How I stopped starting at the beginning

Starting with the End

I do not usually begin at the beginning. I tend to think about what done looks like, then work backward from there.

Imagine you have just spent three days refining a research paper, trying to account for everything you can think of. You have gone through iteration after iteration and arrived at something that feels defensible, maybe even pretty good.

Then you hand it to someone else to read, or to a fresh AI prompt with no context.

Their reaction is the thing you did not want to hear. They land immediately on the most obvious question, the one you somehow missed because you had been too close to the work for too long. Or they do not even get that far before spotting a typo in the section you manually edited at 12:30 the night before.

So how do you avoid that?

Jump to the end first. Decide what done should look like. Decide what a strong result would need to cover, for this paper, in this context.

# Pre-flight Prompt / Validation Prompt 

Think for a better answer.

There is a topic that is very much a buzz today: data centers in space. The question is, is there anything that is valid about thinking about it today? Or is this just hype and political maneuverings? I have an intuition about the answer, but I also respect that there are companies and governments who are seriously looking into it. 

Given that framing, assume I'll do my own research about this. The questions I have for you: what questions would you have of a paper on this subject? What would you want to to know from me and of my research? What would you be curious about and disappointed if I didn't touch on it?

Be as thorough and detailed as you like.

But suppose you forgot and started at the beginning anyway. I did.

You can still recover before you publish. Open a fresh prompt with no memory of the project chats or earlier drafts, then ask the framing question your own way. In my case, I tried to imagine what a thoughtful person, hearing the topic for the first time, would want to know and what they would be disappointed not to see addressed.

That gave me a much better starting point than continuing to polish the draft I already had.

I used the output from that fresh prompt to build a deeper research prompt and a more structured plan. It helped surface gaps, sharpen the outline, and pressure-test the paper before I invested more time taking it further.

Useful reminder for me: sometimes the better move is to begin with the reader who has seen none of it yet.