essentially its becuase we are on a trajectory to nowhere with our current tech. lets say we get to 5nm, then what.
The human brain is amazing. Sure, there are flaws, but from all the struggles and despair we extract brilliance from it.
Probably one reason the prediction of brain-like computers coming true were pushed out to 2050's to 2000's. They realize there's still far more going on than what they know.
essentially its becuase we are on a trajectory to nowhere with our current tech. lets say we get to 5nm, then what.
But at some point does the ability to do math really, really fast allow (at some point) one to simulate the entirety of the chemistry inside a brain? That's one of the more interesting ends of scaling in my opinion.
I know I've already seen one simulation of a ~300 neuron worm's brain.
The problem with that is more that we don't fully understand everything that happens in the brain yet, so we can't just write a program that simulates it.
Actually you don't need to simulate the brain. All you need is to get identical output when the same input is applied. Typical black box problem.
To put in a way pc geeks can understand, the first IBM BIOS was cloned by haveing people make tables of all the possible inputs and all the possible outputs from those inputs. Those tables where then given to another group of people to create the comparable "clean" BIOS. The clone BIOS programmers had no knowledge about how the IBM BIOS functioned.
the common art major physics notion is that "binary computers can think in yes and no, quantum computers can also think in maybe", that is total fantasy and fiction that implies that quantum computers will gain self awareness by virtue of being quantum.
first of all classifying current computer tech as "binary" is obtuse. operations are more complex than merely "yes / no"; The individual components of an individual logic gate are binary based... but that is because that is the most simple and direct way to build them, actual processing elements are built from those simple elements and are far more complex and versatile. And computer communication is commonly reduced to binary, but machine code has more flexibility then merely "yes/no" and and the actual storage often is in more then binary (SSD MLC for example, where each cell has 4 states, aka quadranery).
Quantom computing only gains in efficiency but does not somehow enable you to calculate functions that cannot be performed in current computers (just, perhaps much more slowly; it might make it more practical to perform certain calculations due to speedups).
the reason brains are so different than computers is because computers are constructs that do math (and do it well, and very very quickly), organic brains are not math processors but chemistry factories; we are very much deficient when it comes to math compared to a computer.
Excellent post. The human brain has been repeatedly underestimated, because we fail to realize just how badly computers are playing catch-up.The human brain is amazing. Sure, there are flaws, but from all the struggles and despair we extract brilliance from it.
Probably one reason the prediction of brain-like computers coming true were pushed out to 2050's to 2000's. They realize there's still far more going on than what they know.
haven't gotten to the cosa002.pdf yet, but I agree with most of what the guy/gal is saying in the temporal intelligence article. I have been saying that the GOAI guys are going about everything wrong for years now.
my biggest issue with the temporal intelligence article is that it seems to focus on creating an AI without understanding it (by emulating the neuron's input/output and letting one grow) rather then fully understanding how and why it works the way it does first and then constructing one. For example, we know humans who suffered brain damage would occasionally see major changes in their personality, including becoming psychopathic. Before creating AI it would be important to make sure you know why and how to make it be empathic. futuristic fantasy always assumes that humans:
1. Create AI that is psychopathic.
2. Enslave that AI.
since I am not a moron, I oppose to both.
One of the major things I want to achieve in life is to create AI, and I intend for that AI to be empathic, capable of love, and treated as my own child (just because I would have created it with my hands and brain rather then my reproductive organs doesn't mean it isn't my child).