Agency

As I embark on this journey to develop more agency, I thought I’d write down some thoughts on what the concept means to me. At a meta level, the process of writing down my thoughts is itself part of demonstrating agency, so this exercise is productive even if I don’t refer back to these notes.

First – what is agency? Per Neel Nanda, Agency is doing what is needed to achieve your goals.

That is helpful, but it’d be even more helpful to more fully spec out a model for what agency looks like in practice. That would give a richer sense of what agency is than that that pithy definition.

Here’s what agency looks like in practice:

• Setting a hyperprior that you are capable of generating plausible plans, executing on them, correcting their flaws, and beginning again, and that this process will eventually yield successful results
• Realizing humans underestimate low probability events (per Nassim Taleb) and overestimate high probability events
  • In response to this bias, one upgrades their estimation of “low chance” plans and downgrades their estimation of “sure things”
  • Especially important heuristic for complex, difficult problems
  • Implicitly acknowledges that life is more chaotic than it seems
• Making plans to solving problems and achieving goals
  • Plans need to merely be plausible (“this could work”) and executable, no more
  • Asking the question – “What’s the stupidest and easiest one thing you could do to make even a little progress?” (h/t Emmett Shear)
  • Making multiple plans is good, even “bad” plans (“the dumbest plans that could possibly work” per Emmett Shear)
  • When unexpected problems pop up, making deterministic plans to address them rather than fretting over the fact of the problem or idly ruminating on them or their associated difficulties, required effort, sources, etc
  • Plans should include a detailed problem statement, incorporate your goals, and anticipate likely obstacles
  • Avoiding performative activity without any associated theory or story for why these behaviors will lead to success (full connection between current actions, plans, goals, and states of the world)
• Sensitivity to the things one values or truly enjoys, rather than what one is “supposed” to like or seek
• Taking action without regard to the probability of success
  • Being overly sensitive to success probabilities opens up the risk of “psyching yourself out” and reduces action
• When it’s cheap, taking quick action to rectify issues or optimize situations to your liking (per Tweets From Kat Woods)
  • Related: avoiding status quo bias, or not letting the way things currently are be a lazy justification for accepting current conditions and allowing them to persist
• Dynamically correcting plans and actions in response to real world feedback
  • This should be seen as an inevitable, even joyous part of the process rather than a failure or burden
• Being willing to fail
  • Being too tied to success both curtails action and cuts the odds of success
• Once plans are made, asking for help
  • Assistance from others can help supercharge one’s plans, giving you a leg up right from the star
  • Making lists of people who can either help you directly, or can refer you to people who can help
  • Being shameless in asking for help and making it clear you want help
  • Asking for help without regard to the costs you could be imposing on the help-giver

Having decided what agency is, it’s helpful to turn this into a schematic for what A system for agency could look like:

Transclude of A-system-for-agency.canvas

Next, having defined agency both in words and schematically, one needs a system for operationalizing agency in daily life. An easy place to start is an “agency tracker” – a system for tracking agentic behavior on a daily basis. The goal is to fill up the tracker as much as possible each day, taking every opportunity to exhibit agentic behavior and nothing such instances, similarly to how I track “Wins” in Obsidian.

---

References

Seven ways to become unstoppably agentic

https://twitter.com/eshear/status/1740059627788915050

How to Develop Your Sense of Agency

Transclude of How-to-Develop-Your-Sense-of-Agency#45100d

How to increase your agency: a flowchart

Agency: accepting the world, noticing paths to your goals, noticing what your goals are

• You have a problem: are you working on it?
  • No: Are you in a victim mindset?
    • Yes: Can you answer these questions?
      • What if it were possible?
      • What’s the stupidest, easiest thing you could do to make even a little bit of progress?
      • Why are you so sure you won’t succeed?
    • No: Can you answer these questions?
      • (now) What are you doing right now?
      • (goal) What is a detailed description of the world after you’ve succeeded? What does success look like? What are you actually trying to do? In 12 months, what would you like to be celebrating with a friend?
      • (why) Why do you want to do that?
      • (problem) What is the roadblock? What is the problem in detail? What might make you procrastinate?
      • (solutions) What are some ideas that could possibly work?
      • (plan) Given the best, easiest, or most liked idea, what’s your rough draft of a plan?
      • (next step) What’s the immediate next step?
      • (help) Who or what could help fill in the gaps?
      • (environment) How could you set up your social context and environment to
        • bolster your motivation, and
        • avoid frustrations and temptations?
  • Yes: Can you answer these questions?
    • What is a detailed description of the world after you’ve succeeded?
    • What’s the connection between what you’re doing now, your plan, and the goal?

Remember:

people tend to

• underestimate the likelihood of success for “bad” plans that could work
• overestimate the likelihood of success for sure-thing plans (day-to-day is usually more chaotic than we expect)

Free yourself from the requirement that your ideas must be good. All that matters is that they’re possible. It could work.

ewrite