On singularity… long run trend of science is that new ideas become harder to find faster than our capacity to find new ideas improves… is AI like a deep sea oil rig? Are “ideas” like a deportable resource?

Additionally never mentions the important context that Australia is a leading exporter of fossil fuels. While fossil fuel electricity generation has decreased fast… total FF consumption has decreased much more slowly… and total FF production has not decreased at all.

Misleading headline. Conflates electricity with total energy. Never mentions in article that electricity accounts for only ~20% of final energy use in Australia and probably less than half of useful energy services.

But historically the next idea has become harder to find and the new problem has become harder to solve *faster* than our scientific capabilities have increased. The net result is reduced resource productivity of scientific innovation.

A >10X increase in rate of progress today than say a century ago. In fact if someone time traveled back to 1800 with modern computers and scientific equipment we could probably rediscover all of modern chemistry, physics, biology in only a few decades…

The idea of a singularity depends mathematically on our scientific capabilities increasing faster than our scientific problems get harder. People saying that advanced AI will enable “more than a century of progress in less than a decade” need to explain why we don’t *already* see …

Beyond 2050: From deployment to renewal of the global solar and wind energy system The global energy transition depends on large-scale photovoltaic (PV) and wind power deployment. While 2050 targets suggest a transition endpoint, maintaining these systems beyond mid-century requires...

This new preprint says same stuff: arxiv.org/abs/2502.04205

1. www.sciencedirect.com/science/arti...
2. www.econstor.eu/handle/10419...
3. www.sciencedirect.com/science/arti...

How does the view of china “winning” the renewable energy transition square with the view of unequal ecological exchange theory that views this spatial restructuring of renewable energy supply chains as representing a net appropriation of resources and waste assimilation capacity by US/EU?

Electric vehicles are still only 1/3 of total cars in use. And rapid transformation of one sector in one country does not translate to rapid transformation in all sectors globally. In addition mass buildout of electric cars is gross misallocation of remaining carbon budget.

Mathematically an energy transition that makes resource circularity most feasible would mean that 1) the duration of transition exceeds the average lifespan of renewable power plants and 2) the growth of renewable power is linear.

That depends on many assumptions about resource availability, circularity, substitutability. But the best way to plan for a sustainable renewable energy system is to reduce energy demand thereby allowing growth curves that make circularity most feasible.

Whatever energy transition we do we have to essentially repeat every 20-30 years, indefinitely. Not just an issue of “are there enough minerals for the transition?” but how long can we sustain continuous recommissioning of these systems?

On the rules of life and Kleiber's law: the macroscopic relationship between materials and energy Energy; Mass; Kleiber's Law; Metabolic ecology; Limits; Conservation.

Very similar to this result that looks at scaling of energy use and material stocks. www.cell.com/heliyon/full...

But the current growth dynamics make these waste flows hard to manage. Linear growth of renewables means constant materials flows. Logistic growth means recurring waves. Exponential growth means recurring pulses.

On the rules of life and Kleiber's law: the macroscopic relationship between materials and energy Energy; Mass; Kleiber's Law; Metabolic ecology; Limits; Conservation.

See this related article: www.cell.com/heliyon/full...

Not sure why there would be such a larger discrepancy between the two datasets. But still taking the larger number based on higher quality data (Ember) solar still likely provides small portion of total energy use

There seems to be something wrong with the data shown on OWID. On the page for electricity generation page they show 18.8 TWh for solar but on primary energy consumption page they show only 4.4 TWh for solar apparently after adjusting it upward with substitution method.

This somewhat undercounts renewables. Accounting for higher end use efficiency would require a larger multiple of ~3-4 units of FF substituted by one unit of primary electricity. Best comparison would be to look at energy services. But substitution method is decent as first approximation.

Important to note substitution method isn’t *perfect* and still somewhat undercounts renewables. Substitution methods multiplies primary electricity sources (solar/wind/hydro/nuclear) by about ~2-3 to account for fossil fuels lost in electricity generation.

But this is misleading not only because monthly electricity shares do not translate to annual electricity shares (changes in solar resource and electricity demand throughout the year) but mostly because electricity use is still small part of total energy use.

Compare this to electricity. Low carbon electricity sources contribute almost half of Pakistans electricity generation. And solar is growing rapidly making up 10% in 2024 and 25% in the *first months* of 2025.

Sure I was just saying people often misquote Smil as saying renewables transition will take “centuries” when he actually says global energy transitions take decades. And today solar and wind make up ~5% of global energy use. Still marginal.

What a rapid transformation where solar and wind each account for ~1% of total energy use in Pakistan. Note the substitution method used here accounts for efficiency improvements associated with decarbonization and electrification.

Most importantly electricity is a small share of their total energy use (even adjusting for efficiency) so you can reach higher shares of electricity faster. What matters is the share of total energy use not electricity. Also note Pakistan is a country with rich solar resources.

Did Smil make specific predictions about Pakistan? Or about global transition? In all his books he writes that national transitions are faster. Note several characteristics allow fast growth of renewables in Pakistan…

Most people saying Smil has been proven wrong are often misquoting him. This is his most recent conclusion on energy transition.

Previous energy transitions were limited by infrastructure and end use… renewables don’t face these issues yet. Current growth is basically investment limited and at small shares we can sustain high growth rates.

Agreed. People forget solar and wind are still! marginal globally. Provide ~5% global energy adjusting for efficiency. The rapid growth *at low shares* is mostly made possible by fact that *at low shares* renewables can be added w/out changes to energy infrastructure or end uses.

And this is the actual timeline of previous energy transitions that the compares rise of renewables to. Decades not centuries as many people falsely quote.