4 ms·

Isn't AGI-Hard everything that seems rather simple for animals? I would think just of robotics and navigation in unknown environment. Everything else like Ches

Isn't AGI-Hard everything that seems rather simple for animals? I would think just of robotics and navigation in unknown environment. Everything else like Chess, Text, Art, Games seems to be solvable by current AI systems. Robotics seems much much more difficult.

tuyiown 4y ago
Yes, the problem is to replicate hundred of millions years of acquisition of heuristics that allows cognition on «natural» environment. Those heuristics defines all bricks our cognition is based upon, think of basics things like objects permanence, we just cannot reason without such ability. To me, as long as no AI manage to replicate those bricks, at least some of them, it won't be able to interact with us in ways that helps us efficiently with the outer, real, world.
- zone411 4y ago
  I predict that this will a research area that advances hugely after we solve text-to-video and text-to-3D (so maybe in 2 years?): learning these types of heuristics from billions of hours of videos. Implementing this in robots is another matter, but there has been some good progress in robotic grasping and manipulation, for example.
tgv 4y ago
Chess and other games are solved, because they use well-defined rules, and have a combinatorial explosion that we cannot comprehend. Text and art are not solved. What would solved even mean for those? There are NNs which produce decent output given a simple prompt, but e.g. art is characterized to a large extent by development over time. Stable Diffusion isn't going to develop by itself. Robotics OTOH is much simpler, but it might depend on your expectations. Look at Boston Dynamics.
- MichaelRazum 4y ago
  Agree, solved is not well defined. Let's just say it becomes hard to distinguish it from a human. For robotics, I just think of a "simple" tasks like picking up an object and moving it to some place. Haven't seen yet something like this in robotics, that looks human or animal like.
  - tgv 4y ago
    Not to just contradict, but computers playing chess or video games is also distinct from human play. I'm absolutely not expecting strong AI (as "AGI" used to be called) to be realized within 100 years, but when it does, it will have distinctly non-human qualities. Intelligence is shaped by the constraints of the hardware, so to speak, and a computer's constraints will be different. It's been predicted that the first self-aware AI will develop severe mental problems, almost, but not quite, entirely unlike Marvin in THHGTTG.
    - MichaelRazum 4y ago
      Fore sure self-aware AI will be weired. I was rather thinking on a simple metric where AI seems to be far behind in "skills". We see this with FSD and also home robotics haven't took off, because real world object manipulation and navigation is in my opinion where AI currently fails. If you solve it, then you could maybe claim, ok the AI is more "clever" than a mouse. Currntly wouldn't agree on that. Of course in this area, it is also a hardware issue.

tuyiown 4y ago

Yes, the problem is to replicate hundred of millions years of acquisition of heuristics that allows cognition on «natural» environment. Those heuristics defines all bricks our cognition is based upon, think of basics things like objects permanence, we just cannot reason without such ability. To me, as long as no AI manage to replicate those bricks, at least some of them, it won't be able to interact with us in ways that helps us efficiently with the outer, real, world.

zone411 4y ago
I predict that this will a research area that advances hugely after we solve text-to-video and text-to-3D (so maybe in 2 years?): learning these types of heuristics from billions of hours of videos. Implementing this in robots is another matter, but there has been some good progress in robotic grasping and manipulation, for example.

tgv 4y ago

Chess and other games are solved, because they use well-defined rules, and have a combinatorial explosion that we cannot comprehend. Text and art are not solved. What would solved even mean for those? There are NNs which produce decent output given a simple prompt, but e.g. art is characterized to a large extent by development over time. Stable Diffusion isn't going to develop by itself. Robotics OTOH is much simpler, but it might depend on your expectations. Look at Boston Dynamics.

MichaelRazum 4y ago
Agree, solved is not well defined. Let's just say it becomes hard to distinguish it from a human. For robotics, I just think of a "simple" tasks like picking up an object and moving it to some place. Haven't seen yet something like this in robotics, that looks human or animal like.
- tgv 4y ago
  Not to just contradict, but computers playing chess or video games is also distinct from human play. I'm absolutely not expecting strong AI (as "AGI" used to be called) to be realized within 100 years, but when it does, it will have distinctly non-human qualities. Intelligence is shaped by the constraints of the hardware, so to speak, and a computer's constraints will be different. It's been predicted that the first self-aware AI will develop severe mental problems, almost, but not quite, entirely unlike Marvin in THHGTTG.
  - MichaelRazum 4y ago
    Fore sure self-aware AI will be weired. I was rather thinking on a simple metric where AI seems to be far behind in "skills". We see this with FSD and also home robotics haven't took off, because real world object manipulation and navigation is in my opinion where AI currently fails. If you solve it, then you could maybe claim, ok the AI is more "clever" than a mouse. Currntly wouldn't agree on that. Of course in this area, it is also a hardware issue.