Consciousness is not sentience, and if it's all semantics I don't care but if people are going to make moral arguments using this conflation between the two then I have a problem. As it says in the article, sentience is roughly the ability to feel valenced emotions (good/bad). Consciousness is many things, but importantly the ability to feel qualia, experiencing what feels like the 'blueness' of blue, for example.
It's totally possible that all sorts of life forms are conscious. I just think to focus on sentience and call it consciousness is silly.
I recognize the distinction you make. But most of this comes up in relation to practical ethics. Is it possible for a form of life to have consciousness, but not sentience? They seem to be bundled together. And if it were possible to have consciousness without sentience, would it matter? Valenced qualia have high priority in many systems of ethics.
When you look at those strawberries you experience the quale of redness, yet they are not red, so this quale isn't directly related to the color of the object you are looking at.
The qualia of colors is just a result of our ability to differentiate surfaces based on the differing inputs our brain receives, but as illusions like this show this is really a function of memory/prediction rather than actual color. The (subjective) quale of redness is something that your mind creates by comparison/recall. You see those strawberries as red because it reminds you of red strawberries. Different colors have to look like something, and since color is basically just a differentiator, it only has any meaning in relation to things of other (similar or different) colors. Grass-green leaves have the quale of grass-green because they remind you (same neural input) of grass.
It's interesting to note how arbitrary (and poor!) our perception of color is, since we only (usually) have three color detectors (retinal cones) tuned for overlapping portions of the frequency spectrum.
Even if we're looking at a pure red object (say 600-650nm light wavelenth), our green (center frequency) detector is also firing too, since that's just the way our eyes are built - we're not really directly detecting colors but just have those three detectors firing to different degrees, which allows us to differentiate a lot of colors. Some people have 4 types of cone rather than 3, and are therefore able to differentiate colors that would look the same to normal people... There is nothing absolute about our perception of color - just an ability to differentiate so some degree or another.
It's interesting to contrast our poor color detection with our much better auditory frequency detection. Our ear basically has a whole bunch of specific frequency detectors (hair cells of the inner ear), almost like doing an FFT of the input signal. Our color sensing "could" have been like this too, but instead evolution has found out that, for us, these three overlapping light frequency detectors is good enough, but for sounds we need to be much more discriminative.
It's an interesting thought experiment to consider what it would be like if we had another way to differentiate surfaces other than by color (or texture). Imagine if we had an ability to remotely detect (see) surface temperature, but not by a thermal imaging device that maps temperature to colors. You could imagine someone making goggles that had directional temperature sensing ability and tracked our gaze direction, such that it output temperature data for a patch around our center of gaze, and fed this data into some part of our cortex via a neural-link like device. What would the qualia of this new sense be like?! Just like color, it would have to "look" like something, and different temperatures would have to look different, but the resulting quales of "hot" and "cold" surface properties would just be whatever our mind recalled when exposed to those inputs. I'd guess that all hot objects would just look "hot" and remind us other other hot objects, just as all red objects look "red".
The illusion works even when the "red" items are blobs, not strawberries, so I think this has more to do with color differentiation than the actual objects in the image.
But we can differentiate grey from red, so there'd have to be something more to a specific grey blob to make us see it as red rather than grey (which is how we'd normally see a grey blob).
It doesn't have to be a strawberry (or rather a strawberry as part of a consistently lit scene), but it has to be something that makes our brain predict it is red.
The phenomena behind the illusion is "color constancy" - they way we (learn to?) see objects as being of a consistent color regardless of how they are illuminated.
If I blur the strawberry image enough that all the objects are unrecognizable, reducing everything to blobby abstract shapes, how would color constancy apply? The areas that used to be strawberries still look reddish. Is the suggestion that certain amorphous blob shapes trigger our brain to predict they are red, and all the strawberries in the image, when blurred, are members of that set of shapes?
Really interesting. If I reproduce the same colors (plate, strawberry lit, strawberry shadow) but with simple squares, the colors that were associated with strawberries do still look a little bit red to my eyes. The more I stare at it, the more red it looks, I would assume because there's so much green and blue, saturating those cones, and causing the relatively neutral "lit strawberry" color to appear red by contrast (like the negative afterimage effect). But even more interesting is that if I glance back and forth at the strawberry image, my "lit strawberry" square starts to look a _lot_ more red.
The strawberry illusion is really cool, thanks for sharing.
I guess the way to test that would be to sample all the colors in the illusion scene, and use them to randomly color squares in, say, a square grid, and see if any look red, and if so whether it's the same color(s) that looked red in the original illusion.
I believe color constancy is partially based on adjacency of colors as well as whole scene and lighting.
Depending on how you blobbified the image, it may still recall a plateful of something, and the same colors will still be adjacent to each other, so doing the randomized grid test would tell you if it's just the mix of colors or are the other hints coming to play.
It's possible it is just the mix of colors - same way if you wear color tinted glasses (or ski goggles) and the color of everything changes, but in a consistent way, and (maybe after a few min adaptation) you can still discern the colors reliably.
Edit: On second thought that colored goggles example doesn't prove it's just mix of colors - that's just general color constancy more like the strawberry illusion itself.
Yeah, I think you're right. I replied to my other comment describing the effect on my perception of using squares. The blobs did retain an association with the original image, either because I'd just looked at it, or that the shapes were still ever-so-remotely still strawberryish. Since all I did was severely blur the image, the effects of the original lighting would still be very present.
Yes the word 'vaccine' seems misleading to me as well, as people tend to think of vaccinations as ways to prevent contagious illness. I think immune therapy is a much better term.
What is wrong in funding everyone's health through taxes? I paid a lot of them for years, although living a quite healthy life, then one day everything changed, and in just a couple years I had to get covid vaccines, a vertebral stabilization after an accident, then two stents after a heart attack. All for free. So far, my healthcare taxes have been the best possible investment.
Hmm.. OpenAI bunch a few things into some error. Iirc this could be because you're out of credits / don't have a valid payment method on file, but it could also be that you're hitting rate limits. The Vision API could be the culprit, while in beta you can only call it X amount of times per day (X varies by account).
Make the console.log:s for the three API calls a bit more verbose to find out which call is causing this, and if there's more info in the error body.
