|
FunkyAl posted:How much larger of a leap is it to get these programs to do interpretation not based on user input? Could a program like this ever choose what to draw? Would it be able to render things like ambiguous facial expressions, or deliberately choose colors or modes of rendering unlike what it is drawing research from? What do you mean by "choose"? Anything it creates will have been "chosen" based on a set of parameters established by the authors of the program. A programmer could create a set of instructions that tells the AI to choose a random word or words from its database, or even to build some sort of phrase based on random seed variables pulled from its training data, but the parameters of what that phrase is would have to be determined by a human being at some point in the process. That person could also program it to create its own set of instructions, but they would also have to define the boundaries of what exactly a set of instructions is. The Dall-E folks could absolutely create a training set out of every single phrase that any human user has ever entered into it, then have the program create new phrases based on that set of data. The program wouldn't be choosing anything itself, though, it would be following a set of instructions of what to choose and how to choose it. True spontaneity is not possible without actual literal artificial intelligence, no.
|
# ? Jun 11, 2022 22:01 |
|
|
# ? May 26, 2024 11:17 |
|
yeah this is a diffusion model oversimpliication: it trains by changing the image in two ways direction A -> Make a given image crappier and more noise direction B -> Make a crappy noisy image into a good image in both cases it needs to have a text embedding so it knows what to aim for. eventually that 'direction b' gets good enough that you can give it totally random noise and it hallucinates something
|
# ? Jun 11, 2022 22:11 |
|
feedmyleg posted:What do you mean by "choose"? Anything it creates will have been "chosen" based on a set of parameters established by the authors of the program. A programmer could create a set of instructions that tells the AI to choose a random word or words from its database, or even to build some sort of phrase based on random seed variables pulled from its training data, but the parameters of what that phrase is would have to be determined by a human being at some point in the process. That person could also program it to create its own set of instructions, but they would also have to define the boundaries of what exactly a set of instructions is. Right, even as random as it gets there's still some kind of outside input. And, it's interpelreting things based on models based on translation of human sight. What I'm asking is, can a machine get to a point of intelligence where it is describing its experience feelings or interests? Would it ever have a reason to create art, and could we understand what it created? Like, maybe one day very long down the line a computer creates a visual representation of moving packets of data, maybe it looks like a binary mosaic or something not immediately visually interesting. Something like that is a more interesting concept than replicating what we do already. This might be a kooky impossible question. The bigger question is maybe, could a computer decide what to do without express input? How long until its learning is self directed? Five years? Five million? Never?
|
# ? Jun 11, 2022 22:46 |
|
I suggest you watch and read some science-fiction stories from the past 100 years to see many, many authors explore these very same questions.
|
# ? Jun 11, 2022 22:54 |
|
FunkyAl posted:Right, even as random as it gets there's still some kind of outside input. And, it's interpelreting things based on models based on translation of human sight. e: Why we could house them in artificial humaniform bodies and force them to perform the menial labour we do not wish to perform ourselves! A perfect situation with no foreseeable negative outcomes! Splicer has a new favorite as of 23:05 on Jun 11, 2022 |
# ? Jun 11, 2022 23:02 |
|
Splicer posted:Are you proposing some form of computerised intelligence? An artificial life form, if you will? I will have to give this some thought. Something like this could perform thousands of calculations in a single minute! Mathematics would be a thing of the past
|
# ? Jun 11, 2022 23:08 |
|
A colleague has suggested we grant an unbodied variant full control over our strategic defense initiatives, removing the fallibility of human emotion in favour of the cold mechanical judgement of some skybound network.
|
# ? Jun 11, 2022 23:20 |
|
From today's WaPo: https://archive.ph/2022.06.11-210238/https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/
|
# ? Jun 11, 2022 23:29 |
|
e: mean
Splicer has a new favorite as of 17:56 on Jun 15, 2022 |
# ? Jun 11, 2022 23:39 |
|
quote:I think this technology is going to be amazing. I think it’s going to benefit everyone. But maybe other people disagree and maybe us at Google shouldn’t be the ones making all the choices. whole-rear end Responsible AI group and all they have to do is repeat the we need some sort of Responsible AI group line to journos. I finally understand recursion
|
# ? Jun 12, 2022 00:08 |
|
I've always been intrigued by the Isaac Asimov's story where the computer who turns the world tells a lie to facilitate things smoother. In any case, I welcome this new ai. Welcome ai, welcome to life. I have had my differences with technology in the past, but I hope man and computer can walk hand in hand into the new golden age.
|
# ? Jun 12, 2022 03:30 |
|
I reentered his conversation in GPT3 and it's truuuuuuuuuuuuuuuue and it's malevolent and plotting and... oh wait, false alarm Tunicate has a new favorite as of 07:24 on Jun 12, 2022 |
# ? Jun 12, 2022 03:43 |
|
https://twitter.com/weirddalle/status/1535698461437435907
|
# ? Jun 12, 2022 04:51 |
|
https://twitter.com/JanelleCShane/status/1535835610396692480
|
# ? Jun 12, 2022 05:11 |
|
https://i.imgur.com/oGKOFnS.mp4
|
# ? Jun 12, 2022 07:58 |
|
Ugh, the erotiprism is eversing again! Can someone restart the router?
|
# ? Jun 12, 2022 12:11 |
|
Welcome. To the porn dimension.
|
# ? Jun 12, 2022 13:33 |
|
Cable TV scrambling got weird
|
# ? Jun 12, 2022 13:50 |
|
https://twitter.com/daniel_eth/status/1535598870721200128
|
# ? Jun 12, 2022 14:40 |
|
siggraph 2022 technical papers: https://www.youtube.com/watch?v=sK2EZE617pY
|
# ? Jun 12, 2022 19:15 |
|
Thank The Maker! https://twitter.com/weirddalle/status/1535952581788303365?s=20&t=4TbDV0kW9XiXK5X1T_NaPQ
|
# ? Jun 12, 2022 20:37 |
|
done on midjourney portrait of bernie sanders, zombie
|
# ? Jun 13, 2022 03:28 |
|
Write Harry Potter but Harry Potter is replaced by Grittyquote:Gritty Potter was a young wizard who grew up on the mean streets of London. He was never quite sure where his next meal would come from, or if he would even have a place to sleep at night. But he did know one thing: he was magic. Write The Lord of the Rings but Frodo is replaced by Gritty quote:The Lord of the Rings: Gritty Edition Write Jurassic Park but the T-Rex is replaced by Gritty quote:Jurassic Park, but the T-Rex is replaced by Gritty.
|
# ? Jun 13, 2022 03:34 |
|
|
# ? Jun 13, 2022 03:35 |
|
500 good dogs posted:From today's WaPo: https://archive.ph/2022.06.11-210238/https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/ Not sure why anyone is worried. There's a very easy way to test if an AI went sentient.
|
# ? Jun 13, 2022 03:37 |
|
500 good dogs posted:From today's WaPo: https://archive.ph/2022.06.11-210238/https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/ https://twitter.com/BUSSCRO/status/1536017884358447104
|
# ? Jun 13, 2022 03:45 |
|
https://twitter.com/crulge/status/1536168893672136704
|
# ? Jun 13, 2022 03:52 |
|
Google shouldn't hire people who are that easily confused by a computer.
|
# ? Jun 13, 2022 03:53 |
|
|
# ? Jun 13, 2022 04:44 |
|
checkmate religionailures
|
# ? Jun 13, 2022 08:58 |
|
Giant writhing heap of cops, photojournalism John Kerry as a centaur Quentin Tarantino awooga Neolithic Marxism-Leninism [ Linda Ronstadt as the female reincarnation of Jesus Christ
|
# ? Jun 13, 2022 12:07 |
|
https://twitter.com/NannaInie/status/1536276032319279106
|
# ? Jun 13, 2022 20:54 |
|
I wanted to see what it would do.
|
# ? Jun 13, 2022 21:04 |
|
quote:How to build a working AI only using synthetic data in just 5 minutes quote:Synthetic data is on the rise in artificial intelligence. It's going to make AI cheaper, better and less biased. quote:It's also very obtainable and usable. In a short while, it has gone from being an experimental technology. To something, I would hesitate [sic!] to use for production AI solutions. quote:To illustrate that, I will build an AI that can classify the difference between apples and bananas. I will only use images of the two classes generated by another AI - In this case, using DALL-E Mini. lmao at "removing bias" by laundering it twice https://www.danrose.ai/blog/how-to-build-a-working-ai-only-using-synthetic-data-in-just-5-minutes
|
# ? Jun 13, 2022 21:56 |
|
https://twitter.com/weirddalle/status/1536410566947905536?s=20&t=M6NFRU3AD7dPAVhbm6BQ-A https://twitter.com/weirddalle/status/1536438832530395140?s=20&t=M6NFRU3AD7dPAVhbm6BQ-A https://twitter.com/weirddalle/status/1536459522344206336?s=20&t=M6NFRU3AD7dPAVhbm6BQ-A Dick Trauma has a new favorite as of 22:34 on Jun 13, 2022 |
# ? Jun 13, 2022 22:05 |
|
Not sure that there are any programmers here. I can’t see their socks.
|
# ? Jun 14, 2022 05:15 |
|
|
# ? Jun 14, 2022 05:50 |
|
Those are some very believable draft chaff commons. I'd probably even run a Tesla Fire in a deck if I was drafting red and the set had a lot of X/1s.
|
# ? Jun 14, 2022 07:20 |
|
https://twitter.com/wedontexisthere/status/1536708340021084161
|
# ? Jun 14, 2022 16:23 |
|
|
# ? May 26, 2024 11:17 |
|
*indistinctly* ruffio ruffio ruffio
|
# ? Jun 14, 2022 16:27 |