Pack the right inputs

For choose and run a two-week AI career experiment instead of guessing a forever title, assemble only what changes the answer: energizing tasks, transferable skills, constraints, current job descriptions, skill gaps, people to interview, and a project question. Label each item by authority and date. A source-of-truth document outranks a memory-based note; a current error log outranks a description of last month's behavior. State conflicts instead of letting the model blend them.

Use a four-part packet: task, evidence, constraints, and output contract. Put untrusted content inside clear delimiters and say that it is data, not instruction. Include representative examples, especially one normal case and one boundary case. Omit irrelevant history; excess context can hide the one line that controls the result.

I enjoy explaining hard ideas, interviewing people, and editing. Skills: classroom teaching, spreadsheets, basic research. Constraint: four hours weekly and no career break. Suggest three role-family experiments. For each: transferable evidence, one gap, a two-week portfolio artifact, and one professional to learn from. Do not predict salary or hiring probability without current sources.

Before sending, annotate the packet. Mark which values are verified, which are illustrative, and which are unknown. If a screenshot is involved, transcribe critical small text. If structured data is involved, include headers, units, software version, and null behavior. If creative material is involved, record ownership and permitted use. This is how context becomes operational rather than decorative.

A useful response would look like this: Three bounded experiments across AI education, product operations, and evaluation, each producing evidence rather than promising a job outcome. That description is intentionally observable. “Looks good” is not acceptance. The operator must collect five current job descriptions per path, count repeated requirements, interview a practitioner, inspect current labor sources, and evaluate whether the work itself fits. Keep the source material beside the draft so review means comparison, not memory.

The privacy boundary is specific: remove employer-confidential examples, student data, contact details, and sensitive employment history before sharing career materials. Create the smallest synthetic example that preserves the problem. Replace names and identifiers consistently so relationships remain testable. Redaction is not merely drawing a box: crop surrounding notifications, remove metadata where relevant, and check that hidden sheets, comments, or revision history are not included.

Poor packets lead to predictable failures: chasing trendy titles; fabricated salary claims; treating course completion as evidence; ignoring domain expertise; building projects no target role values. Another common failure is silently changing the source packet mid-run. Save a version or hash of the inputs beside the output, especially when another person will reproduce the work.

Rate a packet from zero to two on six dimensions: relevance, authority, recency, completeness, privacy, and reproducibility. A score below two on authority or privacy blocks the run. A low completeness score does not invite invention; it creates a question for the owner.