We can push energy costs to next to nothing as solar furnaces and abundant wind and solar power are brought online worldwide over the next century.
Unlikely given that alternative energy sources don't pack the same power punch as fossil fuels.
That's true but they are also vastly underutilized currently. Germany went from virtually no solar and wind power usage to wind generating 6% of total electricity and solar generating 3%. They are shooting for total renewable production by 2050 with solar generating 25% of that total. And they are not using solar furnaces (because they really don't have enough sunny days).
While the US probably can't make such a commitment (in large part because it is the country of can't do
having sacrificed it's future to a extremely cynical view of "practicality" completely dominated by very short term thinking) total renewable energy right now accounts for 14% of total electricity production. Even such a visionless, leaderless society has stumbled in the right direction.
It is not a big technological problem to take the majority of the world and switch them to 25% solar electricity generation within a 20 year period. It's just a matter of leadership.
Currently the world produces 20 TW of electricity per year, 6 TW more than in 2000. The large jump is most likely due to 1/3 of China being pulled into the 20th century. Renewables in China account for 17% of it's electricity generation.
Current world electricity production is 11% of the energy that falls on Earth from the Sun every hour. We can actually harness that energy for real (as in, for real from an engineering perspective) to eliminate energy production problems. Kurzweil has also noted that solar power production seems to have begun to grow exponentially a la Moores Law. In this case every 2 years world solar power doubles. However this observation only goes back 20 years whereas Moores Law goes back to 1965 (but it's observations go back to the 50's - so we have a 60 year history of exponential growth in computer memory and processing power).
Of course right now we do need petroleum to power machines. This is because it used to be plentiful and we built machines to run on processed oil. However we have produced machines that just run on electricity in the past (mostly in the early part of the 20th century). Secondly we can also produce synthetic oil now (the Germans had to do so toward the end of WWII). Well beyond that dirty oil from coal production, we can in the next scientific generation (20 yrs) really begin to use large scale synthetic oil directly from constituent chemicals. The basic work has been done for some time but is not commercially viable yet.
Kurzweil has also noted that parallel to Moores Law human utilization of mechanical power also seems to exhibit exponential growth. However I cannot find Kurzweil's documentation of this (it's in one of his books) an d specifically I can't find a reference to the period involved.
You also need fossil fuels to build and maintain alternative energy sources.
Currently because our entire world was built on fossil fuels. But we can wind them down if we make the commitment.
However we will also find that we still need nuclear power and will think through safer nuclear design. Additionally we could also revisit Dr. Gerald K. O'Neil's ideas on large solar arrays in space.
Nuclear power's by-product unfortunately has to be isolated from people and the environment for spans of geological time. Not a wise source of energy.
You also need fossil fuels to build, maintain and decomission nuclear power plants.
I am talking about inherently safe designs that cannot result in another Fukashima or Chernobyl. As for spent nuclear fuel, we can either store it or send it into the Sun. It's a matter of commitment. We can really move to a completely electrical driven economy for this with the vast majority of electricity generated from renewable energy and the rest generated from nuclear sources. Over time electricity costs actually drop to 0.