Skip to content

Final oral defense

The examiner should follow the learner's answers until the first unsupported assumption appears.

Model & first principles

  1. What is the state of the system, and why are those variables sufficient?
  2. What are the governing balances or conservation laws?
  3. What are the dimensions/units of every term in your core equation?
  4. Which relationship is derived and which is empirical?
  5. What is the simplest model that would falsify your current story?
  6. What happens in the zero, infinite, very-small and very-large limiting cases?

Uncertainty & evidence

  1. What is your largest epistemic uncertainty?
  2. What is aleatory variability versus reducible ignorance here?
  3. What observation most changes your posterior belief?
  4. Which measurement is a proxy, and how could the proxy decouple from the goal?
  5. What negative result did you preserve rather than explain away?
  6. Which result has not been independently reproduced?

Causality & intervention

  1. What exactly is the intervention, population/system and outcome?
  2. What confounder would most threaten your conclusion?
  3. Where could conditioning/selection create a collider or survivorship problem?
  4. What feedback does your intervention cause in agents, organisms or institutions?
  5. Which observed association would remain even if your mechanism were wrong?
  6. What experiment best distinguishes your favored mechanism from its strongest competitor?

Optimization & control

  1. What is the objective function, and who chose it?
  2. Show a scenario in which optimizing your metric makes the real objective worse.
  3. What constraints are hard, soft, legal, safety-critical or merely conventional?
  4. What state is unobservable or uncontrollable?
  5. What happens when feedback is delayed, saturated or noisy?
  6. Which safety constraint is outside the optimizer's authority?

Physics & scale

  1. Where do matter, energy, momentum, charge and heat go?
  2. What is the dominant scaling law as size/throughput grows 1000×?
  3. Which transport process becomes limiting at scale?
  4. What fails first: material, thermal, electrical, mechanical, biological or organizational?
  5. Which current law of physics is your design leaning hardest on?
  6. What would require new physics rather than better engineering?

Numerics, computation & AI

  1. How do you know your numerical answer is not a conditioning/discretization artifact?
  2. What happens if the model changes its own tools, prompts or evaluator?
  3. Which part of your AI system is capability and which part is interface/scaffold?
  4. What benchmark shortcut could reverse your conclusion?
  5. What stays outside the self-improvement boundary and why?
  6. If foundation models become 10× stronger, which of your skills becomes more valuable rather than obsolete?

Manufacturing & economics

  1. What changes between one prototype and 10,000 units?
  2. What is the bottleneck process and why?
  3. What drives yield loss, maintenance labor and downtime?
  4. What is the unit cost at scale, and which technical parameter has highest economic elasticity?
  5. What critical input has concentrated supply or long lead time?
  6. What improvement would not matter economically because another bottleneck dominates?

Safety, security & reliability

  1. What is the top hazard even if every component works as designed?
  2. What common-cause failure defeats your redundancy?
  3. What adversary benefits from exploiting the system and what can they control?
  4. What weak signal will your organization normalize after repeated success?
  5. What event forces shutdown, rollback or loss of deployment permission?
  6. Who has stop-work authority when leadership wants to continue?

Power, history & institutions

  1. Who defines success, and who cannot refuse participation?
  2. Who receives concentrated benefits and who bears diffuse costs?
  3. Who owns the data, infrastructure, model, patents or productive assets?
  4. What are the exit, voice and appeal mechanisms?
  5. What historical analogy are you using, and where does it break?
  6. What would a political actor rationally do with your technology that an engineer might not intend?

Ethics, wickedness & communication

  1. Give three incompatible but reasonable formulations of this problem.
  2. What value conflict cannot be resolved by more data?
  3. Which decision should be reversible because the model is uncertain?
  4. What is the strongest moral argument against your preferred deployment?
  5. What disagreement in your team is about facts, what is about values, and what is about identity/status?
  6. What would a person harmed by your system say you failed to understand?