Sponges, Sensors, and the True Cost of Green Tech

The sensors section of the newsletter is interesting but the model won't see reality, it will see the closest, most plausible pattern it was trained on. It's the physical sciences equivalent of an LLM hallucination.

Love it

Today’s newsletter covers:

The Nobel prize winning ‘molecular sponges’ and their surprising uses in biosecurity.

A 200% safer AI-designed car bumper that may never materialise.

Virtual sensors for fusion discoveries and a tough dilemma for future science.

Chinas tariffs and the dirty secret of the rare-earth metals powering our green future.

I feel we need a name for this weird practice of pretending a policy was a climate change policy and then attacking it for that imagined failure:

> Well, the catch is that mining the most popular rare-earth metal, neodymium, emits miles more CO2 than any other metal (which is ironic because they’re a key ingredient to most green technologies). So, it seems Western countries are choosing to outsource this pollution to quietly keep their hands clean of carbon. This policy naturally snowballed into China having a significant market dominance, allowing it to produce even less common rare-earth elements at an unbeatable cost.

Is there any evidence at all that rare earth mining and processing is concentrated in China because the West was trying so hard to be low carbon that rather than, I dunno, build well insulated homes or smaller lighter cars, they took a hard line stance on neodymium processing?