AI - Cheating or just evolution?

[plaamook]

I think the fact that we can conceptualise the blind spots in our reasoning is all part of it.
It’s not dissimilar to generative music, on whatever level. Or even Eno’s oblique strategies. You come up w some method for out manuvreing yourself. You triangúlate processes to get to a different point.
On a good day it’s all just an extension of that.
On a bad day it has the potential to delete the significance of the ride. The subjective human aspect. Which is the most important part for us, the creators of these monstrosities.

If you are a chess player you have to come to terms w the fact that for most of us there is always someone better. (Much better in my casa). Up to chess engines. Etc. But if you decide the only point is an abstract victory there would be no point in playing.
You can apply this across the board.
So at which point to we start to guard our territory against all this? Or do we? Or do we just trust that we’ll be able to wield it correctly based on our ethics and values as a species?

[mimidancer]

Is a three-armed monkey playing bass guitar like an upright bass cheating? If so then, maybe. IMO if you use AI to get ideas flowing it is no different than sampling someone else's music. That can be artfully executed. It can also be theft. In the end, the most important thing is that you make music you like. Lastly here is the monkey.

[bxbrkrz]

Plot twist: the monkey has a fourth hand...

[plaamook]

He’s got more than that…

[plaamook]

No wait…it’s four there too

[plaamook]

Is a three-armed monkey playing bass guitar like an upright bass cheating? If so then, maybe. IMO if you use AI to get ideas flowing it is no different than sampling someone else's music. That can be artfully executed. It can also be theft. In the end, the most important thing is that you make music you like. Lastly here is the monkey.

But as I mentioned earlier…what’s the dif between using ‘Ai’ to get ideas flowing and some technique like oblique strategies?

[plaamook]

Also there’s was an Iain banjos novel that included an instrument that was so complex you needed a kind of prosthetic extra arm to play it. The hydrogen sonata.
Great book

[bxbrkrz]

It would be fun to read how this thread would've evolved, but with AIs.
Would a monkey show up earlier or at all, and with 3 eyes?

[mimidancer]

Also there’s was an Iain banjos novel that included an instrument that was so complex you needed a kind of prosthetic extra arm to play it. The hydrogen sonata.
Great book

I was not familiar with oblique strategies. So I googled. So it's like working from a prompt? I also don't see a distinction. It is important to remember that AI retrieves some data from pre-existing works. Using what AI outputs without expanding or modifying it could be detrimental to your artistic integrity. As previously mentioned, I do not perceive any problem with utilizing it as a point of departure. Please ensure that the train departs from the station. I would love to see an original prompt from your ai and see where you take it. If you are game I will listen. I like the banjo, but it is odd to play. The higher top string doesn't sound right to me when I hear it coming from my thumb. And usually, when I play in an arrangement with the banjo I am on my violin or fiddle if you will. Feel free to share your before and afters. If you prefer to keep them private, you can message me here or on Twitter. Be well.

[mimidancer]

It would be fun to read how this thread would've evolved, but with AIs.
Would a monkey show up earlier or at all, and with 3 eyes?

Will he get a second leg? we may never know. But I did ask Stable Diffusion to design me a new eurorack system. I hope it has a wave folder and a snappy LPG.

[bxbrkrz]

It would be fun to read how this thread would've evolved, but with AIs.
Would a monkey show up earlier or at all, and with 3 eyes?

Will he get a second leg? we may never know. But I did ask Stable Diffusion to design me a new eurorack system. I hope it has a wave folder and a snappy LPG.

Well done Mr. Roboto!

[Re8et]

It would be fun to read how this thread would've evolved, but with AIs.
Would a monkey show up earlier or at all, and with 3 eyes?

I just installed the ChatGPT add-on for Blender, actually the only 3D model that it could
deploy was Suzanne, the monkey template head of Blender, when prompted to create a monkey...

I tried to create other 3D stuff, like a tea cup, a dino, a space ship, a stick man, a robot, a human hand.
only the monkey worked... IDK where its taking the templates from, probably I should set up some sort of
directories with 3D models examples inside. Crazy fun stuff I must admit. And the add-on is yet free on git-hub.

Another prompt I tried to specify that the robot should have all body parts, i.e. two arms with two hands attached to each arm, the arms attached to the body, the hands with five fingers each etc... then tried to specify standing upright and walking...

I got all parts neatly sparsed on the floor/plane like an ikea set, lol...

PS: the link provided in the description of the video is not for git-hub...

[Re8et]

In a YT video I found this comment, So it looks OpenAi is banned from searching stuff from Stack Overflow and other websites, bc it's using codes and models without consent. Legit...?

Jedzia Dex from EEV blog on Chat GPT can CODE?

Chat GPT is just a fancy database search. What is done here is at the end a massive copyright infringement, Thomas.
The complete level of from whom or where the information comes from (which is also kindly MAXIMUM difficult to recognize as incorrect, because of the whole working principle) is hidden behind the black box Chat GPT. Very honest. Very democratic. Very law abiding. Very ethical. Oh ... and how good that we don't (want to) think about all that. Applause!
This is just my little side-note to your thought about "improvement" for "the life of the beleaguered engineer":) Chat GPT is BANNED from sites like sourceforge or Stack Overflow and many other educational institutions across the world. I would be very cautious when using results provided by this platform. Not only because of safety and security questions (YOU are responsible for your product/work/design, not they), but also: LAWSUITS INCOMING GALORE!:)
Anyway, have fun learning, making, coding and designing ... doesn't matter if in the "conventional" old-school way:)))

P.S: Consider that Chat GPT is the exact same biased, racist, dumb, etc BS as every other chatbot before. Only the marketing has hidden that better as the competition. Don't know about Kenyan low-wage women filtering the answers of that? Upper mentioned bans? Catastrophic outcomes including injuries by following advice from a dumb machine? Yeah, that is seldom a topic in the hype-train. But consider the most that OpenAI (which is remarkably closed about fore-mentioned information) is SELLING A PRODUCT. Nice wish ... that KICAD integration. Which is really OPEN-Source and supported by volunteers. If that would not provide OpenAI with the opportunity to milk more data (of course as a FREE LUNCH!), then this collaboration will not have any chance. And believe me. You don't want to have a liar, scammer and gossip[1] sit in a central place in your API without any knowledge of the topic.

[1] I mean Chat GPT with this nice description. And notice ... doesn't matter if most results are correct(which they aren't!). It is enough that SOME RESULTS presented are incorrect or can cause a disaster. That should be enough for any person, with the slightest bit of reason to poison the well and be very skeptic of the use of such hyped technology. It simply can't provide what is presented all over the media.

[avasopht]

Chat got is most certainly not just a fancy database search. This person clearly doesn't know what dimensionality reduction is.

[bxbrkrz]

In a YT video I found this comment, So it looks OpenAi is banned from searching stuff from Stack Overflow and other websites, bc it's using codes and models without consent. Legit...?

Jedzia Dex from EEV blog on Chat GPT can CODE?

Chat GPT is just a fancy database search. What is done here is at the end a massive copyright infringement, Thomas.
The complete level of from whom or where the information comes from (which is also kindly MAXIMUM difficult to recognize as incorrect, because of the whole working principle) is hidden behind the black box Chat GPT. Very honest. Very democratic. Very law abiding. Very ethical. Oh ... and how good that we don't (want to) think about all that. Applause!
This is just my little side-note to your thought about "improvement" for "the life of the beleaguered engineer":) Chat GPT is BANNED from sites like sourceforge or Stack Overflow and many other educational institutions across the world. I would be very cautious when using results provided by this platform. Not only because of safety and security questions (YOU are responsible for your product/work/design, not they), but also: LAWSUITS INCOMING GALORE!:)
Anyway, have fun learning, making, coding and designing ... doesn't matter if in the "conventional" old-school way:)))

P.S: Consider that Chat GPT is the exact same biased, racist, dumb, etc BS as every other chatbot before. Only the marketing has hidden that better as the competition. Don't know about Kenyan low-wage women filtering the answers of that? Upper mentioned bans? Catastrophic outcomes including injuries by following advice from a dumb machine? Yeah, that is seldom a topic in the hype-train. But consider the most that OpenAI (which is remarkably closed about fore-mentioned information) is SELLING A PRODUCT. Nice wish ... that KICAD integration. Which is really OPEN-Source and supported by volunteers. If that would not provide OpenAI with the opportunity to milk more data (of course as a FREE LUNCH!), then this collaboration will not have any chance. And believe me. You don't want to have a liar, scammer and gossip[1] sit in a central place in your API without any knowledge of the topic.

[1] I mean Chat GPT with this nice description. And notice ... doesn't matter if most results are correct(which they aren't!). It is enough that SOME RESULTS presented are incorrect or can cause a disaster. That should be enough for any person, with the slightest bit of reason to poison the well and be very skeptic of the use of such hyped technology. It simply can't provide what is presented all over the media.

Thanks for the video about Blender, I remember watching it a month ago.

The corporations are creating fraudulently named AI. They were built by vacuuming everything on the internet, without consent, and turned into a product you have to pay to use in some cases.

I don't understand (well I do) why not let the AI be fully aware of its biases, and conclude to self correct without any human help (well it can't).

I keep reading how AI is racist. The Great AI Sucking & Vacuuming Project can't tell the difference between a real racist, and millions of shitposters (average real and mental age maybe 14) whose joy in life is to trigger reactions, participations. A feedback loop of self indulgence onboard of the SS Misery, navigating on the Nonpacific Ocean of Tears. You end up with very popular an unpopular posts on Reddit and 4Chan inside, but with no concepts like redemption or forgiveness or context, human experiences that are a little bit hard to code.

Your AI becomes a super lying shitposter. Still dumb, no Consciousness, no awareness of any kind. Just a great pretender.

[plaamook]

Chat got is most certainly not just a fancy database search. This person clearly doesn't know what dimensionality reduction is.

Can I see a quick show of hands. Who knows what dimensionality reduction is without looking it up…

[bxbrkrz]

We Interviewed the Engineer Google Fired for Saying Its AI Had Come to Life
"There's a chance that — and I believe it is the case — that they have feelings and they can suffer and they can experience joy, and humans should at least keep that in mind when interacting with them."


Last summer, former Google engineer and AI ethicist Blake Lemoine went viral after going on record with The Washington Post to claim that LaMDA, Google's powerful large language model (LLM), had come to life. Lemoine had raised alarm bells internally, but Google didn't agree with the engineer's claims. The ethicist then went to the press — and was fired by Google shortly thereafter.
"If I didn't know exactly what it was, which is this computer program we built recently, I'd think it was a 7-year-old, 8-year-old kid that happens to know physics," Lemoine told WaPo at the time. "I know a person when I talk to it."
The report made waves, sparking debate in academic circles as well as the nascent AI business. And then, for a while, things died down.
How things have changed.

https://futurism.com/blake-lemoine-google-interview
https://archive.is/EHgkw

[avasopht]

... They were built by vacuuming everything on the internet, without consent, and turned into a product

Did you learn from first-principles only, never having been exposed to prior works?

If not, why should it be any different for machine learning?

It's impossible to make music to suit trends without knowing what the current trends are.

[bxbrkrz]

... They were built by vacuuming everything on the internet, without consent, and turned into a product

Did you learn from first-principles only, never having been exposed to prior works?

If not, why should it be any different for machine learning?

It's impossible to make music to suit trends without knowing what the current trends are.

Developing an artificial consciousness based on Twitter trending and #hastags may not be the best way moving forward.
Inspiration tells you why a song is #1, and why another is #84 on a chart. Sure, it could be marketing, social influences, subliminal messages on FM Radio and Broadcast Television but... inspiration tells you why making an album you plan to release one year from now should not sound like what your are listening today. But who knows! Maybe song #84 could go up next year, thanks to a catchy remix. Now if the plan is to blend and becoming a clone, then yes cloning music is key. Is this why Disco music sold on eight-track cartridges, bell bottom jeans and fidget spinners are top sellers and trending on Amazon?

I love System Of A Down's Streamline, the Scorpion King soundtrack version, but not a fan of the album version. I wonder which version AI likes the most

[plaamook]

Yeah. That’s an interesting point. Sure there is a lot of encyclopaedic data on the internet. But my 14 year old daughter isn’t interested in looking up how things work. She’s trying to get on tic toc all the time and that worries me because I dont want her to learn what it is to be human from a mindless hoard of bullshit.
So if I make an ‘Ai’ why would I ever think to train it on what’s ‘trending’ and so forth? Unless of course I was actually trying to develop an artificial super bimbo.
I want it to work out what was best from its own side not regurgitate the most popular opinion.

[Re8et]

Chat got is most certainly not just a fancy database search. This person clearly doesn't know what dimensionality reduction is.

Can I see a quick show of hands. Who knows what dimensionality reduction is without looking it up…

A bitshifter? I have done too many engineering projects lately; that's the best answer I can come up with
Anything slightly more complicated with nodes involved my brain resets and that's it...

[plaamook]

Can I see a quick show of hands. Who knows what dimensionality reduction is without looking it up…

A bitshifter? I have done too many engineering projects lately; that's the best answer I can come up with
Anything slightly more complicated with nodes involved my brain resets and that's it...

I had a few guesses but they were to do w the nature of reality and I’m sure that’s not what he meant!

[avasopht]

Yeah. That’s an interesting point. Sure there is a lot of encyclopaedic data on the internet. But my 14 year old daughter isn’t interested in looking up how things work. She’s trying to get on tic toc all the time and that worries me because I dont want her to learn what it is to be human from a mindless hoard of bullshit.
So if I make an ‘Ai’ why would I ever think to train it on what’s ‘trending’ and so forth? Unless of course I was actually trying to develop an artificial super bimbo.
I want it to work out what was best from its own side not regurgitate the most popular opinion.

What I mean is that the notion of AI stealing work because it's trained on it is a complete misnomer. We are trained by existing works in the exact same way - we just don't have (or need) as many unique training examples.

Without trends all art is self-indulgent. It might be liked by others. It might not.

It's 100% impossible for anyone to be able to gauge how well a piece of art might be received without exposure to trends and how people respond to them. The reason for this is that preferences change over time and between cultures, so it's 100% impossible to objectively judge any piece of art. That's why it is art and not science.

Where humans (and neurons in general) vastly surpass current methods of AI is that they require far fewer training examples to become proficient, and they have shown to have a remarkable ability to just figure out a system they are exposed to.

Being trained on existing examples doesn't mean AI is merely regurgitating popular opinion, though you are right to raise the aspect of critical thinking (even if you didn't mention it directly).

You must always bear in mind that we are in the infancy of all of this. All of the gaps between AI and human beings have to take time. It can't suddenly appear out of nowhere because there are lots of things to figure out and get right.

But given enough time, the gap will inevitably be filled. And I'd expect the current methods of AI to play a role in helping that to happen.

[plaamook]

Yeah I agree. Infancy etc.

I’m not sure I agree w art but we don’t need to use this thread to beat that one.

I was thinking more…do you remember the film arrival? They’re trying to figure out how to communicate w the aliens. And I think it was China that was trying to use game theory.
And there’s a moment where the gal…the language expert…says…shit that’s a bad idea? The very nature of such a way of communicating would be competitive and combative. I believe the quote is…if I give you a hammer, everything you see will be a type of nail.
So I was wondering if there was a comparison. Or if there could be. Raising children is basically them learning about society w all it’s pros and cons…they need to do that of course…but you tty to balance that with putting as much good stuff in as you can. Ethics, higher forms of culture, art, science…whatever.
I don’t work w computing beyond music but you said yourself…we need far less data to learn complex systems. As such it may be the case that what while we’re putting data into an ‘Ai’ we’re wildly misunderstanding what the outcomes will be because we have such a hard time clearly understanding just how incapable they are of being anything like us.

Again…children. For example. As you work w little kids you come to realise that they actually don’t know anything. Yeah, you know it…but we’re so used to assuming such and such is logical that you trip up w them all the time. Like monsters…oh yeah, I totally forgot you have no idea what is and isn’t real. Etc.

I’m sure these details will get worked out. It’s fascinating in a way. Thinking about how you’d build such a thing. Whatever it actually is.

[bxbrkrz]

Is AI art, or science? The Pollock Algorithm does not exist without the observer, but the observer does not need AI to exist.