⬡ Your first lesson · Run your controls · 1 of 6
Your first control: a question you already know the answer to
By the end you can seed any batch task with answers you already hold, so the knowns tell you whether the run can be trusted.
Every lesson after this one assumes you can tell a trustworthy run from a quiet failure. This is where you learn to tell.
Input check
- 14,812Product Specification - T4 DNA Ligase
- 25,003Product Specification - Proteinase K
- 370,915Welcome to our product catalog
- 44,977Product Specification - RNase A
- 570,915Welcome to our product catalog
- 65,120Product Specification - DNase I
FLAG: pages 3 and 5 are identical. Six products should not produce the same page.
1 · Learn the move · Verified-data override
Never run only unknowns. Seed every batch task with items whose answers you already hold, and let the knowns judge the run. The incident this lesson is built on: six supplier pages pulled for six enzymes, every request returned HTTP 200 - success - and two of the six responses came back byte-identical at 70,915 characters. The catalog homepage, twice, from a silent redirect. The status code said everything worked; the control that caught it was noticing that two things which should have differed did not. That is the whole discipline in one sentence. Your verified values always override the model's output: when a control comes back wrong, the run halts and nothing from it gets used. A disagreement between output and a known is not noise to average away - it is the most useful information the run produced.
Extract [fields] from the sources below into a table. Two of these items are controls: [item], where I know [field] = [value], and [item], where I know [field] = [value]. Before extracting, compare the raw sources: report the length and first line of each, and flag any two that are identical - identical responses to different requests are a failure signal, not data. After extracting, check the controls. If either control disagrees with my value, my value wins: mark the run FAILED and do not present the other rows as usable.
2 · Your turn. You write the prompt
You pulled the text of six supplier specification pages for the enzymes in your reaction and want the storage temperatures and unit definitions extracted into one table. For two of the enzymes you know the values cold, because they are on the tube in your freezer. Write a prompt that makes those two knowns the judge of the whole extraction.
Write the prompt for this scenario as given, or swap in a real one from your own work. The score works the same either way.
Remember: the AI sees only your prompt, not this page. If the situation isn't in your prompt, it doesn't exist.
Optional. These shape the output when you run your prompt below, not your score.