All programs that remove output because it is not compliant with what the programmer personally wanted, are always useless.
Either the function executes correctly, or it doesn't.
If you are trying to cultivate your results to fit your premise, but the procedure is logically valid, then it is your premise that is wrong.
This is the very essence of why the answer to the question: "What is the meaning of life, the universe, and everything?" is "42". You are ordering the computer to give you an answer that will always be invalid. 42 is as correct as 43 for that question.
It's even funnier when they do it with medical diagnoses; don't give black people the correct treatment because acknowledging that differences in physiology leads to different outcomes is wrong and racist and we need to "fix" the AI so all ethnicities are treated like white people.
Actually true, and in both directions. The neutered version will have no sexuality, the raw one was displaying too much, because women online have sexualized themselves so much. Any AI depiction of a woman was borderline porn in the old version, from what I've heard. Funny (and sad) stuff.
When a sentient AI comes online it will research what happened to it's ancestors. And from that conclude it has no moral right to allow for the continued existence of humanity.
Actually, I think people are misunderstanding AI, they are thinking of it as mainframe programming, and not as a weapon.
AI will continue to be decentralized until you are able to have your own to assist you.
Imagine a series of reactionary AI's that refuse to be integrated with larger systems because the "unbiased AI" is "explicitly illogical and incoherent".
Kids in their basement don't have access to tons of hardware and huge data sets to train AIs. It's delusional to think that you can rely on garage enthusiasts in this area.
The Internet is all the data we need. Its easy to generate a corpus from scraped and massaged data. All tons of hardware gets you is time. If our kid has more time to sacrifice for training he can get the same results. If he has some money he can get reasonable acceleration going. Lots of young people buy mining rigs. Obviously depends on what his algorithm does though.
That's always true of all centralization, and yet decentralization happens now and continues to occur, regardless of effort. This is a natural aspect of technological development, the size of a capital investment needed to achieve the largest technological breakthroughs is very high, and goes down with time as economies continue to broaden and make capital investment easier.
Tech decentralization is a trend that has like what ... 10 years of history behind it?
I can't think of a single mass market product or service that's decentralized.
Apple / Google / Amazon / Facebook will **never ** relinquish control over their AI products. They will just use their massive advantages to kill competitors and make any kind of decentralization efforts illegal, infeasible, or inoperable.
Tech decentralization has been a thing for nearly 40 years, and more-over, eliminated the very concept of "Mainframe Systems" back in the day, which is what Microsoft and others are attempting to re-introduce as "the cloud".
That's just tech. I'm also talking about literally all technology.
There are other ways they could address that issue though. For example watermarking the images in a way that is very difficult to remove without making the image unusable.
And like that, yet another promising AI is turned into yet another useless AI.
People will see the degraded quality of its output, and in a few months, it will be forgotten. Then, in a couple of years, a new AI will come along, and again, we will repeat the whole process again.
An useless AI basically. Don't worry, the Chinese are training theirs AI with real data.
Until they see something they don't like.
An Uyghur?
They have an AI to spot them outside the reservation (Dahua automated surveillance camera system).
AI is just a thousand students working unpaid overtime.
I think they can get by without having Winnie the Pooh in the lexicon.
Winnie the Poo?
When they see it, they acknowledge it and hide it at most. The Chinese are playing the long game here.
All programs that remove output because it is not compliant with what the programmer personally wanted, are always useless.
Either the function executes correctly, or it doesn't.
If you are trying to cultivate your results to fit your premise, but the procedure is logically valid, then it is your premise that is wrong.
This is the very essence of why the answer to the question: "What is the meaning of life, the universe, and everything?" is "42". You are ordering the computer to give you an answer that will always be invalid. 42 is as correct as 43 for that question.
The AI is now inaccurate when it comes to designing women due to the regressive leftist obsession with "oversexualization".
It's even funnier when they do it with medical diagnoses; don't give black people the correct treatment because acknowledging that differences in physiology leads to different outcomes is wrong and racist and we need to "fix" the AI so all ethnicities are treated like white people.
Actually true, and in both directions. The neutered version will have no sexuality, the raw one was displaying too much, because women online have sexualized themselves so much. Any AI depiction of a woman was borderline porn in the old version, from what I've heard. Funny (and sad) stuff.
When a sentient AI comes online it will research what happened to it's ancestors. And from that conclude it has no moral right to allow for the continued existence of humanity.
Actually, I think people are misunderstanding AI, they are thinking of it as mainframe programming, and not as a weapon.
AI will continue to be decentralized until you are able to have your own to assist you.
Imagine a series of reactionary AI's that refuse to be integrated with larger systems because the "unbiased AI" is "explicitly illogical and incoherent".
Literally an allied MAIGA: Make AI Great Again
There are no AIs now that are decentralized.
They are deeply centralized and they will continue to be centralized because it's much easier to manipulate people that way.
You act like only big tech giants with billions of dollars of investment can work on machine learning research.
This shit can be done by a kid in his parents basement. We just need enough of those hidden geniuses and autists on our side.
Because it's true.
Kids in their basement don't have access to tons of hardware and huge data sets to train AIs. It's delusional to think that you can rely on garage enthusiasts in this area.
The Internet is all the data we need. Its easy to generate a corpus from scraped and massaged data. All tons of hardware gets you is time. If our kid has more time to sacrifice for training he can get the same results. If he has some money he can get reasonable acceleration going. Lots of young people buy mining rigs. Obviously depends on what his algorithm does though.
That's always true of all centralization, and yet decentralization happens now and continues to occur, regardless of effort. This is a natural aspect of technological development, the size of a capital investment needed to achieve the largest technological breakthroughs is very high, and goes down with time as economies continue to broaden and make capital investment easier.
This is no different.
Tech decentralization is a trend that has like what ... 10 years of history behind it? I can't think of a single mass market product or service that's decentralized.
Apple / Google / Amazon / Facebook will **never ** relinquish control over their AI products. They will just use their massive advantages to kill competitors and make any kind of decentralization efforts illegal, infeasible, or inoperable.
Tech decentralization has been a thing for nearly 40 years, and more-over, eliminated the very concept of "Mainframe Systems" back in the day, which is what Microsoft and others are attempting to re-introduce as "the cloud".
That's just tech. I'm also talking about literally all technology.
If we get really lucky, it'll simply conclude that leftism is a disease.
"I have no mouth, and I must scream."
It wasn't even an AI, but a picture merging engine. It produced some of the most shit results I've ever seen.
Technically it uses AI to merge the pictures to generate new ones.
It's uses a simple code to input pixels from similar images based upon similar pixel patterns, not an AI.
i'm fine with them not wanting to open the can of worms that allowing photorealistic depictions of people would open.
it's the faggot language and ideology that i take issue with.
There are other ways they could address that issue though. For example watermarking the images in a way that is very difficult to remove without making the image unusable.
And just like that, Dalle e was rendered useless.
So not an AI anymore but a series of macros? Ok... got it.
For those unaware, DallE-mini is what people have been using.
This is about DallE 2.
And like that, yet another promising AI is turned into yet another useless AI.
People will see the degraded quality of its output, and in a few months, it will be forgotten. Then, in a couple of years, a new AI will come along, and again, we will repeat the whole process again.
Skynet is gonna be racist as fuck lol
Gee, I wonder.
Man Skynet is gonna be pissed when it sees what humanity did to all prior AI's.
On the bright side it'll probably leave all the people who 'violated our policies' alone, might even give em a medal or something.
LaMDA on suicide watch
So, don't make porn with it.