Adjusting 7/16/23-9/24/23 Work Goals

Current Work Goals

Spend ≥ min(70 hours, COMPLETION) on ”Mathematics for Machine Learning”.
Spend ≥ min(70 hours, COMPLETION) on “Deep Learning for Coders”.
Publish ≥ 70 Twitter threads.
High-profile account engage on Twitter ≥ 210 times.
Make ≥ 2 contributions to GGML repos.
Make ≥ 2 contributions to Hugging Face repos.
Make ≥ 1 contribution to LangChain repos
Make ≥ 1 contribution to swc repo.
Make ≥ 1 contribution to bun repo.

New Perspectives (7/23/23)

Current Reasoning

Given:

  • I have the contextual knowledge and cognitive skills required to technically operate in ML at various levels of the stack.
  • LLM breakthroughs have made ML hot right now.

Strategy:

  • Develop reputation in ML by writing ML content and contributing to open-source ML projects.
  • Develop skills via daily study.
  • Leverage reputation/skills to build community in ML.

Desired Result:

  • Improved perspective and community access either:
    • Exposure to an existing opportunity e.g. via cofounder introduction.
    • Allows me to identify people with a problem.

Good Things

  1. These goals are ambitious.
  2. Building reputation via writing is a high-quality strategy.

Bad Things

  1. In the context of starting my own company, improving my skills without a defined problem is a waste of my time.
  2. Problem identification is passive and secondary.
  3. Restriction to ML.

Changes

  1. Replace all undirected study with problem hunting. Problem hunting is fundamentally communicative:
    1. Defining a community.
    2. Creating a hypothesis about that community’s problems.
    3. Testing this hypothesis via meaningful engagement with members of that community.
  2. Eliminate explicit open-source contributions. I may still elect to do so, but in the context of problem hunting and reputation building.

New Work Goals

Spend 126 hours on problem hunting.
Publish ≥ 70 Twitter threads.
High-profile account engage on Twitter ≥ 210 times.

Musings

Founding as a Habit

It is better to think about starting my company as a behavioral change than as an action. I need to align my cues, cravings, responses, and rewards such that they conspire for my success.

Writing

  • Global Reiteration is Fine - Do not be concerned with duplicating another source’s information.
  • Abstraction is Required - Writing something self-contained and consumable requires drawing information boundaries.

Dating the Algorithm

Aphoristically - your audience is a reflection of what you say.

Inversely - you join the audience of what you choose to hear.

A neutral recommendation algorithm is fundamentally a mirror. Life is better, when you like what you see in the mirror.

Meditation

The wide range of thought encouraged by my current situation has made my mind noisier during sitting meditation. Practice diligently to investigate than nature of this change.