The same phenomenon also applies to color blending. In a modern context this phenomenon is especially visible in UI elements that blur the background: Up until recently many UIs did not get the blurring right, resulting in greyish dark spots between different colors. I think CSS blurring in most browsers still gets it wrong.
Cuteness makes us want to hug, tease and whirl around the little being (basically an adaptation for the offspring to be encouraged to test social boundaries, individual limitations, and explore how the world works), but if it's too small or fragile for that or just a picture on a screen, then one gets a little angry for not being allowed to do so (like seeing delicious food and not being allowed to eat it). At the same time one is primed for playful action (biting, pinching, squeezing), so anger already is nearby in the space of emotions. This might also be aggravated between different species as the reactions might not match the expectations.
> Yes but consciousness is not in any way a foundation of the universe.
The universe is fractally structured regarding simplicity: It is simple at its foundation upon which we find layers of evidently chaotic processes (e.g. brownian motion, thermodynamics, fluid dynamics), which in turn converge to metastable rule sets that are seemingly simple again (e.g. evolution, neural networks). Unpredictable fluctuation from the chaotic layers below are either exploited as a computational mechanism (e.g. for probabilistic modeling or adaptations of unpredictability in behavior) or averaged out by regulation (homeostasis), so simple rules remain plausible despite the underlying chaos. Those simple rule sets in turn produce complex processes (e.g. psychology, science), which in turn produce simple processes (e.g. game theory, economics). Of course the higher up in this hierarchy, the more unstable rule sets become, e.g. most economic theories make poor predictions, but evolution is an extremely reliable theory.
> Simple answer to complex problems, outside science and engineering, are almost always wrong.
True, but it seems that the universe is simple at its foundation and complex in some of the things that emerge from it.
The standard model fits on a sheet of paper and it commonly makes predictions that are accurate to the ~20th decimal place. Neural networks can be expressed in fraction of that complexity and evolution is even simpler. Computations that emerge from these things are often complex beyond our comprehension, and hence it is a good heuristic to distrust simple explanations in the sciences about emergent phenomena such as psychology, sociology, economy and biology.
The worst aspect of our universe might be that it does not care about our feelings. We have evolved to gain status in terms of competence and knowledge, but most edifices of ideas eventually collapse (beyond ars gratia artis), because there is only one truth and many ways to perceive it. Computationalism renders centuries worth of philosophical, religious and perhaps even mathematical inquiry redundant. People have erected their careers upon claims that can be casually explained away by simple computational mechanisms. It's easy to see how the simplicity could be perceived as insulting. It's a tragedy, really and Bach might be a bit too insensitive about this.
It can be demonstrated that GANs learn to model certain image contents with separate input variables like the presence of certain objects or 3D rotations [0] all by themselves, so there is definitely more going on than simply learning to paint some image patches at certain locations and smoothly interpolating them. An image patch reveals itself directly in the training data, so the image patch could simply be stored in the network weights, but a 3D rotation or the presence of a certain class of objects cannot be learned by simply copying image patches into the weights. A 3D rotation requires at the very least a computation of foreshortening, occlusion perhaps based on a depth map. An object detector requires at the very least a feature hierarchy, perhaps binding computations that relate different object parts to the whole object.
Neural networks basically learn to implement nearly arbitrary computations (up to a certain circuit depth) to produce the desired output, so you can also think of deep learning as program mining of a certain program space that is reachable by the adaptive functions in a neural network. Stephen Wolfram has written about that in his blog and talked about it in one of his podcasts. So it's basically magic much like evolution, and it will probably destroy us because it's simply too powerful.
> almost every image posed on reddit is a form of piracy
This statement seems questionable and needs to be backed up by statistics. Looking at the front-page, the most entertaining items are original content contributed by the users. Clearly, piracy is happening too, perhaps in the total uploads more so than what reaches the front-page, but I distrust the notion that this is associated with substantial economic loss. And in the utilitarian analysis of the problem it matters whether it is substantial because of the implications on freedom of speech that may come from an automated filter.
The Carbon X1 6G is 120 grams lighter and has much more ports. Gonna leave Apple for good. The MacBook Air was an amazing machine for the past 8 years, but Apple lost its way. Too many cargo cults (slimness, few ports, radical changes, secretiveness, to name a few) and too little closeness to their customers (poor repairability, high prices, unwanted features).
Conversely, those algorithms which cannot be converted to sequential possibly only run on exotic Turing Machines like ones with infinite bands or infinite parallelism and those likely don't have a physical equivalent because it looks like our universe only supports finite computation. In other words, the algorithms which violate Church-Turing might not be interesting at all, or only for analysis in an idealized setting.
How out of possibly hundreds of different operations that could take the hex table and and that date in possibly 5 different date formats as arguments, would you find out that it's shift key and YYYYMMDD?
The problem is that more technology also creates even more dangers that we would need to work on the same time. Whether we invest more or less in tech does not seem to affect the outcome, but how much of it is done with foresight and additional safety research.
https://www.youtube.com/watch?v=LKnqECcg6Gw (Computer Color is Broken, MinutePhysics)