Wow. Take a single portrait photo, then match it up with someone on video, and you can make a pretty good 3d avatar of the person in the portrait. Trust me, click through and look at this… it’s going to be really important that people understand the limits of what you can trust with your eyes. I have an older friend who would be instantly fooled by this, I am not sure I could even explain the concept to her.
One can go a very long way with “expert system” software, but so called “hard” AI will not exist with the current concept of computers. As long as there are discrete packets of binary data traveling to discrete places and dead air the rest of the time on that path, hard AI cannot be done, no matter how fast and complex it gets otherwise.
With a large enough database, and access to to test enough possibilities, and rules to not waste time testing dead ends, it could write its own software blazingly fast and certainly give results in a required manner, though the results can be catastrophically wrong in ways a child can see.
Any brain cell (human or otherwise) is constantly acting with the time between pulses carrying the information and the connection between cells acting variably depending of everything else that is happening and no two acting the same, thus there is no dead air, and it is that chorus of everything affecting everything else that is perceived as awareness.
That’s a strong claim that relies on quite a few assumptions. Reminded me of this:
The NYT published this nine weeks before the Wright brothers flew.
Humans are bad at recognizing when they’re on a curve. There’s a great (speculative and somewhat alarming) article about AI progress that I think hits a lot of nails right on the head.
I’m not sure whether or not quantized processing can ever achieve sentience. All I can say is — as I sit here replying to people across the globe via a handheld piece of glass that can access the entirety of human knowledge — I wouldn’t rule it out.
Have thought about this and similar subjects quite a bit. Along the same line, is the possibility that our existence is already not as we believe it to be. Two parallel possibilities are that we are actually part of a simulation and not biological beings, or that we are biological beings where all of our external stimulus is being generated by a massive simulation. Maybe I am a non physical character in a universal scale role playing game, or I am a biological being immersed in a universal simulation. While I do not believe it likely that y’all are unreal, I can’t argue against the possibility. Certainly the speed at which simulations have evolved in my lifetime would suggest that the development of a mathematical simulation indistinguishable from reality is not only possible in a 100+ years but likely.
Regardless of whether it constitutes or, ever will constitute “hard AI” or actual consciousness, the technology we have now that we are calling “AI” is being used in ways that require careful attention to spot complicated interactions and forestall unwanted outcomes.
If you have drive time or, other time spent doing things with low cognitive commitment, there are also some very engaging podcasts with appropriate experts covering some of this, including (and in no particular order):
I recall a basic philosophical issue from a girlfriend who was majoring in philosophy. Either reality can be observed or it can’t. If it can’t there is no point in exploring that, so the alternative not only provides results, but insights into mechanisms.
I see many discussions of nerves acting like wires which they do not. The data is transferred by reversing cascades of sodium and potassium ions. The rate and pattern of those cascades vary a lot but do not cease and when it comes to a synapse the effect is also quite variable and influenced by what is going on at all the other synapses. That is modified and changes what it sends out continuously. Multiply that by a few billion cells and the simplicity and uniform result (either on or off) of digital computers cannot begin to catch the nuances of the way nerves work. One does not have to follow this to the specifics of the end results to see that the differences increase as the complexity increases.
That is why I put in what I did. Now if some basic piece is invented that has the nuances of a nerve cell then the question would need to be revisited, but the basic binary every bit equal to all others but for address cannot get there.
I am of a mind that computional power won’t be a limiting factor. Binary computers are already the past. If you think of quantum computers compared to traditional computers as one single technological advance, in 100+ years there are likely to be dozens of such evolutionary advances. Not just increases in processing power but how we define a computer. Moore’s law will be irrelevant. The most visionary of computer engineers today will be as prescient as a Neanderthal. I believe this will only take 100-200 years but lets say it is 1000 years.
But all that aside… If I am a simulation, my perception of reality is what the program tells me reality is. Wouldn’t even have to be very complicated or computationall intensive. I would have no way to tell the difference.
If I am actually a biological being and all my external stimulus is coming from a simulation then how my brain works doesn’t have to be replicated or even understood. It just needs to be fooled.
As if civilization much less the species is still around to be involved. That is the problem I see with all the magical thinking. Having no idea how a problem can be solved they assume it will be, even as those things where the steps are known the process to accomplish them is whistling past the graveyard while the specific steps are optimistic at best. I am wildly in favor of positive progress and will do what I can but the math does not work in the end.
i hadn’t, that’s cool, too. i’d done some lego custom minifig research a couple of years ago for a work project, but we ended up not using my (really effing cool, if i do say so myself) design for our internal awards wall.