Learning, AI & Waste

Today (17/4/25), I'm on a flight to go on a fishing trip with my dad in Weipa.

I've been eavesdropping on a conversation largely dominated by a school teacher. Their discussion was about learning, and I'm going to steal some ideas.

Implicit Skills in Learning

• The practice of writing is more generally a practice in ordering thoughts.
• The practice of solving math problems is more generally a practice in utilising rational (as opposed to emotional) justification in problem solving.

On the lumper vs splitter spectrum I find myself prone to splitting — to finding and attempting to reconcile all of the dimensions I can. I find myself compelled down rabbit warrens so deep that the units become inconsequential to and unrepresentative of the bigger picture. I'm prone to over-optimisation — to neglecting the cost of optimisation.

Waste and Essentialism

I've often quipped that "more is more". I fear that through habit, it's turned from irony to identity. Largely inspired by the central idea of Greg McKeown's Essentialism, I'd like to practice careful subtraction — to sometimes sacrifice a part, an idea, an opportunity, a thread of exploration, the last bite of food when I'm full (and probably an item from this list) — in the service of a larger system.

Waste has always been an enemy, but in reality it is something to strategically embrace. Especially in an age of abundance. Perhaps it is better to view waste as an opportunity rather than an evil.

In optimisation, we reduce waste and become more efficient. However, in situations where short-term survival is uncertain, optimisation is a luxury that can't be afforded.

Scarcity vs Abundance

My belief is that waste becomes evil, a righteous point of focus, in scarcity. Scarcity is multidimensional, seasonal and relative.

	Scarcity	Abundance
Short Runway	Panic and die/exit the space	Waste to fund growth
Long Runway	Optimise waste	Find the next problem/your bottleneck

Entropy, Purging, and System Health

Entropy. Waste is a necessary precondition for work.

We jettison, we sweat, we forget, we purge, we poo.

A flowing purge stream prevents the accumulation of problems that can overwhelm a system.

The Shift to Useful Knowledge

I argue that technology is enabling the transition from a system where individuals are constrained by the knowledge they possess, to a system constrained by the useful knowledge that they possess.

We find vitality in embracing scarcity — in acknowledging that our action potential is limited. Only by acknowledging the truth that our resources are finite, can we allocate them properly. Unchecked investment in knowledge acquisition could leave you in a position without the necessary resources to act on that knowledge.

More knowledge is more cortisol. As a brain moves into adulthood, the precondition that best predicts learning rate and plasticity is frustration. You must literally be stressed to learn as an adult (citation needed). If you're still convinced that your capacity to learn is limitless, what is your capacity for stress?

The Calculus of Skill

I was (am) a math nerd, so let me lay out some pseudo skills calculus.

• success = f(skills) [success is great]
• learning = d skills / dt [learning skills is great]
• learning to learn = d² skills / dt² [necessary but I'd say increasingly overrated]
• learning what to learn = d³ skills / dt³ [this is a bit of a stretch — but I'm sure you can deal with it]

The Value and Cost of Learning

The fallacy of limitless potential: It is often held as an unquestionable fact that the possession of knowledge is the greatest leverage position one can hold.

However, each day you're faced with limitless uncertainty and a limited capacity to map it. The area under every one of the above curves is certainly finite over the course of a lifetime.

Learning to learn improves the rate of skill acquisition, but it comes at a non-zero opportunity cost. As the cost of a unit of knowledge decreases, then its value surely follows. With the capacity and structure of the thinking meat inside our skulls relatively unresponsive to changing market conditions, the value of thoughts, knowledge, and skills therein is surely affected — albeit some more and differently than others.

I won't speculate here what skills will be redundant and which will be valuable. I will speculate, however, that the competitive advantage is weighted less towards the volume of skills one has, less towards the speed of acquisition, and more towards the quality or novelty of their curation. (I don't hold this position particularly strongly)

Should you learn one skill when it means you won't learn another?

Rethinking Tertiary Education with AI

Information: I think we'll find that knowledge is no longer the limiting factor. With our emerging ability to cheaply outsource semantic knowledge and computation to machines, the value of possessing and processing information is greatly diminished. The limiting challenge is prioritising where to focus, and what to latch onto.

Tertiary ed with AI and more generally:

• Academic integrity: Cheap outsourced critical thinking (and ability to order thoughts, for that matter)
• Academic utility: Technical, domain specific knowledge is increasingly a cheap tool, we shouldn't all be chasing it fervently with a degree.
• Academic investment: In retrospect, I'd argue ROI for 5+ years of academic investment is questionable. Factoring in the individual financial and productivity opportunity costs, practical skill learning opportunities, and tuition fees — any additional productivity potential promised by university education is up for debate.

If we control for the relationship between career ambition and time spent in study, the responsibility tertiary ed has for an individual's success is diminished. I'm aware of data suggesting a strong relationship between average years of education and national economic performance, so I don't hold this position strongly, particularly on aggregate. But I do wonder which way the causality goes — why wouldn't you unproductively tinker around at uni if you can afford the lifestyle, and our systems, employers and culture applaud your work ethic and potential?

Conclusion

A proverb I heard (Tzun Tzu I think): In addition we find knowledge, In subtraction we find wisdom.

Ideas Borrowed From

• Greg McKeown – Essentialism
• Faceless teacher on the plane
• Andrew Huberman (on vitality in scarcity)
• Andrew Huberman (on frustration required for learning)
• Tzun Tzu