The Secret QKP Master Plan

The main limiting factor in human progress is the number of knowledgeable people. That's the base rate that scientific progress gets multiplied by. All else being equal, a society with twice as many knowledgeable people will progress about twice as fast.

The process that generates knowledgeable people is learning. Unfortunately, the modern system for learning (high school, university, etc), is expensive, inaccessible, and dubiously effective.

Historically, many of the greats ended up teaching themselves. Einstein and Newton were both self-taught; more recently, Elon famously taught himself rocket science. Making efficient systems to help people teach themselves seems like the best way to put a kink in this curve of human knowledge.

But self-teaching is hard. Lots of techniques that seem good at first turn out to be largely useless. Two people can put the same amount of time into a subject and learn wildly different amounts based on their study methodology. It's only after you know something that you can see the best way to learn it. People lose motivation, get lost, can't get started, learn the wrong thing.

The goal of the Quantized Knowledge Project is to do for self-learning what the Model T did for cars and what apt did for installing linux packages: make it predictable, uniform, and efficient.

Here's the plan (just between you and me).

Phase 1

1. Create and popularize a standard format for quantized knowledge. Each card must be able to represent an arbitrarily complex piece of knowledge, like "how to play this chess endgame" or "how to write this proof" or "how to use GDB to debug this process." It should be easy to generate and parse, and should integrate well with git.
2. Use this format and make flagship decks for three fields (Poker, Japanese, Rust).
3. Recruit a small cohort of learners for each. Talk with all of them. Make sure their incentives are aligned. Run them through the decks over the course of a month or two, and make sure that they do indeed learn uniformly. Find a fair way to assess them at the end and publicly report the results. Which ones achieved their stated goals?

Phase 2

1. Write a lot about how to quantize knowledge effectively. Recruit others to do the same. Try to disperse the skill of making good flashcards into the broader community.
2. Work with the community to build high quality decks for skills which don't require expert input.
3. Figure out how to make money, perhaps by charging for advanced spacing optimization. Use that money to start commissioning content.

Phase 3

1. Carve out a profession: knowledge quantizer. Employ textbook authors, professors, and professionals to make decks as a first class medium for knowledge distribution.
2. Curate a very small number of very high quality decks in each area. (Think like how we have gcc and clang, or nginx and apache.)
3. Replace universities as the gold standard of knowledge acquisition, and dramatically increase humanity's efficiency at learning.