The Art of the Open Question

Separate the situation from the question with tags - <situation> for the standing facts, <question> for what you actually want to know, <dimensions> for what the answer must cover - and put the instruction last. The move that does the most work here is asking it to restate the real question before answering: a closed question usually hides a decision, and seeing that decision named back to you often changes what you ask next.

• Split <situation>, <question> and <dimensions> into separate tags.
• Ask it to restate the real question before answering.
• Invite it to say your framing is wrong.
• Ask for the uncertainty and the disagreement, named.
• Keep a long piece in an Artifact; put standing facts in a Project.

<situation>
I am opening an upscale cafe called 'Kondapur Mornings' in Kondapur, Hyderabad, next month. Kondapur sits next to the IT corridor - young professionals, young families, a lot of people on odd working hours. My marketing budget is small and I have to commit most of it in the first 90 days. I have no existing audience.
</situation>

<the_question_i_asked_badly>
"Is social media marketing a good choice for my new cafe?" I know that is a closed question and that the answer is yes. It is not the question I need answered.
</the_question_i_asked_badly>

<dimensions>
What a useful answer covers:
- Which customer segment to target first, and why that one rather than the others
- Which platforms, judged on what they are actually good at for a neighbourhood food business - and include Google Business Profile and WhatsApp in that judgement, not just Instagram and Facebook
- A first-month content plan, week by week, with specific posts rather than content categories
- Three metrics worth watching, two vanity metrics to ignore, and the number at which I should conclude something is not working
- What organic social will not do, and what else the same money could buy
</dimensions>

<task>
Before you answer, give me <the_real_question>: one short paragraph stating what you think the decision underneath my question actually is. If my framing is wrong - if "which platform" is the wrong first question for a business like this - say so there, plainly, and answer the better question instead.

Then <answer>, working through the dimensions above.

Then <assumptions>: everything you had to guess because I did not tell you - budget size, seating, hours, competition on the same stretch. List them so I can correct the ones you got wrong.

Then <where_i_am_least_confident>: which of your recommendations rests on general marketing principle rather than anything specific to Hyderabad neighbourhood cafes. I would rather know which parts to treat as generic advice.

Under 900 words for <answer>. Weight it towards the content plan and the metrics.
</task>

<question>
I want to understand the Vijayanagara Empire's place in South Indian history. I do not want a verdict on whether it was important - I want to be able to hold a conversation about it and know where the edges of my knowledge are.
</question>

<dimensions>
1. Historical significance in context: what changed politically, economically and in religious patronage because the empire existed. What it consolidated and what it interrupted.
2. Architecture: Vijayanagara compared against the preceding Hoysala dynasty. Materials, plan, scale, ornament, function. Named examples on both sides - Hampi against Belur and Halebidu. What the differences reveal about the two states, since that is the real reason to compare buildings.
3. The Battle of Talikota, 1565: why it is treated as a pivotal turning point rather than one military defeat among many. Separate what the battle destroyed from what was already failing.
</dimensions>

<task>
Work through the three dimensions.

Then <historiography>: where historians genuinely disagree. Name at least two live disputes - the empire's religious character and the suddenness of the post-1565 collapse are two candidates - and set out what each side argues rather than giving me a settled answer.

Then <confidence>: go back over your own answer and mark any date, figure, attribution or inscription you are not confident about. Specific numbers and named sources are where this kind of question goes wrong, and I would much rather you write "commonly given as 1565, and that one is firm" and "the troop figures in the chronicles are unreliable" than present everything at the same level of certainty.

If I ask for further reading, give me authors and subject areas rather than exact titles unless you are sure of the title.
</task>

<project>
I am planning a historical fiction novel set at the peak of the Kakatiya dynasty in Telangana - the Rudrama Devi and Prataparudra period. I asked myself whether a historical novel was a good idea, which is a closed question and a waste of one. The real work is finding the specific book.
</project>

<task>
Brainstorm five distinct concepts. Distinct means different in protagonist, scale and register - not five court intrigues with different names. At least one protagonist must have no access to power at all: a tank-builder, a weaver, a temple musician, someone whose relationship to the dynasty is entirely from below.

For each concept:
<protagonist>Name, station, and the thing they want and cannot have.</protagonist>
<conflict>The pressure that does not resolve on its own.</conflict>
<what_if>The hinge that makes this a novel rather than a chronicle.</what_if>
<the_weakness>What is wrong with this concept - thin plot, unsympathetic lead, too much exposition required, too close to something that already exists. Every concept has one. Name it without softening it; this is the part I will actually use.</the_weakness>

Then <commercial_assessment>: rank the five for today's market, treating the Indian English-language market and the international market separately, because they want different books. Say what the ranking rests on and flag where you are reasoning from general publishing patterns rather than anything specific.

Then <research_burden>: what I would have to research before writing each one, and which concept leans hardest on historical detail that may be thin or contested.

On sources: do not invent chronicles, inscriptions or scholarly works. If you are not confident a specific source exists, describe the kind of source I should go looking for and say that is what you are doing. A plausible-sounding but non-existent inscription would cost me weeks.

Put the five concepts in an Artifact so I can cut it down to two over the next few turns.
</task>