Links #21
Renewables love nuclear, heatwaves are solvable, economies love welfare and more.
1.
The 2035 report details how we can get to a 90% renewable grid by 2035 using solar, wind, and batteries for 70% of our energy, with nuclear (15%), hydro (5%), and natural gas (10%) making up the rest.
They argue that this grid would be just as dependable and energy would cost slightly less. Seems reasonable, though I think 2040 is a better timeline.
Understanding Solar Energy by Bryan Potter goes into the math of solar for grid energy. Solar can clearly provide 40% of energy with minimal cost declines.
EDIT: see also Can We Afford Large-scale Solar PV? - by Brian Potter
I can’t help thinking that more nuclear power would make this process so much easier. Nuclear already provides ~20% of U.S. energy needs today but that number is set to fall as power plants get decommissioned. A grid with 30% nuclear would only need 55% wind and solar. That’s 3.5x higher than today, but trivial with current renewables growth.
Even better, nuclear makes squeezing that last 10% of natural gas much easier. Nuclear is an excellent compliment to renewable energy generation. It’s great for baseload power and can provide cheap district heating. It might even be feasible to use the steam turbine in decommissioned coal and natural gas plants.
One improvement that would make nuclear particularly enticing is thermal energy storage. Moltex Energy is combining molten salt nuclear reactors with molten salt thermal storage so that the reactor can run consistently but dispatch energy to the grid when it’s needed.
Nuclear can also provide goods that renewables have a hard time providing like supercritical water or neutrons for transmutation. Enhanced geothermal or fusion could fill these roles as well.
2.
Factcheck: Climate change is not making extreme cold more common. I continue to believe that heatwaves are the biggest current problem that climate change causes. Other problems like variable rainfall (floods and droughts), rising tides, and wildfires are easier to address. Attributing other extreme events is iffy, climate tipping points are unlikely, and the AMOC is not about to collapse. Also interesting is the fact that colder places are warming more due to polar amplification. This will tend to lower heating demand in high latitude regions, though cooling demand will still rise everywhere.
There are some interesting ways to address heatwaves. India is warming more slowly because of pollution from sulfur aerosols. Painting roofs white helps too. Large-scale irrigation can lower heat deaths in nearby cities. Between cities, there’s a substantial amount of adaptation to local climate, in ways that aren’t entirely understood.
I wonder if elements from Islamic architecture like thick walls, windcatchers, solar chimneys, fountains, evaporative cooling, and ground-source cooling might come into style in hot and dry regions. And then there’s a grab bag of other ideas like personal cooling devices, salt cooling vests, AC, trees, city-wide sunshades, and pneumatic canopies.
Related: Does Directed Innovation Mitigate Climate Damage? Evidence from U.S. Agriculture
Also related: adding thin gorilla glass panes between double pained windows could make them highly insulating, reducing heating needs in the winter. Perhaps making them reflective would help keep houses cool.
3.
Optimal Unemployment Insurance by Nicholas Decker covers a study about the effect of unemployment benefits on workers. Do they search more to find a better job, or do they take longer (or take more risky jobs) because they know they are insured? It finds that 60% of the increase in unemployment duration comes from searching for a better job, which makes everyone better off.
The study reminds me that many welfare programs can be justified on the basis of economic efficiency:
Baby bonuses, which pay for themselves in future tax revenue.
Subsidizing education increases productivity, a partial public good
Social security redistributes a child’s future income to their parents in a way that avoids bargaining problems.
Defense, police, firefighters, and domestic disaster relief can all be efficient forms of insurance1. Health insurance could become universal Genetic Insurance in the limit.
See this series by Trevor Chow for more2.
A state focused purely on economics would do exactly this amount of welfare and stop. The next step is to consider purely-altruistic redistribution to raise welfare. Altruism is a public good, so it makes sense to fund it collectively. However, I worry that voters would choose a level of redistribution that hampers growth, particularly since the money comes from inefficient taxes. Instead, a subsidy for a public goods funding mechanism focused on redistribution seems better.
Everything else
This short presentation by David Dalrymple on FlexHEG contains some remarkable claims such as “ … DARPA showed through the HACMS project that it is possible to create software systems that are completely free of exploitable bugs.” In the limit of capability, the defender “just wins” at the software level, especially with AI automation. He turns to the hardware level where things are much harder, especially if attackers have physical access3. I hope secure hardware is feasible, especially with FPGA’s, open source, inspectable hardware4, nuclear batteries, and multiparty computation. My hope is that AI will make designing, simulating, verifying, and inspecting secure hardware feasible so that everyone can be secure to basically any adversary.
Convolutional Differentiable Logic Gate Networks
Differentiable Logic Cellular Automata
Introducing Roche Sequencing by Expansion (SBX) Technology
Engineered commensals for targeted nose-to-brain drug delivery. Cool idea, though I’m not sure why you wouldn’t just administer a peptide solution.
A torpor-like state in mice slows blood epigenetic aging and prolongs healthspan. Interesting not only for the ability to induce hibernation, but the connection between sleep, aging, and epigenetics.
Endocisternal interfaces for minimally invasive neural stimulation and recording of the brain and spinal cord. Demonstrates a neural interface that passes up through the spine to reach various parts of the brain. It’s better than directly accessing the skull and avoids the explantation issues of stentrodes. I remain skeptical that BCI will offer dramatic improvements in computer interfaces, and noninvasive methods seem far more scalable, but I’m glad people are trying new techniques.
Extending women's fertility: the last frontier argues that much of the opportunity gap for women comes from a cultural and personal preference for women to care for their children. Fertility technology can move their childbearing years later, making it possible to build a career first before having children.
What is Earth's Carrying Capacity? Under various assumptions, the Earth can support hundreds of billions to hundreds of trillions of people. This is not to say that that is desirable, but to point out that “… especially with demographic trends as they are, the world will not face any genuine limit to growth under any plausible scenario and under any time frame that is relevant for making policy. The most rational course of action is to put such anxieties aside and focus on real problems facing the world today, none of which are overpopulation or overconsumption.”
Matt Yglesias points out that Misinformation mostly confuses your own side.
The paper Subversive Conversations is good. The general idea being that two players (or a group of players) can come to agreement about the state of the world and a course of action while under threat from an evesdropping party with different goals.
Digital Herd Immunity and COVID-19 shows how digital contact tracing could stop disease spread even with low vaccine uptake.
I like the looks of Liu Jiakun’s West Village in Chengdu. It combines a cyberpunk and solarpunk style and points at some of the stuff I discussed in my post about AV’s and cities.
I’m toying with the idea of generic disaster insurance. Basically you would pay a monthly fee to have the right to claim a (say) $100K loan at any time. If triggered, you would get the money immediately and pay back the loan according to the terms. That way you can always have cash in case of a major accident or surprise litigation.
Though I disagree with many of the proposals there, it’s getting at a similar idea. I think the focus on inequality and tax revenue over the welfare of the worst off is misguided, but that’s a topic for next time.
The rest of the talk is quite wonky. The specification and applications are interesting in their own right, but very niche.
More links on the Andrew “Bunny” Huang’s project: Open Source is Insufficient to Solve Trust Problems in Hardware, Infra-Red, In-Situ (IRIS) Inspection of Silicon, IRIS (Infra-Red, in situ) Project Updates, Can We Build Trustable Hardware?, BlueHat IL 2019 - Andrew "bunnie" Huang - Supply Chain Security: "If I were a Nation State...”.
What makes addressing drought more tractable than addressing heat waves in your mind?