9 ms·

“ It actually knows what a giraffe is.” No. You know what a giraffe is, Dall•E simply creates pixel groups which correlate to the text pattern you submitted.

“ It actually knows what a giraffe is.” No. You know what a giraffe is, Dall•E simply creates pixel groups which correlate to the text pattern you submitted. Watching people discuss a logical mirror scares me that most people are not themselves conscious.

gfodor 4y ago
Not exactly. If you're using the word "simply" you probably don't quite understand what exactly DALL-E is doing. It's not just mapping correlations between words to pixels. The intermediate representations and textual attention are what make DALL-E interesting.
jvanderbot 4y ago
Precisely this! You don't know what a giraffe is, either. But you can talk about a giraffe to yourself, using words with correlating imagery that you can "see" by feeding through your visual cortex, and that's close enough.
- astrange 4y ago
  You know more of what a giraffe is than DALLE; you can imagine one as a 3D model, for instance. It can't do this, it can only render different 2D images of one it's seen in photos. Of course, this is still impressive that it knows how to remove the backgrounds from those. Examples of prompts that don't really work are "spherical giraffe" and "the underside of a giraffe". We'll need more 3Dness to generate other media like models and video, so it should be coming.
  - jvanderbot 4y ago
    I agree. One thing I've noticed with most image generation, including faces, is that it makes the same mistakes that a novice artist makes. It draws eyes where eyes should be, mouths where mouths should be, but doesn't actually shape them correctly for the 3D head and viewer perspective. It can perfect a specific viewpoint with lots of training (like a novice artist who perfects drawing realistic faces and not much else). But my interpretation of much of the artifacts are that they are just missing structural cues.
snapcaster 4y ago
How do you know you know what a giraffe is?
- twoxproblematic 4y ago
- logifail 4y ago
  > How do you know you know what a giraffe is? Perhaps a more interesting question could be: [how] do we know what consciousness is?
adamsmith143 4y ago
Comprehension does not imply consciousness or vice versa...
gojomo 4y ago
There's a lot of people who make this same argument – DALLE/GPT/etc is just a 'mirror' or 'parrot' – but they rarely make convincing supporting arguments. They just assert it as axiomatic, whistling-past all the ways that they themselves – unless they believe in supernatural mechanisms – are also the product of a finite physical-world system (a biological mind) and a finite amount of prior training input (their life so far). I'm beginning to wonder if the entities making this argument are conscious! It seems they don't truly understand the issues in question, in a way they could articulate recognizably to others. They're just repeating comforting articles-of-faith that others have programmed into them.
- mlcrypto 4y ago
  Yep human brain is just neurons firing or chemicals activating
- martindbp 4y ago
  Yes, and DALLE clearly knows what a giraffe is, maybe not in the same way as we do, but it has placed the giraffe in some latent space close to other animals like it but far away from say, cars. But yes, there is plenty it doesn't know about giraffes because it's only been fed images and captions, while we can watch documentaries about giraffes and apply knowledge we've acquired about other animals and the world through our senses.
  - drdeca 4y ago
    I think I’d say that it knows what a giraffe looks like and some aspects of their behavior. I’m not sure if it (in a sense) “knows that a giraffe is an animal” though I wouldn’t be too surprised if it does. Though, I guess maybe that is all “knowing what a giraffe is” is?
    - saynay 4y ago
      At some point, these discussions always get to some question like "what is 'knowing'?" Would a person who only ever read articles and looked at pictures of giraffes have a better understanding of them than Dall-e does? At some level, probably, in that every person will have a similar lived experience of _being_ an animal, a mammal, etc. that Dall-e will never share. Is having a lesser understanding sufficient to declare it has no real understanding?
EricBurnett 4y ago
Thought experiment: between you and the AI, which would do a better job depicting a giraffe skeleton? A giraffe in it's natural habitat? Their favorite tree to eat? Species on the genetic tree closest to giraffes? If we assume this AI or a successor can win that evaluation, in what way would you say you know what a giraffe is better than the AI?
chaxor 4y ago
There are far too many that try to use this argument without really knowing what they're talking about. *Don't mistake skepticism for knowledge* This is a major problem on this site and elsewhere.
elcomet 4y ago
> You know what a giraffe is, Dall•E simply creates pixel groups which correlate to the text pattern you submitted. How would you tell the difference though? Can you think of a test to distinguish between those two abilities ?
rhdunn 4y ago
There's an interesting Google TechTalk on YouTube called "The Next Generation of Neural Networks" from 2007 [1]. In that video, there is a demo [2] that showcases the neural network recognizing numbers when given a drawing of a number as inputs, but more interesting is the follow on where the demo shows fixing the output to a given number and letting the neural network generate the "input" image to show what it is thinking that number can be. That is a strong indication to me that that particular neural network has a good understanding of what the particular number glyphs look like, even if it does not know what those are conceptually or how they relate to each other and to mathematics -- that is, that the neural network would not be able to work out what number 42 is, how it relates to 37 (i.e. 37 < 42), or how to manipulate those numbers (e.g. 2 + 7 = 9). Dall-E will likely be similar in that it is effectively doing that perception step where you fix the text description from the classifier output and run that in reverse to show what the neural network is "seeing" when it is "thinking" about that given output. So it won't be able to describe features of a giraffe, or information about where they live, etc. but it will be able to show you what it thinks they look like. [1] https://www.youtube.com/watch?v=AyzOUbkUf3M https://www.youtube.com/watch?v=AyzOUbkUf3M [2] https://youtu.be/AyzOUbkUf3M?t=1293 https://youtu.be/AyzOUbkUf3M?t=1293

gfodor 4y ago

Not exactly. If you're using the word "simply" you probably don't quite understand what exactly DALL-E is doing. It's not just mapping correlations between words to pixels. The intermediate representations and textual attention are what make DALL-E interesting.

jvanderbot 4y ago

Precisely this! You don't know what a giraffe is, either. But you can talk about a giraffe to yourself, using words with correlating imagery that you can "see" by feeding through your visual cortex, and that's close enough.

astrange 4y ago
You know more of what a giraffe is than DALLE; you can imagine one as a 3D model, for instance. It can't do this, it can only render different 2D images of one it's seen in photos. Of course, this is still impressive that it knows how to remove the backgrounds from those. Examples of prompts that don't really work are "spherical giraffe" and "the underside of a giraffe". We'll need more 3Dness to generate other media like models and video, so it should be coming.
- jvanderbot 4y ago
  I agree. One thing I've noticed with most image generation, including faces, is that it makes the same mistakes that a novice artist makes. It draws eyes where eyes should be, mouths where mouths should be, but doesn't actually shape them correctly for the 3D head and viewer perspective. It can perfect a specific viewpoint with lots of training (like a novice artist who perfects drawing realistic faces and not much else). But my interpretation of much of the artifacts are that they are just missing structural cues.

snapcaster 4y ago

How do you know you know what a giraffe is?

twoxproblematic 4y ago
logifail 4y ago
> How do you know you know what a giraffe is? Perhaps a more interesting question could be: [how] do we know what consciousness is?

adamsmith143 4y ago

Comprehension does not imply consciousness or vice versa...

gojomo 4y ago

There's a lot of people who make this same argument – DALLE/GPT/etc is just a 'mirror' or 'parrot' – but they rarely make convincing supporting arguments. They just assert it as axiomatic, whistling-past all the ways that they themselves – unless they believe in supernatural mechanisms – are also the product of a finite physical-world system (a biological mind) and a finite amount of prior training input (their life so far). I'm beginning to wonder if the entities making this argument are conscious! It seems they don't truly understand the issues in question, in a way they could articulate recognizably to others. They're just repeating comforting articles-of-faith that others have programmed into them.

mlcrypto 4y ago
Yep human brain is just neurons firing or chemicals activating
martindbp 4y ago
Yes, and DALLE clearly knows what a giraffe is, maybe not in the same way as we do, but it has placed the giraffe in some latent space close to other animals like it but far away from say, cars. But yes, there is plenty it doesn't know about giraffes because it's only been fed images and captions, while we can watch documentaries about giraffes and apply knowledge we've acquired about other animals and the world through our senses.
- drdeca 4y ago
  I think I’d say that it knows what a giraffe looks like and some aspects of their behavior. I’m not sure if it (in a sense) “knows that a giraffe is an animal” though I wouldn’t be too surprised if it does. Though, I guess maybe that is all “knowing what a giraffe is” is?
  - saynay 4y ago
    At some point, these discussions always get to some question like "what is 'knowing'?" Would a person who only ever read articles and looked at pictures of giraffes have a better understanding of them than Dall-e does? At some level, probably, in that every person will have a similar lived experience of _being_ an animal, a mammal, etc. that Dall-e will never share. Is having a lesser understanding sufficient to declare it has no real understanding?

EricBurnett 4y ago

Thought experiment: between you and the AI, which would do a better job depicting a giraffe skeleton? A giraffe in it's natural habitat? Their favorite tree to eat? Species on the genetic tree closest to giraffes? If we assume this AI or a successor can win that evaluation, in what way would you say you know what a giraffe is better than the AI?

chaxor 4y ago

There are far too many that try to use this argument without really knowing what they're talking about. *Don't mistake skepticism for knowledge* This is a major problem on this site and elsewhere.

elcomet 4y ago

> You know what a giraffe is, Dall•E simply creates pixel groups which correlate to the text pattern you submitted. How would you tell the difference though? Can you think of a test to distinguish between those two abilities ?

rhdunn 4y ago

There's an interesting Google TechTalk on YouTube called "The Next Generation of Neural Networks" from 2007 [1]. In that video, there is a demo [2] that showcases the neural network recognizing numbers when given a drawing of a number as inputs, but more interesting is the follow on where the demo shows fixing the output to a given number and letting the neural network generate the "input" image to show what it is thinking that number can be. That is a strong indication to me that that particular neural network has a good understanding of what the particular number glyphs look like, even if it does not know what those are conceptually or how they relate to each other and to mathematics -- that is, that the neural network would not be able to work out what number 42 is, how it relates to 37 (i.e. 37 < 42), or how to manipulate those numbers (e.g. 2 + 7 = 9). Dall-E will likely be similar in that it is effectively doing that perception step where you fix the text description from the classifier output and run that in reverse to show what the neural network is "seeing" when it is "thinking" about that given output. So it won't be able to describe features of a giraffe, or information about where they live, etc. but it will be able to show you what it thinks they look like. [1] https://www.youtube.com/watch?v=AyzOUbkUf3M https://www.youtube.com/watch?v=AyzOUbkUf3M [2] https://youtu.be/AyzOUbkUf3M?t=1293 https://youtu.be/AyzOUbkUf3M?t=1293