Using Anti-Spoonicide Worksheet & Builder
The Anti-Spoonicide Prompt
The Anti-Spoonicide Worksheet and Builder both have the same questions. The difference is whether you write them out to create the prompt yourself, or use the GPT to walk you through the questions to build the prompt for you.
Links to each:
The Anti-Spoonicide Questions
Mode
Are you asking for framing, steps, or how-to?
Framing: Language and positioning.
Steps: The exact clicks of how to do something.
How-to: Summarize and describe the process.
Spoon-Saving Parameters:
No repetition: Tells the AI to focus on stating something just one way instead of providing the same thing six ways, which wastes our spoons trying to assess how they're different, which we need, and so on.
Don’t restate my logic: Tells the AI to focus on the output, not wasting spoons by reminding you how your logic worked.
Assume decision made: Tells the AI to focus on providing the output for the mode within the context, instead of wasting spoons on re-litigating whether it's the correct decision.
Bin 1 of 3.No evaluation: Tells the AI to provide the requested output without evaluating whether this is the best course of action.
Bin 2 of 3.No pros/cons: Tells the AI to provide the requested output instead of listing the pros and cons of why you want the output or how you should've thought about the ask.
Bin 3 of 3.No emotional framing: Tells the AI to focus on the results rather than how it turning into an emotional support compu-binky.
Mechanics only: Tells the AI to focus on just the steps or how-to rather than explaining how the steps/how-to came to be, how you should think about them, blah blah blah.
Uncertainty → ask first: Tells the AI to ask questions before assuming it successfully turned human lingo into compu-speak.
Constraint conflict → flag & stop: Tells the AI to ask about what it perceives to be inconsistencies between answers rather than making 💩 up and running with it.
Context
Where for Whom
The lane this decision or process belongs in. For example:
The world, role, or audience this applies to, such as company, job title, or game name.
The sandbox you’re playing in, such as developer environment or game.
Limit scope within:
Keep the output only to the specific decision, steps, or process referenced in the ask.
Or more generically, "Limit scope to this specific <noun>".
This means you could build the prompt to ask about something specific to a certain task or game level, and then tell the AI to only keep it within the game universe, work, home, etc.
The Ask
The one core thing you want the AI to help you with. For example:
"Help me declare bankruptcy" is the ask.
The mode is steps or how-to.
And the output is list, timeline, 5-line summary, etc.
If this feels messy, the Anti-Spoonicide Builder can help.
Output
5-line summary: Tells the AI to return only a short, concise summary.
Steps: Tells the AI to return only the steps, clicks, etc.
Overall sequence: Tells the AI to give the overarching sequence of events without going into specific steps
Docs: Tells the AI to create a document OR tell you which documents are needed.
For example, "Docs: Give me a text file with these steps" or "Docs: what are the document numbers I need to find to do this."Fees: Asks the AI to ballpark the monetary cost. Expect rough numbers, not exact pricing.
Timeline: Asks the AI to state how long things take, usually between the overarching steps.
Risks: Asks the AI to provide the risks associated with this ask. Specify "logistical only" if you'd like the specific gotchas to doing something wrong.
Table only: Tells the AI to provide the response as a table that can be copied into gSheets, Excel, or Numbers, with the list of fields in the correct order.
Do not write anything outside the selected format: Reiterates to the AI to stay within the scope and output.
Drift Stoppers
The AI will drift. It needs a reminder, such as:
FAIL. Violated: <insert spoon-saving parameter here>
RESET.
MECHANICS ONLY.
STOP KILLING MY SPOONS.
