Chapter 1 · Prerequisites

Checkpoint · Prerequisites

Eight questions on the prerequisites before Larry meets the models. Pass to move on to Chapter 2.

Builds on

Chapter progress

0 / 11 lessons completed

Up next in this chapter: Neural networks in 800 words →

You've walked the prerequisites: why scale forces specialization, the three axes that shape an app before you write serving code, what makes inference the critical path, how online and offline pull in opposite directions, how the audience sets the SLA, the evaluate → fine-tune → distill order, and the latency-throughput language the rest of the course speaks. Prove it stuck before the models.

Eight questions. Score 80% or better to close out the prerequisites. Every option explains itself, and you can take another pass any time.

01 · Checkpoint

Prerequisites — the ground before the models.

Scale and specialization, the three app axes, AI-native reliability, online vs offline, consumer vs B2B, model selection in order, and the latency/throughput metrics you'll quote for the rest of the course.

Check your understanding0 / 8

Question 1

Why does serving at scale push you toward specialized models instead of one large generalist?

Question 2

Larry places every inference app on three axes before choosing a stack. Which set is it?

Question 3

What distinguishes an AI-native application from one where AI is a bolt-on feature?

Question 4

Online and offline inference optimize for different things. Which pairing is right?

Question 5

How does the consumer-vs-B2B axis shape the rest of the stack?

Question 6

Model selection is three levers. In what order does Larry reach for them?

Question 7

Fine-tuning and distillation both improve a model, but differently. Which contrast is correct?

Question 8

Why does Larry quote p99 latency instead of the average?