LLMs for science, your driveway is a battery, filtering plastic from the water
3 reasons to be optimistic.
Welcome back Nerds. It’s been a minute.
I’d like to say I was busy, mostly just procrastinating and letting other things get in the way.
3 reasons to be optimistic:
Most of the time, problems are easy to identify.
At least they are easier to define than solutions for big hairy problems like climate, pollution, healthcare, or what the future will look like.
They are easier to define because they already exist.
But there are reasons to optimistic. Here are three reasons to be hopeful about the future. Also, these are seeds for a future you can imagine, or better yet create.
Ex-Google CEO Eric Schmidt is launching a nonprofit to bring focused on bringing chemists and scientists together with those with deep expertise in AI. The goal is to fundamental transform how science is done.
Samuel Rodriques, founder of the Applied Biotechnology Laboratory at the Francis Crick Institute; and Andrew White, a University of Rochester professor and a pioneer in the use of artificial intelligence in chemistry, will spearhead the effort.
Why It’s Important: To have a real world impact, LLMs have to move beyond a productivity hack. And Silicon Valley needs to move beyond ad driven models and the creator economy. Combined with IBM’s work with NASA & Hugging Face, and BIO-GPT out of Microsoft, and this gives us three strong examples of work being done to put the next generation of AI tools to better use.
Scientists at the University of British Columbia’s BioProducts Institute found that if you mix tannins and wood dust you can create a filter that traps 99.9% of all microplastic particles present in water. In mouse models, the filter prevented the build up of microplastics in organs.
Why It’s Important: Plastics are everywhere. virtually all tap water is contaminated by microplastics, and we’re now now finding plastics embedded in the heart. The problem is real, and more than just an ugly mess in the ocean. Who knows what the long term effects will be. But the filter created here is a great illustration of something which is economical, cheap, and biodegradable ready to be scaled up at municipal treatment facilities.
Your house foundation could serve as one giant battery. Researchers from MIT’s Wyss Institute for Biologically Inspired Engineering have put together a supercapacitor made from two of humanity’s most ubiquitous materials: cement and carbon black. After mixing carbon black into a cement powder and adding water, the carbon creates a fractal like tree structure within the cement as it curres. Soaking the material in an electrolyte like potassium chloride causes charged particules to accumulate on the carbon. Two pieces of this together with an insulating layer or empty space between creates a powerful supercapacitor. A 45 cubic meter block of this doped concrete would have enough capacity for 10kwh of energy, about the daily electricity usage of a household.
Why It’s Important: Although it is CO2 intensive to make, concrete is fairly ubiquitous. This could be a relatively inexpensive way to put foundations to better use, maybe in new home construction, but also in roads and larger facilities attached to the grid. I’m also willing to bet that this is less CO2 and resource intensive than mining and refining lithium and other metals for traditional batteries.
What I’m reading / watching:
Insightful video about the insights that can come from a thinking and looking at systems holistically. Paradoxically, when you look at the big picture, it’s easier to simplify.
Geeking out on mushrooms:
Two must watch, both from the godfather Paul Stamets:
And one that was a fascinating shorter overview of the history of how various civilizations have used fungi, and what we might use them for today.
Ideas in the hopper (feedback appreciated):
I’ve been thinking and reading about complexity of late. From
I have a renewed interest in articulating what regulatory nextworks need to look like in the next paradigm. Mainstream / common denominator mechanisms work well in times of stability. But when things are fluid and move rapidly, those systems fail. But that doesn’t mean we can move without regulation. Think Gamestop, and Coinbase and FTX. What’s a fluid approach here that works?One of the core hypothese of this newsletter is that it’ss more critical that ever to develop your curiosity muscle and think across disciplines if you want to become more resilient and help solve big problems, possibly make a living in the process. I’m looking to create a system or set of questions for this, to put in the reps to develop a more open mind and ask better questions. Can it be as simple as forcing yourself to draw connections between topics (nodes)? This reminds me of Herman Hesse’s The Glass Bead Game.
The fringe is where we find the seeds of the future or future possible. Can we leverage a GPT to monitor the future?
If you’ve made it this far, I’d love to hear your feedback.
Was this great? So so? What would you change? All comments welcome.