5 ms·

I think most people would find it more difficult to draw a “serkle nufding” than a “spotted dog”. At the very least, they would need an explanation that you’re

I think most people would find it more difficult to draw a “serkle nufding” than a “spotted dog”. At the very least, they would need an explanation that you’re looking for whatever pops into their head for the nonsense. Again, that indicates that human intelligence works in a fundamentally different way from DALLE, which has absolutely no idea what is nonsense. So, in this incredibly important task (recognizing nonsense) there’s an asymmetry: people can act like DALLE, but DALLE can not act like humans. To me, that totally breaks any equivalence between DALLE and human intelligence.

simondw 4y ago
> they would need an explanation that you’re looking for whatever pops into their head Of course. And the equivalent of that explanation is baked into DALL-E, in the form of its programming to always generate an image. > but DALLE can not act like humans No, not generally, but I don't think anyone has claimed that.
- bglazer 4y ago
  What else always produces an image, no matter what you present to it?
  - simondw 4y ago
    Lots of things, but I'm not sure why you're asking that. Could you connect that back to the point you're making?
    - bglazer 4y ago
      Sorry for being obtuse. I was referring to a mirror. The point being that people take objection to DALLE being characterized as a mirror but it behaves similarly in this regard
dtech 4y ago
Then again, if you lock a human in a room, tell it "draw me things based on the notes you receive or you will be killed", and give it a note "serkle nufding", you bet the human will draw something. Given the model has only 1 input and 1 output and training is essentially surviving that order, it's not dissimilar.
saynay 4y ago
I am not really sure that this is a basis to say human intelligence is fundamentally different. If you ask a human to draw a picture based on nonsensical input, they probably wont. That might not be a difference in process, but in agency. So is a single-purpose AI equivalent to the entirety of the Human Experience? Of course not. But can it be similar in functionality to a small sliver of it?
londons_explore 4y ago
I just got Dalle-2 to draw a "serkle nufding". It gave me 4 pictures of rather odd looking food. I too assume that random words are usually food, if only because restaurant menus seem to have such a love of making up foreign sounding words.
- dane-pgp 4y ago
  > a love of making up foreign sounding words One great example of this phenomenon is "Häagen-Dazs".[0] Admittedly that's a brand name, rather than a specific dish, but I assume that Dalle-2 would generate an image of ice cream if given a prompt with that term in it (unless there is a restriction on trademarks?). [0] https://funfactz.com/food-and-drink-facts/haagen-dazs-name/ https://funfactz.com/food-and-drink-facts/haagen-dazs-name/
caconym_ 4y ago
When I read the words "serkle nufding" I pictured in my head, with zero perceptible delay, a weird little goblin guy with pale green skin and long ears and a pointy hat.
- sarchertech 4y ago
  Serkle Nufding sounds like a word. What about “xywjlkbtywlsjjsnskksbrjsksbsysjsnsjdkdmdnd”?
  - caconym_ 4y ago
    Honestly? The top half of a signpost bearing the name of a Welsh village in white on green (likely an artifact of my actual local signage standards). The tops of some broadleaf deciduous trees, possibly ash, are visible in the background, but the perspective is from below (as one usually looks up at a street sign, also probably an artifact of my local experience with signs rather than a true representation of Welsh signage) and so the background is dominated by a blue sky. This is probably because your string is a low-entropy keyboard-mash.
  - astrange 4y ago
    That sounds like Welsh, but pointier. https://en.wikipedia.org/wiki/Bouba/kiki_effect https://en.wikipedia.org/wiki/Bouba/kiki_effect
gojomo 4y ago
I've not tested this, but I suspect you can get DALLE to create pictures that humans are more likely to describe as 'nonsense' by adding 'nonsense' or 'nonsensical' to the prompt. That'd indicate that it absolutely does have an idea about 'nonsense' and can recognize, & reproduce within its constrained output, 'nonsense' that's largely compatible with human ideas of what 'nonsense' is. Asking it to produce noise, or raise an objection that a prompt isn't sufficiently meaningful to render, is a silly standard because it's been designed, and trained, to always give some result. Humans who can object have been trained differently. Also, the GPT models – another similar train-by-example deep-neural architecture – can give far better answers, or give sensible evaluations of the quality of its answer, when properly prompted to do so. If you wanted a model that'd flag nonsense, just give it enough examples, and enough range-of-output where the answer your demanding is even possible, and it'll do it. Maybe better than people. The circumstances & limits of the single-medium (text, or captioned image) training goals, and allowable outputs, absolutely establish that these are different from a full-fledged human. A human has decades of reinforcement-training via multiple senses, and more output options, among other things. But to observe that difference and conclude these models don't "understand" the concepts they are so deftly remixing, or are "just a very complex and impressive mirror", does not follow from the mere difference. In their single-modalities, constrained as they may be, they can train the equivalent of a million lifetimes of reading, or image-rendering. Objectively, they're arguable now better at composing college-level essays, or rendering many kinds of art, than most random humans picked off the street would be. Maybe even better than 90% of all humans on earth at these narrow tasks. And, their rate of improvement seems only a matter of how much model-size & training-data they're given. Further: the narrowness of the tasks is by designers' choice, NOT inherent to the architectures. You could – and active projects are – training similar multi-modality networks. A mixed GPT/DALLE that renders essays with embedded supporting pictures/graphs isn't implausible.

simondw 4y ago

> they would need an explanation that you’re looking for whatever pops into their head Of course. And the equivalent of that explanation is baked into DALL-E, in the form of its programming to always generate an image. > but DALLE can not act like humans No, not generally, but I don't think anyone has claimed that.

bglazer 4y ago
What else always produces an image, no matter what you present to it?
- simondw 4y ago
  Lots of things, but I'm not sure why you're asking that. Could you connect that back to the point you're making?
  - bglazer 4y ago
    Sorry for being obtuse. I was referring to a mirror. The point being that people take objection to DALLE being characterized as a mirror but it behaves similarly in this regard

dtech 4y ago

Then again, if you lock a human in a room, tell it "draw me things based on the notes you receive or you will be killed", and give it a note "serkle nufding", you bet the human will draw something. Given the model has only 1 input and 1 output and training is essentially surviving that order, it's not dissimilar.

saynay 4y ago

I am not really sure that this is a basis to say human intelligence is fundamentally different. If you ask a human to draw a picture based on nonsensical input, they probably wont. That might not be a difference in process, but in agency. So is a single-purpose AI equivalent to the entirety of the Human Experience? Of course not. But can it be similar in functionality to a small sliver of it?

londons_explore 4y ago

I just got Dalle-2 to draw a "serkle nufding". It gave me 4 pictures of rather odd looking food. I too assume that random words are usually food, if only because restaurant menus seem to have such a love of making up foreign sounding words.

dane-pgp 4y ago
> a love of making up foreign sounding words One great example of this phenomenon is "Häagen-Dazs".[0] Admittedly that's a brand name, rather than a specific dish, but I assume that Dalle-2 would generate an image of ice cream if given a prompt with that term in it (unless there is a restriction on trademarks?). [0] https://funfactz.com/food-and-drink-facts/haagen-dazs-name/ https://funfactz.com/food-and-drink-facts/haagen-dazs-name/

caconym_ 4y ago

When I read the words "serkle nufding" I pictured in my head, with zero perceptible delay, a weird little goblin guy with pale green skin and long ears and a pointy hat.

sarchertech 4y ago
Serkle Nufding sounds like a word. What about “xywjlkbtywlsjjsnskksbrjsksbsysjsnsjdkdmdnd”?
- caconym_ 4y ago
  Honestly? The top half of a signpost bearing the name of a Welsh village in white on green (likely an artifact of my actual local signage standards). The tops of some broadleaf deciduous trees, possibly ash, are visible in the background, but the perspective is from below (as one usually looks up at a street sign, also probably an artifact of my local experience with signs rather than a true representation of Welsh signage) and so the background is dominated by a blue sky. This is probably because your string is a low-entropy keyboard-mash.
- astrange 4y ago
  That sounds like Welsh, but pointier. https://en.wikipedia.org/wiki/Bouba/kiki_effect https://en.wikipedia.org/wiki/Bouba/kiki_effect

gojomo 4y ago

I've not tested this, but I suspect you can get DALLE to create pictures that humans are more likely to describe as 'nonsense' by adding 'nonsense' or 'nonsensical' to the prompt. That'd indicate that it absolutely does have an idea about 'nonsense' and can recognize, & reproduce within its constrained output, 'nonsense' that's largely compatible with human ideas of what 'nonsense' is. Asking it to produce noise, or raise an objection that a prompt isn't sufficiently meaningful to render, is a silly standard because it's been designed, and trained, to always give some result. Humans who can object have been trained differently. Also, the GPT models – another similar train-by-example deep-neural architecture – can give far better answers, or give sensible evaluations of the quality of its answer, when properly prompted to do so. If you wanted a model that'd flag nonsense, just give it enough examples, and enough range-of-output where the answer your demanding is even possible, and it'll do it. Maybe better than people. The circumstances & limits of the single-medium (text, or captioned image) training goals, and allowable outputs, absolutely establish that these are different from a full-fledged human. A human has decades of reinforcement-training via multiple senses, and more output options, among other things. But to observe that difference and conclude these models don't "understand" the concepts they are so deftly remixing, or are "just a very complex and impressive mirror", does not follow from the mere difference. In their single-modalities, constrained as they may be, they can train the equivalent of a million lifetimes of reading, or image-rendering. Objectively, they're arguable now better at composing college-level essays, or rendering many kinds of art, than most random humans picked off the street would be. Maybe even better than 90% of all humans on earth at these narrow tasks. And, their rate of improvement seems only a matter of how much model-size & training-data they're given. Further: the narrowness of the tasks is by designers' choice, NOT inherent to the architectures. You could – and active projects are – training similar multi-modality networks. A mixed GPT/DALLE that renders essays with embedded supporting pictures/graphs isn't implausible.