Opportunities
News
A background of gravitational waves pervades the cosmos (via @QuantaMagazine)
The Terraformer makes cheap natural gas from sunlight & air (via @TerraformIndies)
Obituaries
Henry Petroski, “America’s poet laureate of technology,” author of To Engineer is Human, The Evolution of Useful Things, and other books
John Goodenough, age 100 (!) “His discovery led to the wireless revolution and put electronic devices in the hands of people worldwide”
Links
Great SpaceX video about iterating towards reusable rockets (via @pronounced_kyle)
Scott Alexander on the “illusion of moral decline” (but see Matt Clancy’s pushback)
Three auto maintenance philosophies, from @stewartbrand’s book in progress
As the opportunity cost of time increases, so does our concern about wasting time
The “undercurrent of tacit knowledge” in any scientific literature (by Tyler Cowen)
GLP-1 drugs are amazing for weight loss, says Derek Lowe (via @curiouswavefn)
Why you can ignore the “aspartame causes cancer” headlines (by @StuartJRitchie). Related, the IARC puts bacon in the same category as plutonium
Queries
What use cases is hydrogen best for? And what are the front-runners for long-term energy storage?
What are the best films about capitalism, business, and market economics?
Why does DNA use base 4 rather than base 2 for encoding information?
How to roll back as much bad regulation as possible in one tweet?
Any indicators of misuse threat of adversarially fine-tune-able 10^25 LLMs?
What intellectuals refocused New Left types towards remaking the physical world?
Quotes
Real history is not politics but the story of those who make a better life
Academics and nonprofits act as if someone else will translate their papers into policy
Tweets & threads
A biodefense roadmap from Kevin Esvelt. Excellent work
Chicago went all in on railroads and Cincinnati went all in on canals
A path to prosperity for small cities within one hour of a major city/airport
Update from prediction market Kalshi on regulated election markets
I push back against Dustin Moskowitz’s list of AI safety “strawmen”
Many good ideas are killed for reasons unrelated to their promise
As we make progress on social problems, we broaden the definition of the problem. A case of the Shirky Principle
In 1844, the Secretary of the Navy was killed during a military demonstration
Most attempts to disrupt education do project-based learning, which is overrated
SF can’t even find actual collisions to cite against self-driving cars
“We can’t answer every question.” “No, but I think we can answer any question.”
I like Casey Handmer, but...
US natural gas today: $5/MCF
CO2 from air: >$8/MCF
H2 from electrolysis: >$37/MCF
As those posts say, biomass conversion and electrochemical reforming of methane seem much more economically practical.
FYI, Terraform Industries published a new blog post about the cost of their hydrogen electrolysis, which is significantly less than you’re linked estimate of around $7/kg. They claim they can get $3/kg using current legacy solar, but with current solar cost trends continuing (which I know you’re skeptical of) could reach $1.12/kg. They accomplish this by optimizing the cost/efficiency trade-off of the electrolysis setup.
I’m assuming Casey Handmer is being overly optimistic, but I can’t tell by how much.
I don’t think Terraform can produce 50% efficient durable electrolyzers that cheaply. So, who’s right?
You’ve seen my blog. I linked to a post that linked to some papers on electrolysis costs.
Casey also has a blog. Here’s the background of their initial team:
So, how would you go about evaluating who’s right, here?
I have a suspicion about why DNA is base 4 vs base 2.
It has to do with the binding strengths of tRNA to mRNA during translation. Base 2 increases the length of codons needed to code for 23 amino acids from 3 to 5, which may lead to less fidelity in the resulting amino acid sequence. That is, partial matches have higher binding strength compared to base 4, which increases the chances the wrong tRNA is loaded. This could also make longer amino acid sequences harder to create—an accidentally loaded stop codon terminates translation!