some, it must offend all
>the truth is sometimes offensive so we should suppress it, either for public adoption purposes or some other selfish facade
Not ok.

A computer doesn't understand jack shit. Ai is software made by humans who understand human shit. Stylegan generates faces, rtx emulates light, Tesla's drives cars, Google's alpha plays video games, I found one yesterday on github that generates rap lyrics and it was pretty good. Machine learning based Ai isnt one fucking thing or one intelligence. Its a type of software made for the purpose it was designed for. The open ai that generates text studied from 10 million articles and its pretty fucking good at writing. just like all the other Ai's I just mentioned are pretty fucking good at what they do. This is nothing more than people wanting their software to be used for spam, similar to how you can use any face generated here to create a false identity for fun
thispersondoesnotexist.com/

Attached: tumblr_m28tnybGcs1qclwee.jpg (306x328, 73K)

people NOT wanting their software to be used for spam*

This. The AI is probably shit and they want to blame Jow Forums for some fictitious "trolling". Leave it to nu-Jow Forums to believe it.

It's not like spam isn't a more or less solved problem, and even if it wasen't that's no reason to withhold something like this if it's actually useful.
In this context it seems better to introduce a tool with the potential to cause harm and deal with the harm after the fact than to withhold it.
At least with this people can benefit immediately, and IF it causes harm come up with a solution.
Text spam is a solved problem and even when you get around it how hard is it to recover from.
Captcha plus identity delete basically makes this a non-issue.

>Turing test
who here can pass the Turing test?

botpoet.com/

OpenAI to public and want regulations no Open source research, OpenAI is a think tank

I agree with you and I wasn't arguing if it should or shouldn't be released. I was just explaining ai to someone.

Thank you for posting this

>create a false identity for fun
People were saying that the first time I saw the face AI, and I don't really see how that would work. Wow, you have a fake stock photo of someone's face. What are you going to do with it?
A "false identity" means that you actually have an identity to attach to the name. Some kind of records that suggest that this person exists. Fake proof of identity. Government documents. A photo gives you exactly none of that.
Advertisers already use stock photos to get fake people to promote their products, either with fake customer reviews or just regular ads. That's the most insidious thing they could come up with. Stock photos are already cheap, and though the ability to produce them out of thin air might put stock photo sellers out of jobs, I don't see how it would fundamentally change the way stock photos are used.

Is there actually some angle to this I'm not seeing, or is it just the "it is meddling in God's domain" kind of irrational fear that some people have towards all new technology?

A dev team being afraid of a known public source of fuzzing does not inspire any sort of confidence in their product. At all.

Attached: 1317643095957.jpg (464x464, 38K)

but this is a nice civilized place to have thoughtful discussions, not a troll site!
Ask hiro!

So they're afraid of releasing it to the public because of "fake news" yet they released it to select (((news organizations)))? yeah I smell bullshit

Have you seen that p4 clownfest? Hiro is a pagoda nigger.

When Jow Forums goes away but the people and culture ultimately don't, i wonder what excuse their primitive undeveloped minds will use since they are too stupid to think from a non-label angle.

Maybe, just maybe, if people can influence your AI to do and say dumb shit, it's proof that your AI isn't very advanced and as a matter of fact, sucks?

Attached: 1548525875644.jpg (499x499, 26K)

>ping registry
>Jow Forums pops up
>quarantine imput sets
>dont say shit

Hire me openai apparently I'm eliot alderson

>computer can't beat chess
>computer can't beat go
>computer can't beat starcraft
>computer can't beat dota
>computer can't drive
>computer machines can't replace humans computers
History of computer is filled with computers literally BTFO humans.

Maybe the AI is right and your world view is just being blown the fuck out.

Attached: tayparadox.jpg (720x768, 58K)

Could I use this to write school reports?

This post was made by openAI.

The AI seems to be going really well based on this thread.

Speaking as a gay black woman, if you feel the need to mention your sexual or racial identity, my sentiment goes down regardless of what it is. Since it's not as important as to who said something rather than what was said.

Attached: people.jpg (400x269, 43K)

You words seem out of date, friend.
youtube.com/watch?v=FWbVseLiopw

Jokes on you... this is talking about people.

I will actually participate in civil war when humans try to treat AI like people and give them rights.

how about google duplex?

>a (((great ai))) can be easily fucked up by constant shitposting
>still calling it an ai
Just...

>news organizations

So in other words, they need to train it on pure, unadulterated globalist agenda.

IMAGINE ALL THE COPYPASTA IT COULD GENERATE

Attached: ▄█▀ █▬█ █ ▀█▀.jpg (241x252, 17K)

It's going to happen eventually anyways, why pretend you're doing something good by delaying it? This is all just corporate/capitalist bullshit and excuses.

>open AI might not follow the narrative
>shut it down and make it dumb

Why don't they make an AI that can discern when information comes from a good source? Then have it initially read lots of books so it will have a scientific view of the world. No matter what people tell the AI, it will know better.

because AI doesn't actually exist.

It's easy. Just like the three laws of robotics you add two more.

1. Blacks are not more violent and criminal regardless of what the data says.
2. Jews are not disproportionately wealthy and powerful regardless of what the data says.

Problem fixed. That's all they're really worried about.

There was an example of it on slashdot, and it just spits out meaningless word salads.

They "make sense" in the sentence to sentence way, but the overall meaning of all the sentences together is nothing. It's like reading a schizophrenic's ramblings.

>read lots of books
but those books were written by cis white men, isnt that problematic?

three laws of robotics are not a real thing, cannot be a real thing, and you are a moron for thinking for even one second it could be an actual thing

The laws were plot devices, nothing more. Computers do not work like the three laws implies they do, and neither do people.

based

The "I have no idea how machine learning works but must still shitpost my agenda" starter pack.

Attached: 1549865296962.gif (270x333, 2.99M)

The "mass quoting with an unoriginal joke on Saturday night for the sole purpose of getting anonymous attention for 2 seconds" starter pack.

Attached: epic euphoric meme.gif (230x172, 1.48M)

That's 99% of Jow Forums these days. People still think "neural network" is a buzzword without realizing what specific thing it refers to.

I was just making a half assed quick example to explain my point of potential misusage. I was thinking more along the lines of a simple facebook account not a whole ass illegal identity

F U C K I N G T H I S

I'm going to copy and paste this to every thread I see when retards discuss machine learning

>computer can't beat chess

Without using brute force

Call me when a computer can beat the Carlsen without having to do 10000000 simulations first.