Ideas do not die of being wrong
Go back through the ideas you dropped in the last two years. Very few of them were disproved. You did not run the numbers and conclude it would not work. You just never found the edge to start from, and the energy leaked out over a few weeks until it was gone.
That is a planning failure, not a judgment failure. The gap between "this is interesting" and "I am doing this on Tuesday" is the whole game, and nothing about having good ideas prepares you to cross it.
The bad news is that the gap does not close by thinking harder about the idea. The good news is that it closes with a fairly mechanical process.
Step one: say the idea out loud until it stops moving
Before anything can be planned it has to hold still. In your head an idea keeps shifting shape, and every time you revisit it you are working on a slightly different one. Saying it out loud fixes it in place.
It also exposes the soft parts. You will hear yourself say "and then people would obviously want to pay for it" and hear how much weight the word obviously is carrying. That is not a detail you catch while typing, because typing lets you skate past it.
This is what the call in oThought is for. You talk, it asks the next question, and it saves the idea when it becomes one rather than making you stop and write it down. The method for the capture part is in how to capture ideas by voice without typing, and the whole session, start to first step, is on AI brainstorming, out loud.
Step two: find the assumption that would kill it
Every idea rests on a small number of load-bearing beliefs. Usually one of them is doing most of the work, and usually you have not looked at it directly.
Ask: what would have to be true for this to work, and which of those things am I least sure about? Not the hardest to build. The one where you are guessing.
Common shapes it takes:
- People have this problem often enough to care.
- They are already paying to solve it some other way.
- I can reach them without spending money I do not have.
- The technical part that looks hard is actually the easy part.
Name it in one sentence. If you cannot get it into one sentence, you have two assumptions and you should pick the scarier one.
Step three: design the smallest test
Now build the cheapest thing that tells you whether that assumption holds. Not a prototype of the product. A test of the belief.
The bar is low on purpose: it should fit in a week, cost close to nothing, and produce an answer you would actually act on. Ten conversations with people who have the problem. A landing page and a hundred visitors. Doing the service by hand once, badly, for one person.
A good test has a stated outcome before you run it. "If fewer than three of ten say they have tried to solve this already, I drop it." Without that line you will interpret whatever happens as encouraging, because you want to.
Most ideas should not survive this step, and that is the point. Killing an idea in a week is a win. The expensive failure is the one that takes six months to become obvious.
Step four: let the research happen without you
Some of what you need is knowable before you test anything, and looking it up is where evenings disappear. After a call, oThought runs research on the saved idea in the background and attaches what it finds with the source, the date it was retrieved, and a confidence label.
Those three fields are the useful part. A claim with a source you can open is something you can act on. A confident-sounding summary with nothing behind it is a guess wearing a suit, and telling the two apart at a glance is what stops you from planning on top of something soft.
The idea also lands in a Mastermind Map next to your other ideas, which sometimes changes the plan entirely. Two ideas you thought were separate turn out to be one product with two entry points, and that is a better starting position than either had alone.
Step five: commit to one idea properly
At some point breadth stops helping. One idea has to get the depth, and the others wait.
That is what a Deep Pack is for. It is a one-time purchase that runs planning agents on a single idea you choose and delivers a multi-document PDF you can work from. It is not automatic and it is not applied to everything, because depth on twelve ideas is just breadth with extra steps.
What commitment looks like in practice is narrower than people expect. One idea, one assumption, one test, one date. Everything else in the map stays captured and stays available, and none of it needs attention this week.
If the deeper reason this all runs through a phone call rather than a text box is interesting to you, that is why call-first, not chat-first.
What this looks like in a week
Monday you call and talk the idea out until it stops moving. By the end of the call it has a shape and it is saved.
Tuesday the research is sitting there with sources attached, and one of them changes your view of the assumption you named.
Wednesday you write the test and the kill condition. Thursday and Friday you run it.
Then you either have an idea worth a Deep Pack or one you can drop without regret. Both outcomes beat the version where it sat in a notes app for a year while you wondered about it.
Questions
How long should the first step take?
A week or less. If your first step takes a month, it is a project, not a test, and you will learn nothing until the month is over.
What does the Deep Pack actually produce?
It is a one-time purchase that runs planning agents on one idea you select and delivers a multi-document PDF covering that idea in depth, so you have something concrete to work from rather than a summary.
Can I plan several ideas at once?
You can capture as many as you like, and the map keeps them connected. Depth is deliberately one idea at a time, because planning everything in parallel is how nothing gets started.