8 ms·

"Admits" seems a bit harsh, like he was doing his best to hide something from the world. By all accounts, it looks like he had written out a reasonable explana

"Admits" seems a bit harsh, like he was doing his best to hide something from the world. By all accounts, it looks like he had written out a reasonable explanation as to why he predicted solar would not become cost effective. At the end of the day, if something doesn't make financial sense, it won't last long without subsidies from the government. The assumptions behind his predictions changed and therefore so did the conclusion. Seems pretty reasonable. Like he says, solar power's intermittency is still an issue in non-sunny places, but in areas with lots of sunlight, the landscape of power generation is changing for the better.

beat 10y ago
In other words, he's being a scientist, and adjusting his conclusions to suit the available evidence. Which, to people not familiar with the scientific method, is shocking.
- Joof 10y ago
  Most people are more familiar with the standard argument of 'nuh-uh' when things change.
- r0muald 10y ago
  Yes, but only to a certain extent, I think. A truly honest, and most importantly, useful analysis wouldn't be limited to admitting that previous research did result in wrong predictions. I'd like to see a breakdown of what was wrong in the previous prediction, and to be sure that it's understood well, I would like to see also a true prediction based on modelling and not limited to .. observing what is happening with solar power right now. Am I too demanding of a scientist? Adjusting my predictions based on what I observe, he was wrong in 2008 and I have a low expectation that his new predictions will turn out correct.
  - btilly 10y ago
    You are indeed being too demanding, as can be seen when you know the general pattern of how the prediction was made. Most industries, most of the time, show exponential improvement over time on whatever their primary measure of quality is. Moore's law is the most famous simply because its speed has had such dramatic effects, normally improvement is smaller, for example the 7% increase/year in battery capacity for a fixed price that Tesla's future business plans rely on. See this graph of the cost of memory storage in hard drives: http://www.jcmit.com/disk2015.htm http://www.jcmit.com/disk2015.htm. (Note that an exponential curve is approximately a straight line on a log graph.) Every so often there will be breakthroughs which show up as sudden jumps. After the breakthrough is assimilated you'll often see a pause because research plans were disrupted by the breakthrough, and then the curves start going again. Sometimes at a different speed. As an example on the above graph, the sudden order of magnitude price drop between 1995 and 2000 is due to the commercialization of Giant Magnetoresistance (GMR). This pattern of technological advance is well-known, but the jumps are not predictable in advance. However even after such a jump, your best bet for projecting continued improvement over years to decades is the same methodology.
Amorymeltzer 10y ago
I particularly liked his point about bringing processes with a high electricity demand to those areas and cutting down on cost across the board. He gives aluminum, ammonia, and desalination as examples, which could do a lot for reducing costs of production while still overall "greening" the process. If they can take advantage of the cycling in prices, I can see regions like the American Southwest selling this as a business incentive.
- fovc 10y ago
  It's not a novel idea. E.g., Dubai produces a lot of aluminum, since importing ore and exporting Al is cheaper than exporting natural gas. [1] https://en.m.wikipedia.org/wiki/Dubal https://en.m.wikipedia.org/wiki/Dubal
  - ZeroGravitas 10y ago
    Hydro power and aluminium are a traditional combo too.
- ThrustVectoring 10y ago
  The transportation fuels example is even more exciting, IMO. Transportation fuels are basically simple hydrocarbons that burn with oxygen to turn into CO2 and H2O. There's only engineering, capital, and energy costs that need to be overcome - then we can run that process in reverse and turning our current transportation infrastructure carbon-neutral.
  - afarrell 10y ago
    > run the process in reverse Does the fact that CO2 makes up a very small portion of the atmosphere make this hard?
    - semi-extrinsic 10y ago
      One of the big financial problems for CCS (carbon capture and storage) is that CO2 is essentially worthless. If a market would suddenly arise for concentrated CO2, I do not think we would have much of a problem to fill demand.
mtgx 10y ago
Or maybe it was just another case of blindness to exponential growth curves and the law of accelerated returns. Many thought solar is inevitable, even when most were skeptical. Just like many thought the same about EVs (although it's becoming trendier to believe in EVs now, too, especially in light of Model 3's pre-orders, but it wasn't always the case). Imagine if the founders of Solar City also thought like this guy. Suffice to say they probably wouldn't be owning one of the largest US solar companies in the US right now.
- Animats 10y ago
  "Imagine if the founders of Solar City also thought like this guy. Suffice to say they probably wouldn't be owning one of the largest US solar companies in the US right now." Solar City did think like that guy. Solar City is heavily dependent on subsidies and the ability to sell power to utilities at retail, not wholesale, rates. Solar City pulled out of Nevada when Nevada took away some of those benefits.[1] Their business model requires subsidies. The big utility-scale plants are now profitable at wholesale rates. Even in Nevada.[2] That's what matters. [1] http://www.slate.com/articles/business/the_juice/2016/01/solarcity_and_vivint_pulled_out_of_nevada_is_solar_power_losing_at_the_state.html http://www.slate.com/articles/business/the_juice/2016/01/sol... [2] https://en.wikipedia.org/wiki/Copper_Mountain_Solar_Facility https://en.wikipedia.org/wiki/Copper_Mountain_Solar_Facility
LanguageGamer 10y ago
To what extend can the decreases in cost be attributed to the subsidies? The subsidies shift the demand curve enough to incentivize R+D... or is my Econ 101 viewpoint missing something?
- btilly 10y ago
  Supporters of the subsidies will claim 100% of the credit. Opponents will claim it would have happened regardless. Both sides will put forth impressive looking studies proving their case. My belief is that it wouldn't have happened nearly this fast without both subsidies, and a widespread awareness under investors that there are going to be government incentives for a huge renewable market. But figuring out the amount is impossible.
ryangittins 10y ago
Agreed! His predictions were spot-on given all the information he had to work with. It's not like this guy was rabidly against solar power or anything; in fact it seems like he's pretty delighted to be wrong. Who wouldn't be? Solar's getting cheaper! Everyone seems to be projecting their own egos onto this guy. Their minds say, "If I were proved wrong and had to admit it, I'd be mortified." These people aren't scientists. They're being emotional. Again, props to this guy.
mrb 10y ago
"it won't last long without subsidies from the government" This is a fantastic example of how government subsidies can help. Had there been no subsidies, the demand for solar power would have been smaller, the industry would not have been compelled to compete that much on improving solar power, and this Harvard researcher's prediction (solar not cheap enough in the near future) might have turned true. He said these subsidies make no sense because solar is far from being competitive, yet they are the very reason solar became competitive.

beat 10y ago

In other words, he's being a scientist, and adjusting his conclusions to suit the available evidence. Which, to people not familiar with the scientific method, is shocking.

Joof 10y ago
Most people are more familiar with the standard argument of 'nuh-uh' when things change.
r0muald 10y ago
Yes, but only to a certain extent, I think. A truly honest, and most importantly, useful analysis wouldn't be limited to admitting that previous research did result in wrong predictions. I'd like to see a breakdown of what was wrong in the previous prediction, and to be sure that it's understood well, I would like to see also a true prediction based on modelling and not limited to .. observing what is happening with solar power right now. Am I too demanding of a scientist? Adjusting my predictions based on what I observe, he was wrong in 2008 and I have a low expectation that his new predictions will turn out correct.
- btilly 10y ago
  You are indeed being too demanding, as can be seen when you know the general pattern of how the prediction was made. Most industries, most of the time, show exponential improvement over time on whatever their primary measure of quality is. Moore's law is the most famous simply because its speed has had such dramatic effects, normally improvement is smaller, for example the 7% increase/year in battery capacity for a fixed price that Tesla's future business plans rely on. See this graph of the cost of memory storage in hard drives: http://www.jcmit.com/disk2015.htm http://www.jcmit.com/disk2015.htm. (Note that an exponential curve is approximately a straight line on a log graph.) Every so often there will be breakthroughs which show up as sudden jumps. After the breakthrough is assimilated you'll often see a pause because research plans were disrupted by the breakthrough, and then the curves start going again. Sometimes at a different speed. As an example on the above graph, the sudden order of magnitude price drop between 1995 and 2000 is due to the commercialization of Giant Magnetoresistance (GMR). This pattern of technological advance is well-known, but the jumps are not predictable in advance. However even after such a jump, your best bet for projecting continued improvement over years to decades is the same methodology.

Amorymeltzer 10y ago

I particularly liked his point about bringing processes with a high electricity demand to those areas and cutting down on cost across the board. He gives aluminum, ammonia, and desalination as examples, which could do a lot for reducing costs of production while still overall "greening" the process. If they can take advantage of the cycling in prices, I can see regions like the American Southwest selling this as a business incentive.

fovc 10y ago
It's not a novel idea. E.g., Dubai produces a lot of aluminum, since importing ore and exporting Al is cheaper than exporting natural gas. [1] https://en.m.wikipedia.org/wiki/Dubal https://en.m.wikipedia.org/wiki/Dubal
- ZeroGravitas 10y ago
  Hydro power and aluminium are a traditional combo too.
ThrustVectoring 10y ago
The transportation fuels example is even more exciting, IMO. Transportation fuels are basically simple hydrocarbons that burn with oxygen to turn into CO2 and H2O. There's only engineering, capital, and energy costs that need to be overcome - then we can run that process in reverse and turning our current transportation infrastructure carbon-neutral.
- afarrell 10y ago
  > run the process in reverse Does the fact that CO2 makes up a very small portion of the atmosphere make this hard?
  - semi-extrinsic 10y ago
    One of the big financial problems for CCS (carbon capture and storage) is that CO2 is essentially worthless. If a market would suddenly arise for concentrated CO2, I do not think we would have much of a problem to fill demand.

mtgx 10y ago

Or maybe it was just another case of blindness to exponential growth curves and the law of accelerated returns. Many thought solar is inevitable, even when most were skeptical. Just like many thought the same about EVs (although it's becoming trendier to believe in EVs now, too, especially in light of Model 3's pre-orders, but it wasn't always the case). Imagine if the founders of Solar City also thought like this guy. Suffice to say they probably wouldn't be owning one of the largest US solar companies in the US right now.

Animats 10y ago
"Imagine if the founders of Solar City also thought like this guy. Suffice to say they probably wouldn't be owning one of the largest US solar companies in the US right now." Solar City did think like that guy. Solar City is heavily dependent on subsidies and the ability to sell power to utilities at retail, not wholesale, rates. Solar City pulled out of Nevada when Nevada took away some of those benefits.[1] Their business model requires subsidies. The big utility-scale plants are now profitable at wholesale rates. Even in Nevada.[2] That's what matters. [1] http://www.slate.com/articles/business/the_juice/2016/01/solarcity_and_vivint_pulled_out_of_nevada_is_solar_power_losing_at_the_state.html http://www.slate.com/articles/business/the_juice/2016/01/sol... [2] https://en.wikipedia.org/wiki/Copper_Mountain_Solar_Facility https://en.wikipedia.org/wiki/Copper_Mountain_Solar_Facility

LanguageGamer 10y ago

To what extend can the decreases in cost be attributed to the subsidies? The subsidies shift the demand curve enough to incentivize R+D... or is my Econ 101 viewpoint missing something?

btilly 10y ago
Supporters of the subsidies will claim 100% of the credit. Opponents will claim it would have happened regardless. Both sides will put forth impressive looking studies proving their case. My belief is that it wouldn't have happened nearly this fast without both subsidies, and a widespread awareness under investors that there are going to be government incentives for a huge renewable market. But figuring out the amount is impossible.

ryangittins 10y ago

Agreed! His predictions were spot-on given all the information he had to work with. It's not like this guy was rabidly against solar power or anything; in fact it seems like he's pretty delighted to be wrong. Who wouldn't be? Solar's getting cheaper! Everyone seems to be projecting their own egos onto this guy. Their minds say, "If I were proved wrong and had to admit it, I'd be mortified." These people aren't scientists. They're being emotional. Again, props to this guy.

mrb 10y ago

"it won't last long without subsidies from the government" This is a fantastic example of how government subsidies can help. Had there been no subsidies, the demand for solar power would have been smaller, the industry would not have been compelled to compete that much on improving solar power, and this Harvard researcher's prediction (solar not cheap enough in the near future) might have turned true. He said these subsidies make no sense because solar is far from being competitive, yet they are the very reason solar became competitive.