← Back to Kevin's newslettersPublished: 2021 September 12

Decarbonization follow-ups

Last month’s newsletter on decarbonization was the first time I’d researched climate and energy. The conclusion (in-so-much as a set of notes / loose book review can have one), was more “hands gesturing vaguely towards nuclear energy” than some specific call to action. I’ve been digging further to find a Kevin-sized opening into the space, but so far have come up empty.

On the research/policy side, I’ve found some additional color:

but while these points seem plausible, they don’t point towards any angle I have leverage on.

On the technical side, I spoke with two of Seattle’s fusion startups about whether my computerist skillset and experience building, e.g., wind farm operations software would be useful in their efforts. However, both were in the “we’re building our next prototype, do you know any graphite anode experts?” physics/engineering development stage rather than the “can you build a software team to improve our duct-taped spreadsheets-in-Dropbox process?” operations stage.

So, until I find some angle-of-attack, I’m tabling the energy supply carbon footprint problem top the back of my mind.

Microfluidics / lab-on-the-cheap

There is one topic that’s been stuck in the front of my mind for a while: The intersection of “maker” manufacturing and synthetic biology. Both have had dramatic capability advances and cost reductions over the past decade.

The same dynamics that let me goof around with luxury mechanical keyboards while I was in Taiwan apply just as much to lab equipment. There’s an emerging Lab on the Cheap grad student cottage industry improvising centrifuges, PCR thermocyclers, microscopes, etc. using 3d printers and repurposed consumer electronics hardware.

The first-order take — why buy a $50k qPCR thermocycler when you can buy a $5k one from Kickstarter — is (presumably) compelling for a sufficiently cash-constrained researcher. See also:

I get why academics / educators play here (anything cheap, off-grid, third world clinic wins TED talk bingo), but I don’t find it personally compelling (beyond the excuse to mess with industrial design / small run manufacturing, of course). The dent such projects will make in traditional markets will probably be similar to the one Desktop Linux has made in Apple or Microsoft’s bottom lines.

The more interesting question is: What will such cost decreases make possible? Postdocs who are (understandably) afraid to modify a $100k shared-across-labs instrument have no trouble hacking on a “low road”, $100 device. A lot more shit has been (literally) shoved into Raspberry Pis than MacBooks.

Cheap equipment also unlocks scale: Google famously built their first production server from commodity parts rather than buying enterprise-grade servers from IBM.

Instead of buying a single $50k precision machine, what can you do with $50k worth of less-precise PCBs and 3d-printer resin? I.e., why do one experiment with a 95% reliable pipeline instead of 1000 experiments with a 70% reliable pipeline for the same cost and turnaround time?

Now that my partner and I have decided to stay in Seattle through the winter (and retrieve my tools from storage!), I’ll start exploring these broad trends within the context of few projects.

I’ll likely start by replicating this continuous-flow, PCB-based microfluidic PCR thermocycler, as the technical skills are familiar and the necessary heating/cooling/optical testing won’t require a wet biology lab.

Not that I have any particular use for PCR myself or feel like there’s a huge demand for $100 microfluidic thermocyclers. (Though that’d be a pleasant surprise.)

Rather, my hope is that working on a concrete problem will familiarize me with strengths and tradeoffs of the microfluidic approach and serve as a foundation for conversations with molecular biologists about how such nascent technologies might be applied to their research.

If you’re day-dreaming about unattainable lab equipment or are itching to collaborate on electrical engineering or 3d-printed mechatronics, drop me a line! (Especially if you are in the Seattle area!)

Misc. stuff