How to:

Optimizing for standardization

What do we optimize for?

It's kind of crazy how much compute we'd have if we just bothered to optimize software instead of relying upon endless dependencies. The Apollo mission was run with less compute than a smart phone. But we don't-—why? Well for one we don't need to.

But there's a bigger reason:

The same reason that if we are going to build a car, we don't innovate the screws. Standardization-and the abstraction it enables—is the most underrated invention of the modern world. It's also crazy how much inefficiency there would be if we had to reinvent the wheel every time. The idea that I can reliably get a precisely-made and yet mass-produced machined screw of a standard size anywhere across the world is one of the most radical technologies to ever be developed.

See, we have optimized our software, as we have our cars and the assembly lines used to build them. But not for size or efficiency but for interoperability and genericism.

Ideas

Optimizing for standardization

It's kind of crazy how much compute we'd have if we just bothered to optimize software instead of relying upon endless dependencies. The Apollo mission was run with less compute than a smart phone. But we don't-—why? Well for one we don't need to.

But there's a bigger reason:

The same reason that if we are going to build a car, we don't innovate the screws. Standardization-and the abstraction it enables—is the most underrated invention of the modern world. It's also crazy how much inefficiency there would be if we had to reinvent the wheel every time. The idea that I can reliably get a precisely-made and yet mass-produced machined screw of a standard size anywhere across the world is one of the most radical technologies to ever be developed.

See, we have optimized our software, as we have our cars and the assembly lines used to build them. But not for size or efficiency but for interoperability and genericism.

Updated continuously • Last edited on
11.24.23
Ideas

Optimizing for standardization

Updated continuously •
Last edited on
11.24.23

It's kind of crazy how much compute we'd have if we just bothered to optimize software instead of relying upon endless dependencies. The Apollo mission was run with less compute than a smart phone. But we don't-—why? Well for one we don't need to.

But there's a bigger reason:

The same reason that if we are going to build a car, we don't innovate the screws. Standardization-and the abstraction it enables—is the most underrated invention of the modern world. It's also crazy how much inefficiency there would be if we had to reinvent the wheel every time. The idea that I can reliably get a precisely-made and yet mass-produced machined screw of a standard size anywhere across the world is one of the most radical technologies to ever be developed.

See, we have optimized our software, as we have our cars and the assembly lines used to build them. But not for size or efficiency but for interoperability and genericism.

Read more