Jake
@ngmi.co
techbro with philosophy degree working on global state machines that sync as fast as physics allows
100k BTC or USD? Galaxy facilitated a $9 billion dollar bitcoin sale in July, it’s easy to move six figures worth of btc around financial institutions
November 14, 2025 at 2:08 AM
100k BTC or USD? Galaxy facilitated a $9 billion dollar bitcoin sale in July, it’s easy to move six figures worth of btc around financial institutions
Unless every artist has never seen art before their first art, it’s as much a xerox as a model trained on others art. The retrieval and synthesis of a an artificial neuron mimics that of a meat based neuron. If I steal a million pictures to learn how to identify a hat, I’m doing what a model does.
November 14, 2025 at 1:14 AM
Unless every artist has never seen art before their first art, it’s as much a xerox as a model trained on others art. The retrieval and synthesis of a an artificial neuron mimics that of a meat based neuron. If I steal a million pictures to learn how to identify a hat, I’m doing what a model does.
I feel like artistic intent is the bare minimum of every prompt. People searching for something in the latent space until they find what they are after, like a photographer trying to get a picture of an eagle.
November 14, 2025 at 1:10 AM
I feel like artistic intent is the bare minimum of every prompt. People searching for something in the latent space until they find what they are after, like a photographer trying to get a picture of an eagle.
so do all these images live in like a file system somewhere? how does it know how to take bits from each one and combine them? how does it make the mashup visually coherent?
November 13, 2025 at 9:26 PM
so do all these images live in like a file system somewhere? how does it know how to take bits from each one and combine them? how does it make the mashup visually coherent?
I truly am a technology brother so I take no offense, and I have enjoyed the conversation, I think these discussions keep us sharp 🤝
November 13, 2025 at 6:41 PM
I truly am a technology brother so I take no offense, and I have enjoyed the conversation, I think these discussions keep us sharp 🤝
I think thats fair. If I take a picture and tell people it's a hyper realistic oil painting, I'm obviously a liar. If I use AI to generate an image, and tell people I made it with GIMP, I'm obviously a liar there as well.
November 13, 2025 at 6:33 PM
I think thats fair. If I take a picture and tell people it's a hyper realistic oil painting, I'm obviously a liar. If I use AI to generate an image, and tell people I made it with GIMP, I'm obviously a liar there as well.
I think if the 6th grader knew what was happening after they submit their prompt they would see it as a beautiful process
I would actually recommend any 6th grader using AI watch this and develop an intuition for the underlying mechanisms and make more art
www.youtube.com/watch?v=iv-5...
I would actually recommend any 6th grader using AI watch this and develop an intuition for the underlying mechanisms and make more art
www.youtube.com/watch?v=iv-5...
But how do AI images and videos actually work? | Guest video by Welch Labs
YouTube video by 3Blue1Brown
www.youtube.com
November 13, 2025 at 6:31 PM
I think if the 6th grader knew what was happening after they submit their prompt they would see it as a beautiful process
I would actually recommend any 6th grader using AI watch this and develop an intuition for the underlying mechanisms and make more art
www.youtube.com/watch?v=iv-5...
I would actually recommend any 6th grader using AI watch this and develop an intuition for the underlying mechanisms and make more art
www.youtube.com/watch?v=iv-5...
It would not be a Jackson Pollock painting anymore than if I splatter paint on a canvas. It would be _your_ work of art, not Pollock's.
One of my old Phi professors was actually a hardcore Benjamin scholar, wrote on him through the lens of turning politics into aesthetics.
One of my old Phi professors was actually a hardcore Benjamin scholar, wrote on him through the lens of turning politics into aesthetics.
November 13, 2025 at 6:26 PM
It would not be a Jackson Pollock painting anymore than if I splatter paint on a canvas. It would be _your_ work of art, not Pollock's.
One of my old Phi professors was actually a hardcore Benjamin scholar, wrote on him through the lens of turning politics into aesthetics.
One of my old Phi professors was actually a hardcore Benjamin scholar, wrote on him through the lens of turning politics into aesthetics.
before I became a techbro with a degree in philosophy i was an industrial design major with a minor in art history
theres many artists i deeply enjoy and lots of designers i took inspiration from
i often see this dichotomy that you can either like tech or art, i think you can do both
theres many artists i deeply enjoy and lots of designers i took inspiration from
i often see this dichotomy that you can either like tech or art, i think you can do both
November 13, 2025 at 6:16 PM
before I became a techbro with a degree in philosophy i was an industrial design major with a minor in art history
theres many artists i deeply enjoy and lots of designers i took inspiration from
i often see this dichotomy that you can either like tech or art, i think you can do both
theres many artists i deeply enjoy and lots of designers i took inspiration from
i often see this dichotomy that you can either like tech or art, i think you can do both
Yes
Jackson Pollock: his style is proto diffusion model, generate noise, derive meaning
Jules Chéret: his integration of text and image to create art in 1800s color lithograph
A.M. Cassandre: his abstraction of basic forms to create emphasis on geometric characters
Giacomo Balla: radical motion
etc
Jackson Pollock: his style is proto diffusion model, generate noise, derive meaning
Jules Chéret: his integration of text and image to create art in 1800s color lithograph
A.M. Cassandre: his abstraction of basic forms to create emphasis on geometric characters
Giacomo Balla: radical motion
etc
November 13, 2025 at 6:12 PM
Yes
Jackson Pollock: his style is proto diffusion model, generate noise, derive meaning
Jules Chéret: his integration of text and image to create art in 1800s color lithograph
A.M. Cassandre: his abstraction of basic forms to create emphasis on geometric characters
Giacomo Balla: radical motion
etc
Jackson Pollock: his style is proto diffusion model, generate noise, derive meaning
Jules Chéret: his integration of text and image to create art in 1800s color lithograph
A.M. Cassandre: his abstraction of basic forms to create emphasis on geometric characters
Giacomo Balla: radical motion
etc
like people are still very much involved when using models, just as all the prior are human+machine creations, so my question is when do we start to decide that human+machine combos are not art... (just as photographs as human+machine was considered not art in the 1800s)
November 13, 2025 at 5:57 PM
like people are still very much involved when using models, just as all the prior are human+machine creations, so my question is when do we start to decide that human+machine combos are not art... (just as photographs as human+machine was considered not art in the 1800s)
to make sure i understand your point, paint brushes, 3D animation software, adobe after effects, etc... these are all art enabling tools right?
and what you're suggesting is that transformers and diffusion models are not art enabling?
is this because the former doesn't exhibit predictive behavior?
and what you're suggesting is that transformers and diffusion models are not art enabling?
is this because the former doesn't exhibit predictive behavior?
November 13, 2025 at 5:55 PM
to make sure i understand your point, paint brushes, 3D animation software, adobe after effects, etc... these are all art enabling tools right?
and what you're suggesting is that transformers and diffusion models are not art enabling?
is this because the former doesn't exhibit predictive behavior?
and what you're suggesting is that transformers and diffusion models are not art enabling?
is this because the former doesn't exhibit predictive behavior?
what is the nature of the machine?
pretend i know how cameras work but not neural nets
I prompt "wide-angle, low-slung fisheye shot, golden-hour light flaring off a marble bank facade. In the foreground, a gritty skater balances on his board. 35 mm film grain, subtle vignette."
what happens next?
pretend i know how cameras work but not neural nets
I prompt "wide-angle, low-slung fisheye shot, golden-hour light flaring off a marble bank facade. In the foreground, a gritty skater balances on his board. 35 mm film grain, subtle vignette."
what happens next?
November 13, 2025 at 5:43 PM
what is the nature of the machine?
pretend i know how cameras work but not neural nets
I prompt "wide-angle, low-slung fisheye shot, golden-hour light flaring off a marble bank facade. In the foreground, a gritty skater balances on his board. 35 mm film grain, subtle vignette."
what happens next?
pretend i know how cameras work but not neural nets
I prompt "wide-angle, low-slung fisheye shot, golden-hour light flaring off a marble bank facade. In the foreground, a gritty skater balances on his board. 35 mm film grain, subtle vignette."
what happens next?
it's not binary, you can know how to build a model and know how to draw, you also don't need to work for anyone (e.g. I was an industrial design major, drawing and sketching every day years before I implemented a model)
Anyone can do this, there's no gatekeeping:
karpathy.ai/zero-to-hero...
Anyone can do this, there's no gatekeeping:
karpathy.ai/zero-to-hero...
November 13, 2025 at 5:38 PM
it's not binary, you can know how to build a model and know how to draw, you also don't need to work for anyone (e.g. I was an industrial design major, drawing and sketching every day years before I implemented a model)
Anyone can do this, there's no gatekeeping:
karpathy.ai/zero-to-hero...
Anyone can do this, there's no gatekeeping:
karpathy.ai/zero-to-hero...
I do agree that the substrate for computation is different between an AI and a human, however I disagree that that differences of that substrate (one meat, the other silicon) is the boundary of what makes something an art or not.
news.harvard.edu/gazette/stor...
news.harvard.edu/gazette/stor...
Artificial intelligence may not be artificial — Harvard Gazette
Researcher traces evolution of computation power of human brains, parallels to AI, argues key to increasing complexity is cooperation.
news.harvard.edu
November 13, 2025 at 3:39 PM
I do agree that the substrate for computation is different between an AI and a human, however I disagree that that differences of that substrate (one meat, the other silicon) is the boundary of what makes something an art or not.
news.harvard.edu/gazette/stor...
news.harvard.edu/gazette/stor...
If the definition of art being used here is the requirement of human biology producing a work with no mechanism behind it, then I agree. However that’s not how I’d define art, so I disagree.
November 13, 2025 at 3:30 PM
If the definition of art being used here is the requirement of human biology producing a work with no mechanism behind it, then I agree. However that’s not how I’d define art, so I disagree.
it’s not plagiarism if the retrieval and synthesis comes from wet meat neurons and is if it comes from dry silicon neurons? If I see 1000 pictures of hats and contextualize it in my brain’s substrate it’s free game, but it’s stealing if we contextualize “hat” a million times with a neural net?
November 13, 2025 at 10:20 AM
it’s not plagiarism if the retrieval and synthesis comes from wet meat neurons and is if it comes from dry silicon neurons? If I see 1000 pictures of hats and contextualize it in my brain’s substrate it’s free game, but it’s stealing if we contextualize “hat” a million times with a neural net?