Can solar costs keep shrinking?

>decentralization does not work.

This is not quite true, because the vast majority of solar generation is consumed on site, avoiding the transmission and distribution costs of delivering electrons that would normally be necessary. Which is just one of a whole motley of dynamics working in favor of solar at larger scales: the arc of solar generation over a 24hr period almost perfectly coincides with actual market demand over the course of the day, with the exception of the afternoon/evening "duck curve", so it's actually relieving pressure on peaking generation.

The important fallacy here is assuming that counterbalancing for the cycles of solar generation requires new investments, when in fact it's relieving pressure on infrastructure that already exists and is already serving those exact counterbalancing purposes. This is in addition to the benefit of offsetting alternative forms of base load generation.

To be clear you are right in an important sense about a pretty fundamental thing. There is indeed a tipping point, and when we reach that tipping point of grid penetration all of the points you have raised will indeed become not merely relevant but crucial. And I forget the exact number, but my understanding is we're nowhere near that tipping point right now. I want to say around 20% of the overall grid being generated from solar power is the tipping point but I'm not sure if that's accurate.

It's kind of like the argument sometimes people want to make about taxes which is that if you overtax it is a drag on the economy, which is hypothetically true but it's true at a given tipping point and it's a tipping point that we're not anywhere near, which doesn't tend to stop the advocates from bringing it up all the time.

It also neatly corresponds with the mass rollout of nuclear power so it's possible it's just the classic measurement issue of "primary energy" vs "useful energy", sometimes called the primary energy fallacy, which makes fossil fuel based systems appear 4x (or probably more going back in time) better than they actually are.

The range is so wide though. Look at a particular narrow band of the GDP per capita and see how wide the energy consumption per capita is for even a very narrow slice of GDP per capita. The dispersion gets even wider as you get up towards the high end of the GDP axis.

The author asserts that we should see 5x GDP/c if we had 5x power usage per person and their own graph shows that that's not the case because it's a flawed assumption that ignores increases in efficiency and transitions away from energy intensive manufacturing to service based.

The best evidence for that is that the GDP/c didn't fall off when we fell off the HA curve. In fact I took the GDP per capita and Energy consumption per capita data from world data bank and the ratio between the per capita GDP vs the per capita energy consumption has been going up steadily since we stopped following the HA curve.

Between 1960 and 1970 the ratio between GDP and Energy Consumption per capita was essentially static at .74 then after 1970 the ratio begins to increase showing we're producing more per unit of consumed energy at a nearly linear rate. By 2014 which is the last year they had the Electric power per capita data the ratio was all the way up to 4.2. Eyeballing it the relationship is almost perfectly linear each year we get a little better at producing GDP for each kWh we consume.

I even redid the calculation based on raw energy use in kg oil equivalents and it gets even more drastic. 1960 to 1970 it goes from .53 to .69 GDP/kg oil equivalent [0]. Then after 1970 the rate increases quite distinctly going from .69 to 1.58 in 1980, 3.11 in 1990, 4.5 in 2000, and 6.79 in 2010.

It's pretty clear from the data that we're getting better at producing things with the same amount of energy. It's an assumption that simply making more power would increase the amount of things made.

Electricity use per capita: https://data.worldbank.org/indicator/EG.USE.ELEC.KH.PC?locat...

Energy use per capita in Kg oil equivalents: https://data.worldbank.org/indicator/EG.USE.PCAP.KG.OE?locat...

GDP per capita in current USD: https://data.worldbank.org/indicator/NY.GDP.PCAP.CD?location...

[0] nice.

I used to be radically pro-nuclear, but after seeing some of the data of solar growth and realizing we could just scale things up (like building iphones), I'm fully on board with the Elon strategy of solar + powerbanks for storage.

I regularly bring this up: modern fridges are generally continuous drive (which means they run off DC at some level anyway). And they run all the time(ish).

Waiting for a (mass-)manufacturer to build one with a 48VDC input that anyone can self-install a solar panel and plug-in but switches to a mains-fed 48VDC source as required.

Doesn’t need a battery (the fridge/freezer is one!), gets you some operation during outages, no license required to wire up. Takes advantage of whatever solar is available during the daytime without really wasting any.

Would also love a non-permanently installed doohickey that doesn’t require an electrician to legally install that will gladly push solar into a power strip and handle all the intertie stuff itself.

It shouldn’t be a “whole house or nothing” and a “completely islanded circuit or nothing” dilemma when wanting to partly solarize something. Which I guess Germany has allowed but regulations a big impediment in most places.