LLM Fundamentals · What You Can Do Now

Don't stop at "I get it." Lock That One Job Down first

After the fundamentals, the usual trap is this: you feel like you learned a lot, then you close the laptop and your hands have nothing new to show.
So from here on, every chapter ends with a page like this. It answers one question: after this chapter, what can you actually do today.

The whole course has one hands-on spine: ship an Agent that actually gets work done. Six milestones, one square each. You're on the first square.
M0
Know what you want it to do for you
M1
Talks like a human, reliably
M2
Actually gets work done
M3
Better or worse, measurably
M4
Runs long without losing memory
M5
The process is reproducible
See it first · How tight does a requirement have to be before it counts

Most people hand AI a task and stop at step one: "Help me handle the weekly report." That sentence isn't wrong — it just can't be accepted. If you don't know what output counts as success, you also don't know whether to change the prompt. Push three more steps and it becomes a real requirement.

From a wish to a requirement you can accept

01
A wish

"Help me handle the weekly report." You have an idea. That's all it is

02
Lock input and output

In: a week's scattered notes. Out: three takeaways plus a next-week plan

03
Define what counts as right

All three sections can go straight to your boss. Not one word to rewrite

04
Pick a plan that holds down hallucination

If it has to cite real numbers, go RAG. Don't let it invent from memory

Stuck at step two. This requirement still can't be accepted
Stop at "help me handle the weekly report" and you jam at step two: you haven't even decided what the input is, or what the output looks like. The last two steps never get a chance to run. "I tried it, didn't feel right" usually starts here.

Why step four gets its own callout

This chapter you learned four ways to hold down hallucination: change the prompt, RAG, tune Temperature, add evals. Which one you pick depends on where your job will go wrong. This is not a four-way preference quiz. Need to cite specific numbers and clauses? Go RAG. Format keeps drifting? Change the prompt and drop Temperature. Running it long-term and afraid it'll quietly get worse? You need evals. Picking the wrong one is fine. Picking none at the start means you leave hallucination for the user to find.

Hands-on checklist · Pick one, start it, tick it off

This Chapter's Hands-on Checklist

0 / 3 done

Write that job in four lines

15 min Everyone

Follow the four steps above. Write down the one job you most want to hand to AI: one wish, what the input is, what the output looks like, what counts as right. Sticky note, notes app, anywhere — but it has to be written down, not just an idea in your head.

What counts as done
Read the four lines to someone who doesn't know your job. They can restate "what AI has to deliver." If they can't, steps two and three are still too vague.

Try it on 5 real inputs

1 hour Want to test feasibility first

Don't use made-up examples. Dig up five real materials and throw them at it, one at a time, same prompt every time. The point is to see which kind of input makes it start making things up. How good the answers are can wait. Too long? Missing info? Internal jargon? Write down the class that fails.

What counts as done
You can say one concrete sentence: "The moment the input has X, it starts inventing." A vague "sometimes it's not accurate" doesn't count.

Draw a human–AI boundary

Half a day Ready to actually build one

Split the job into "AI does" and "you do." Write down what the handoff looks like. Example: AI drafts and flags every number it's unsure about; you only check the flagged ones. Where you draw the line matters less than this: it has to be checkable.

What counts as done
You can answer: if AI's half goes wrong, at which step and by what method would you catch it? If you can't, the line is drawn too far back.

Once it's written, save it to the Build Log

The Build Log stays with you through the Vibe Coding Methodology part. Fill all six milestones and you'll have a complete Agent design brief. It lives in your own browser. You can export Markdown anytime.

Fill in M0