when I quit

What will I do when I don't have a job?

• I don't feel that I have a clear direction. I want to learn and explore. There are lots of goals. But I also know that the price of everything really worth having is commitment. I need to be choiceful.
• There are many things I'd like to learn, pursue, or work on:
  • Personal skills: speaking and relating
    • can take an ultra speaking course, go on dates, etc
    • Participate actively on twitter
    • Engage with people online and try to build real relationships
    • Volunteer for Trevor Project to help gay kids.
  • Consciousness and meditation:
    • take a meditation class, do a retreat, do a lot of reading
    • make it my project to become 10x as happy, Nick Cammarata style
  • Writing
    • essays about life or the world
    • for myself, for a blog, for friends
    • or distill.pub style teaching articles
    • a book about models, fleshing out many models
    • Handbooks for building valuable skills. Write your own self-help book.
      • for example: https://www.julian.com/ on weightlifting, writing, etc
  • Travel and cultural perspective
    • Learn to speak Chinese
    • Travel southeast Asia
    • Travel China
    • Travel Africa (several times obvs)
  • Literacy and reading
    • History
    • Revisit the great philosophers
    • Modern thinkers. Work through my Goodreads list. Strauss. Rose of Paracelcus. Things Hidden since the foundations of the world. Other books that people say have changed their lives.
    • Great essays. inner ring. What else? essays to reread
    • Start a page of books and essays I'd recommend to a younger me.
  • AI work
    • what is the state of the field?
    • what are the implications of really believing that AI will work?
    • technical reading / work in RL
    • technical reading / work in language modeling / transformers / reasoning
    • technical reading / work in AI Safety
      • Catch up with where the field is now. What are people at OpenAI / Deepmind doing? Am I excited about this?
    • Non-technical connections between AI safety and how to be a good person. Popularize the theme that love is value alignment. (of course, part of being a good person is being an effective person---rationality is moral---which is the rest of AI but also very much not something you get by studying AI textbooks)
  • Science
    • Learn physics properly
    • Read more about ML in biology
  • Drugs, depression and psychedelics
    • understand the chemistry and neuroscience of drugs
    • write an explainer of depression. what I didn't know about it.
    • computational models of depression.
    • what do psychedelics teach us about ML.
  • Finance
    • I want to understand finance at the executive level.
    • I don't need to be a quant.
    • But: how does one build a company? What do you need to know? How do I think about structuring incentives?
  • Community
    • I can't / shouldn't do any of this alone. conspiracy is a thing for a reason
    • Find friends online.
    • Go to in-person meetings.
    • Hire a coach or tutor online.
    • Reach out to relevant experts. Ask serious questions. Build relationships.
  • Teaching
    • Teach an ML class
    • Teach about Bayesian modeling and probabilistic programming
    • Write a curriculum for a class
    • Find a friend(s) and alternate teaching each other.
  • Habits of self improvement
    • Regular journaling
    • "Reserve 15 minutes per week to send appreciation emails to anyone you really enjoyed chatting with that week."
    • Ask people generative questions.
    • Set aside time for prediction as a model-building exercise.
  • Value creation
    • There should be a way for me to create value (and believe I'm creating value) beyond writing code for Google.
    • The world is ultimately not that complicated and is rife with opportunities for improvement.
    • The point of all the other things is for me to figure this out.
    • I will eventually need another job. I'd prefer it not be an actual job.
    • I don't need to become rich. I can create value independently, and not capture it. If I become a Gwern or a Scott Alexander, that's fine.
    • I don't want to pretend. Running a business would keep me honest. I'd get an external metric to optimize.
  • All the usual personal stuff
    • Fitness
    • Existing friendships
    • Music
    • Family
• Now I know that I can't even hope to accomplish everything on this list. And I definitely can't wake up every day and choose a new thing to push on. I need choices and structure and focus.
• I could be college-like. I could choose 4-5 things at a time and alternate between them. That gives me a balance.
• I think it might be better to do one thing at a time, deeply. If I'm going to work on finance, I should do a deep dive. I can give myself a fixed time period. Maybe I do a one-week or two-week deep dive on each of a few subjects. The goal is to write something. At the end I have a series of posts.
• Discipline is useful but by itself it won't take me anywhere. I need genuine curiosity. I might not have genuine curiosity for everything I listed above, and that's okay. But delving into something might help me become curious.
• I want and deserve an exploratory phase. I want to know that I'm not trapped. I want to allow thoughts to marinate, and to allow time to figure out my real goals and improve my process.
• But I don't want to explore forever. Or rather---there are some things you can only learn by going deep. So I will need to commit myself to a longer project.
• Ultimately, the goal is to get excited about something. I want to be able to say to myself and others that I'm doing what I'm doing because I believe in it. I want to have a short-term plan and a long-term vision.
• Don't forget to find time to play. Doing something without caring where it leads can sometimes be the path towards something interesting. This PG essay is one take on this: http://paulgraham.com/own.html