Some rough notes from the first day of class in MIT’s How to Grow Almost Anything on changing the world!

Growing up, I think computers were kind of amazing because you could affect immediate change with an objective truth: the compiler. I found the computer as a quick way to create and bring things into the world, so that became my primary outlet. What are the other ways you can affect change in the world?

NOTE: These categories don’t have strict lines between them. Interpolating them is an open area of research.

Types of Change

Bits

Software is eating the world.

(I still don’t fully understand what this quote means), BUT it’s true. I don’t think people should learn computer science because it can be a high paying job or because it’s an entirely new way of thinking but simply because it gives you a broad array of ways to solve issues. Algorithmic thinking actually does change how you view the world and I think there’s something magical about being able to quickly spin up websites for friends or to bring ideas to life.

Atoms

The extent of my experience here is woodworking, but if you want to change the world we all inhabit, then you’re going to have to touch atoms at some point. Engineering and the more hands on fields all fall under this branch to me. If you do computer engineering, then your job is usually translating bits in some way to affect atoms.

Base pairs

Genetic engineering, gene drives, and getting biological systems to do your bidding. Using evolution to grow things, CRIPSR, and biological circuits.

Governance/Law

Society runs on norms and agreed upon conventions. Figuring out what of those is most productive is an extremely high leverage way to change the world. It’s kind of interesting how although security protocols and agreed upon ways of acting are perhaps low status things to try and change, but they are the ocean that we’re immersed in. Some vague things that inspire me are SafeTensors in AI and how biosecurity is handled and talked about at large conventions.

While this is perhaps less exciting and flashy than doing deep technical work, it’s the work that has to be done to make sure we don’t actually destroy ourselves with technology and progress.

Politics

I have no clue how anything gets done in politics. Is there a lot of room for improvement? Maybe? I’d love to talk to someone about this and learn more. Why is U.S. politics completely broken?

Education

If you have knowledge, it should be shared! I think people underestimate how much even sharing little bits and pieces can inspire others. I have a softly held belief that all education should either be quite authoritative (i.e. here are the axioms that make up set theory) or very loose (here’s my half explored research idea), and anything in the middle is bad. Authoritative sources are good to polish and make it so you can bring someone up to speed as quickly as possible. Scrappy education is good to expand someone’s imagination and get them to make things themselves. Nothing is quite as motivating as being handed half a thing and realizing that you’ll have to bring the thing to life yourself.

It’s okay to oscillate between the two at a regular cadence, but if each section if confused as to whether it’s fully polished or perfectly incomplete, I think it’s just poor educational material. If you have some other teaching objective that is orthogonal to this fake axis that I’ve presented, I’d love to hear about it!

Magic hammer

Everything is a nail when your hammer is magic.

In a way, you can view ChatGPT and the push of modern autoregressive models as a universal translation between language and all the other forms of change. You can write a program, design a drug, and change minds all through {insert your favorite human language}. If we get cool robots, then you’ll also have your commands change atoms.

Although I’ve presented quite a few ways I think you can change the world, my biased view is that bits are the most powerful on the list currently. AI in education (MathAcademy, KhanAcademy, etc.), protein-folding with machine learning, Polis to create more efficient ways of governance. Everywhere I look, I can see how bits are being used to change the world in all the other ways mentioned. Can you translate governance to software through incentives? Probably, I just can’t think of many good examples. (If anyone has fun examples going different ways or other ways to move the world, let me know!)