Leaderboard

I don’t think the issue of *beating* Carnot efficiency arises, because this system is not a heat engine, I.e. it does not rely on converting heat to work. So it isn’t limited by Carnot efficiency any more than, say, a hydroelectric turbine, is it?

It was just the abstract which is linked at the homepage of Imagination Engines which you linked for us. (linked at bottom of left-hand column) https://www.worldscientific.com/doi/abs/10.1142/S2705078521500053 Am still trying to get at the full paper, without a paywall.

Thanks, @Alex_Krycek. I went off and did a lit search, too, found some of the ones you posted, including Thaler's paper in Journal of AI and Consciousness. I want to look at some critical responses (not just from my imagitron...) to some of his claims like Which is a pretty bold claim. Whenever we go from so-called weak AI ( simulating specific tasks found in human intelligence, through pre-programmed algorithms ) to strong AI (Turing test-worthy, responsive to novel situations, showing general intelligence and self-programming growth), some people get all metaphysical about it. If Thaler wants to assign his DABUS subjective feelings, while retaining a "parental" ownership of all the patents, then he is indeed having his cake and eating it. Too much, the magic DABUS. If the machine reaches a point where it can make awful wordplay based on The Who lyrics, then I think it should find itself an android body so as to walk itself into a courtroom and request a cut of the proceeds.

Curious how long an AI's copyright would last.

I think it is more a question of ethics here. Lets say a person makes a new type of vaccine. We know that the person has put a lot of work into developing the vaccine so we reward them with a patent for their contribution to cover the expense of the developing the product. This is in a lot of ways relatively fair. But when we allow AI to be an inventor there are other issues that arise. For example if I use AI to generate every possible version of vaccine that could ever exist then it unfairly eliminates competition from people. This is no different than a totalitarian regime having control over what you can do and would stifle further innovation from others that don't have access to this AI. I think cost is also a major factor to take into consideration. If it takes a lot of money for an AI to generate the new vaccine then it may necessitate awarding a patent to the company that created it otherwise there is no incentive to do so. But if the price is low then there will be an influx of frivolous patents.

I was speculating that Thaler's move is driven by some idealism regarding future AGI and not purely his own financial benefit. Financially, as you've noted, it makes more sense for him to be the inventor and DABUS a tool of cognitive enhancement. Perhaps someday he'll be a "father, " like Dr. Daystrom on Star Trek...

I'm just trying to provide evidence, to you (a scientist), that evil doesn't exist in reallity and humans, any more than it does in any other animal. Our intelligence can hide the truth, ironically backed up by our emotions; we assume it separates us from (the animal's) causality and automatically gives us a choice to ignore our training. The ability to think past our emotions, is our true strength. The consequences of not recognising that as a strength, leads to regression, into a more animalistic approach to justice.