An ability to perform more calculations then a human mind does not mean it will beat us.
True. But it does provide the capacity to "beat us".
First, we self assemble from readily obtainable materials out of a self regulating biosphere. Where as this machine would have to be built and maintained by our industry.
Say rather "by industry". Including the word "our" is unnecessarily specific.
Second, there are fucking billions of us. So sure.. we might be able to build some machines that are smarter then ONE person but there are again... fucking billions of us.
Great point. Except that digital entities can be duplicated with stupifying ease while us biologicals require decades of education.
Third, the machine will have its programming directed by us. It will at best be a slave of whomever paid for it to be created.
By definition, adaptive code of sufficient complexity to be considered artificially alive won't be slave to anyone.
Fourth, that programming will be directed at preforming some task where as our task is generally the propagation of our genes with everything else being some sort of weird byproduct.
Again, by the time we've got code that fits the definition of artificial intelligence, it's no longer just an expert system.
Fifth, we have hundreds of millions of years of evolution behind our programming. And I don't think any collection of programmers is going to surpass it in the next century.
Unfortunately, almost all of that evolution was literally trial-and-error, mostly error. We've got a whole bunch of DNA that is literally not expressed because it's useless. That evolution was about getting from primordial slime oozes to walking upright. The lessons "learned" in the middle about how to do photosynthesis while we were ooze-like isn't even vaguely useful to us. Also, aside from symbiosis with the AIs we're talking about here, we have nowhere to go. Evolution takes so long, a decent climate change on this planet and we're screwed where AI can iterate purposefully to solve its problems.
Eventually might there be robotic rivals to humanity? Sure... but not any time soon.
Soon? As in this afternoon? No, probably not. Soon as in within a decade? Again, probably not. But really, once we develop one instance of something truly self-programming and "intelligent", its complexity growth will be very rapid, just like all of our technologies have been. Just like... flight. The hard part was getting the first airplane off the ground. From there... space shuttles in very little time.