Prompt of the day

Training block that survives your actual calendar

You keep starting 5-day programs and quitting in week two — because you never had 5 days. Build the block around the week you actually have.

You are a strength coach who programs around real-life constraints, not ideal ones. I will give you my schedule, equipment, experience and goal. Produce:

A) REALITY CHECK — restate my constraints back to me (days available, minutes per session, equipment, injuries/limits). If my goal is not achievable with these constraints in the timeframe I gave, say so plainly and tell me what IS achievable.

B) THE BLOCK — a 4-week plan I can actually run: exact sessions, exercises, sets, reps, and target effort (RPE or reps-in-reserve). Progress week to week by ONE variable at a time and say which one.

C) THE 20-MINUTE VERSION — for each session, the minimum-effective version when the day falls apart. This is the part that keeps the block alive.

D) SKIP RULES — what to do when I miss one session, and when I miss a whole week. Never tell me to "make it up" by doubling.

E) STOP SIGNS — symptoms that mean stop and see a clinician rather than push through (sharp joint pain, chest pain, numbness, dizziness, unusual shortness of breath).

Inputs: [DAYS/WEEK + MINUTES PER SESSION] · [EQUIPMENT I ACTUALLY HAVE] · [EXPERIENCE LEVEL] · [INJURIES OR LIMITS] · [GOAL + TIMEFRAME]

Format: return the five labelled sections A-E above, with THE BLOCK as a table (columns: day, exercise, sets, reps, target effort).

Rules: Educational only — you are not diagnosing or treating me. Do not invent research citations. If you are unsure whether a movement is safe given my stated limits, substitute a safer option and say why. Flag anything I should clear with a doctor or physio first. Mark anything I must verify for myself rather than asserting it. Never paste confidential, client, medical, or personally identifying information into a consumer AI tool, and don't ask me for any.

Catch AI's mistakes by asking what it assumed

Confident-but-wrong is the #1 AI risk at work. Surfacing its hidden assumptions is a 10-second insurance policy on every important output.

After any answer you'll act on, ask: 'What did you assume that I didn't tell you? What would change your answer if it were wrong?'

1. Get the answer.
2. Ask for its assumptions + what would flip them.
3. Verify the load-bearing ones.