How quickly is solar panel technology expected to advance? If this technology is going to follow the path of the personal computer and allow a pocket-sized panel to power a city in the near future, I wouldn't want to be an early adopter stuck with today's lumbering panels on my roof.
Computers have advanced because they benefit from moore's law - and let me belabor the point a little because so many people get this wrong.
Moore's law is an observation, that guides R&D spending for the big dogs; the observation is this: every ~2 years, give or take, the number of transistors you can place cheaply in a certain area doubles. That is, if today's chip has 1 billion transistors in 100 square millimeters, two years from now the same price buys you either 2 billion transistors, or half the silicon area (price) buys you the same 1 billion. This is only a law while it lasts, but it has lasted about 40-45 years, and will probably last a while yet - all doom and gloom predictions have thus far failed to materialize (though undoubtedly there are limits to the way we do things now, we change the way we do things as time goes on; traditional planar scaling ended at 90nm about 5 generations / ~10 years ago, but you haven't noticed.)
What does not really go down is the price of silicon itself. In fact, the price of silicon wafers only really goes down when wafers transition sizes - 8 inch to 12, or the upcoming 12 to 18. Interesting side note: each transition has knocked out about half the competition, because dealing with bigger wafers means massive expenses.
So you see, moore's law benefiting computers is economics more so than physics.
Solar panels, however, don't benefit from smaller transistors. Solar panels are silicon but the cost of silicon is... well, the cost of obtaining, purifying, growing, cutting, processing, and shipping sand, more or less.
Efficiency gains in solar panel energy production don't follow moore's law - entirely different beast. (If they did, we'd have nearly infinite energy in a couple decades, just as we now have seemingly infinite compute power compared to two decades ago.)
Advancement in solar panels is as follows, roughly speaking:
- Efficiency, small improvements, using different materials and geometries. There're black solar panels, which are far more efficient than your usual panels, but also prohibitively expensive for all but the most critical applications (space).
- Cost reduction; high volume manufacturing means cost all along the supply chain can reduce; time and money make this happen.
- Weight reduction: cutting panels thinner makes them weigh less; less material means less cost, and you also get less cost since they're easier to transport and install.
- Power conversion efficiency in surrounding systems - that is, better and cheaper inverters that last longer.
- Sun-following and weather-predicting gets more power and helps grids balance better.
- Form factors for different needs. For example, foldable flexible solar panel sheets that can be rolled up and carried in a backpack. Downside: far less power generated.
- We'll most likely start putting controllers right into the solar panel silicon - microchips embedded into each cell to improve efficiency by rerouting power around poorly-performing cells (also see solar tracking and weather prediction.)
The short of it all is this: if you put a solar panel up on the roof today, the panels in 20 years will be cheaper, lighter, smarter, but not nearly efficient enough to make your old panels useless.