Put your two tables next to each other

Before reading on, look at your own two tables from 3.4 and 3.5. Write down three things
that were the same and three that were different.

What was the same

Almost certainly all of this:

  • The four beats. Read, plan, act, check. Both did it. Neither invented a fifth beat.
  • Both needed a clear request. Neither could read your mind. The clear request from
    Module 2 worked in both.
  • Both needed checking afterwards. Neither earned blind trust.
  • Both could make a mistake and recover. Same self-correction.
  • Both could have done something you did not want. Same reason to use git.

What was different

  • Where the work appears. Terminal text, or an editor panel next to your files.
  • How the plan is shown. Text going past, or a document you approve first.
  • How approving works. Per command, or per batch.
  • How much you see. One shows you nearly everything by default, one summarises more.
  • What checking looks like. A terminal command, or a browser opening.

Now sort your own list

Take your three differences. For each, ask: would this still matter if both tools got a
software update tomorrow?

Almost everything in the “different” list is a design decision that could change next month.
Everything in the “same” list has been true since agents existed and will still be true when
both these products have been renamed.

The point of the whole module

You are not learning Claude Code. You are not learning Antigravity. You are learning the
loop, and using two tools that happen to run it.

The person who only learns one tool has to start again when their company standardises on a
different one. The person who learns the loop opens the new tool, finds the four beats
within ten minutes, and carries on.

Which one should you use for the rest of the course?

Whichever you preferred. Genuinely.

From Module 4 onward the instructions are written to work in either. Where something is
truly tool-specific, it will say so.

If you have no preference, use the one where you could most easily see what was happening.
Right now, seeing clearly matters more to you than going fast.

Try this before the next lesson

  1. Compare your two git diff outputs. Did the two tools produce meaningfully different
    code from the same request?
  2. Which tool made it easier to catch a mistake? That is a real reason to prefer one.
  3. Delete one of the two folders. Keep the one you will continue with.
PreviousNext