Diagrammatic loops: Startup tempo reduction paradigms

[New? Start here: https://meditationstuff.wordpress.com/articles/]

[This is a guest post.]

Warning: 1st draft, half-baked ideas, poor yEds, proceed at your own peril
There are two life-or-death domains in which prediction is impossible, the actor is part of the environment and pre-set goals don’t exist. One is the startup environment. The other one is your life. I suspect that understanding startups leads to understanding your life (and when I say “understand” I actually mean: deliberately craft otherwise impossible, desirable realities). In this post I will lay out the current state of some of my thoughts about startups.

 

Winning

“It’s somewhat the same thing with instrumental rationality or any sort of OODA-loop, you want to be running on an internal cadence, faster than reality, not driven by reality.” 

This quote sensitized me to Venkatesh Rao’s book “Tempo”. I didn’t appreciate the book too much, but it led me to the OODA-loop and to Tempo and for that I am grateful.

The basic idea of tempo, and how it relates to winning (in non adversarial setting) is that you need to reduce your tempo, that is, your decision-making cycle time needs to get shorter and shorter. Imagine a decision-making cycle that goes like this “Desired World State -> Action -> New World State -> Compare to Desired World State -> If same, over; else, go back to “Action””. The faster you can iterate through this cycle, the faster you will reach your Desired World State, your goals. That is tempo reduction.

The genius in it is that actually adding things to the cycle can make the overall cycle be much faster: it can make it “faster than reality”.

 

Startup paradigms

In startups there are 2 ways to do this, here illustrateD with a sole agent (a startup can be modelled as an agent). One, focused on acting, highlights future unpredictability and thus wants to  augment certainty through various measures or contacts with reality (fail fast. think do get feedback, iterate). The other one focuses on fixing the quality of your model of the world/yourself/paths to goals/goals by thinking from first principles.

 

Abstract Models

Basic Model

As an illustrative model (from AI), you can see a startup as an Agent that has beliefs (about the world state), desires (of how the world state ought to be) and intentions (how to bring those about).

0 – BDI: Act, feedback

Basic Idea:

  1. Agent exists in World
  2. Agent desires Desired World State
  3. Agent Acts
  4. Agent compares Actual World State to Desired World State
  5. If equal then end, else back to 2.

one feedback

THINKING APPROACH

In this section I summarise increasingly complex (and with increasingly short tempo) feedback loops.

Key: feedback loops that are shorter in tempo allow more cycle iterations and therefore allow you to get more done in less time.

  1. Agent desires World State
  2. Agent imagines what particular Act would bring
  3. If particular Act would bring Desired World State then act, else, back to 2.
  4. Agent compares Actual World State to Desired World State
  5. If equal then end, else back to 2.

 

1- Think, feedback, Act, feedback

2 feedback

2- Think, feedback, Think about thinking, feedback, Act, feedback

3 feedback

This seems to be what is going on in industry (except you need to add competitors and all the interactions those open)

 

3- Think, feedback, Think about thinking, Act, feedback, think, Act, feedback

4 feedback

Science seems to follow this model (distributed over many institutions, groups, individuals)

 

ACTING APPROACH

In this section I summarise increasingly fast (and with increasingly short tempo) feedback loops.

1- Think, feedback, Act, Feedback

2 act

  • Think about how to get feedback as fast as possible.
  • Figure out better measurement/feedback methods. That is, if you want to know if A is the case it might be better to 1) use tool X instead of tool Z, 2) figure out thing A’ which is entailed by thing A which is way easier to measure.

 

2 – Think, feedback, Act faster, Feedback

one act

Act faster:

Note that there seem to be limited tricks you can use to get feedback faster : One is to act faster, to get a short tempo (Not sure how to design this). The other one is to get to indicators that are easy to measure that measure something you care about which is hard to measure.

 

Back to Reality

The Lean Startup Methodology is the epitome of faster. You have an idea, build it as fast as possible, measure as fast as possible, and learn as fast as possible. On learning faster they actually mean better: getting more information (uncertainty reduction) of the same data. The other ones do focus on going faster.

lean-startup

Philosophy of Startups

Notice the promise of better thinking: you don’t need self-sacrifice (which probably fails to be sustainable [might explain high burn out and failure rate] if you can just outsmart everyone by constantly going levels up.

Notice the word “Methodology”. Methodologies involve paradigms which entail epistemologies. The current lean startup be fast vs 0 to 1 think from first principles is a class instance of the empiricism-rationalism (knowledge from experience vs knowledge from thinking) debate in philosophy. (If this seems out there do remember that Peter Thiel studied philosophy and advocates thinking from first principles.)

I’m not entirely sure where I wanted to go with all this (or rather, where all this wanted to take me/us to). The model is generative: It’s value is simplifying reality to generate ideas.

Future Material To be added
  • Startups live in adversarial environment, you do not (to some extent)
  • I think that fragile, robust and anti fragile (NNT) have something else to add here.
  • Tempo explains globalisation: steal and save R&D time, reverse engineer first principles and build from there (China). This can probably be used for individuals.
  • I suspect learning effectuation thinking might be the most meta (and thus multiplicative) possible hack. (Mark seems to disagree on this)
  • Get deeper into 0 to 1
  • Get deeper into lean startup methodology
 
 [CLICK to SUPPORT this blog and BUY utterly unique health, fitness, SEX, intimacy, communication stuff.]
Advertisements

One thought on “Diagrammatic loops: Startup tempo reduction paradigms

  1. (tempo = frequency = 1/period, so tempo should increase, period should decrease)

    “Think about how to get feedback as fast as possible.” Yup.
    “The other one is to get to indicators that are easy to measure that measure something you care about which is hard to measure.” Yup.

    You might like Rao’s post about “Product-Driven” versus “Customer-Driven.” He criticizes the lean startup movement, and his thinking about tapping anomie really resonates with me:

    http://www.ribbonfarm.com/2014/04/24/product-driven-versus-customer-driven/

    This book is fluffy and too expensive, but it has some nice ideas:

    http://www.amazon.com/Managing-Unknown-Approach-Uncertainty-Projects/dp/0471693057

    From the promo bullets:

    * Trial-and-Error Learning allows for redefining the plan and the project as the project unfolds
    * Selectionism pursues multiple, independent trials in order to pick the best one at the end

    The best book I know of about rapid prototyping talks about it in just one section:

    http://www.amazon.com/The-Art-Game-Design-lenses/dp/0123694965/

    There are some very creative ideas about getting fast feedback, in the domain of software games, to be sure, but it was very stimulating.

    Effectuation:

    Environmental isotropy – everything looks mostly the same in every direction, don’t know what in the environment is important
    Goal ambiguity – hard to pin down exactly what you want
    Enaction – your actions change the environment and your goals
    Exaptation – “a trait can evolve because it served one particular function, but subsequently it may come to serve another”
    (and a few more principles)

    My last post was mainly about dealing with environmental isotropy and goal ambiguity, in effectuation terms.

    You distinction between your “basic model” and “thinking approach” remind me of the distinction between “tracking” and “registration” in Brian Cantwell Smith’s On the Origin Of Objects:

    http://www.amazon.com/On-Origin-Objects-Bradford-Books/dp/0262692090/

    With regards to your models, it strikes me that an actor can act without any expectation of reaching the end state in that particular action step.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s