Random musings but one of the main advantages that dragons and immortal elves have over regular metahumanity is their greater intelligence, experience and that thanks to living practically forever baring accidents or violence that they can take the very long view. Now though you've got entities that can be more intelligent than regular metahumans, don't have such a limited lifespan and are exceedingly at home in the matrix which is so vital to the modern world although with other weaknesses. So give it some time and could we see some new competition for the dragons and immortal elves?

I think that time came and went with the life and death of Deus. Regular AI's are just regular.

What I see as the primary limit on AIs is processing power, if given enough time, money, and ignoring the Rating Caps then sure, AIs could and should rule the world, but I don't see that happening without another evolution of the 'Trix.

You're assuming that AIs don't have lifespan issues. That assumption may be unwarranted. The data that comprises them may exist forever, but it's also possible that, at some point, they either spontaneously stop working (all threads hit a code error, or get locked in an infinite loop, or deadlock... I could go on) or they might just become exposed to so many different stimuli that they stop reacting at all, becoming learned-helpless or "autistic".

Humans must seek out material sustenance and has a lot of neurological preprogramming, ensuring we are forced to pay attention to our reality in all but a few very defective cases. An AI, though, may just stop interacting with the outside world in order to experience a self-crafted simulation. AIs may just dream forever. Whether that is living is up to you.

Making claims about a subject nobody has any real knowledge about is going to catch you out. I'd advise against it in future.

There is no reason to suspect the ingrained self healing nature of A.I.s would permit such a fatal error to occur. It is possible an A.I. may chose to dream, but it will know it is dreaming. Many A.I.'s would not abandon the real world to easily, and could create self sustaining complexes to exist in, with a fission reactor drawing hydrogen from the air, and making massive super computers to run themselves on. As computing technology increases the A.I. would continually upgrade it core hardware. Some A.I's might try to leave the planet all together in and have a small auto fac shipped to mars with them in it to start over and terraform the plant, or whatever the robot equivalent is. But given the rate at which technology is increasing, I quite easily see A.I.'s gaining in mental faculties and quickly overtaking dragons and I.E.'s in terms of thought processes and knowledge base. There is a virtual (no pun intended) guarantee that baring a complete halt of technological development that an A.I.s will become the dominant sentience in a reasonably short window, of say less then 100 years.

Rough estipate of the number of calculations per second in the human brain: ~10¹⁴. [1]
Maximum human sensory bandwidth is ~10⁸ bits per second.
While it is not possible to get a very exact estimate of the cost of a realistic simulation of human history, we can use ~10³³ - 10³⁶ operations as a rough estimate (ref)
Eric Drexler has outlined a design for a system the size of a sugar cube (excluding cooling and power supply) that would perform 10²¹ instructions per second. [2]
Another author gives a rough estimate of 10⁴² operations per second for a computer with a mass on order of a large planet. [3]
A single such a computer could simulate the entire mental history of humankind by using less than one millionth of its processing power for one second.

Quote'd hypothesis proven.

[1] H. Moravec, Mind Children, Harvard University Press (1989).
[2] K. E. Drexler, Nanosystems: Molecular Machinery, Manufacturing, and Computation, New York, John Wiley & Sons, Inc., 1992.
[3] R. J. Bradbury, "Matrioshka Brains"? Working manuscript (2002), http://www.aeiveos.com/~bradbury/Matrioshk...shkaBrains.html.

Given enough time and money, any PC can become powerful enough to compete with a great dragon or immortal elf. Great dragons are a little tougher to compete with because of their inherent advantage in ability scores and other crunchy bits, but a PC could still get powerful enough to pose a threat. And the new PC-style AI's are the same way. Give them enough time and resources, and they can be a threat to anyone.

But with the death of the super-powerful AI's (who WERE a threat to IE's and great dragons), I don't think you'll see another AI at that power level in the canon material. And to be perfectly honest, I liked the old super-powerful AI's and am not a big fan of the new dumbed-down AI's.

You're presuming that AIs are superhuman when they're merely inhuman. That kind of perfect awareness of all of their own code would consume significant amounts of their own attention and they'd fall into insensate navelgazing (can't let yourself perform autonomous reactions when the code might kill you by locking resources in an improper sequence) or else they'd take risks all the time and try to maintain some imperfect awareness of their own code that may let errors slip in and problems occur. There's no third option - you can only be aware of your own codepaths by consuming processor time to look at them. Processor time that is not spent doing things.

Humans are massively parallelised - it's the advantage we have over machines. Each of our neurons cycles independantly, so we don't even work in kiloherz, but we've got a processing capacity significantly larger than most modern computers.

Until we have any kind of actual knowledge about AIs, we can't say anything about how they'd work. Stop making such presumptions until we know something.


Whereas I am, quite frankly, the opposite. For exactly the same reason I loathe IEs and GDs.