Give the same problem to ten capable people working separately, and many of them will reach recognisably the same idea within weeks of each other. This is not coincidence, and it is not copying. It happens because capable people read the same research papers, face the same limits and start their reasoning from the same place.
Everyone working in a field draws on the same shared, public knowledge: the same published work, the same methods and the same observations. If an idea can be reached by reasoning from public information, then more than one person can reach it.
We tested this on ourselves. We took twelve ideas we were genuinely excited about and had an outside reviewer check each one. The reviewer’s only job was to find a reason to reject it.
Almost all of them had already been thought of by someone else, usually years earlier, and written up in places the people building the idea had never read. The ideas were good, but with very few exceptions they were not new.
This does not make the ideas worthless. It means the value was never going to come from the idea itself. What still holds up after an outside check is not “did we think of it first?” It is the things other people cannot reach by reasoning from public information:
- a position in a field that nobody else holds;
- data that nobody else can produce;
- a proof that only your particular method can produce;
- engineering depth that took years to build and that nobody can recreate in a weekend just by thinking about it.
This changes how we judge our own work. Excitement about an idea does not show that others cannot copy it, so we do not treat it as evidence of that. Every possible new direction goes through the same outside review described in the field report on this site.
The ideas that pass are, by design, the ones whose value never depended on the idea alone. What is genuinely new is rarely found in the knowledge everyone shares. It is found in the few areas where nobody else is working, which is a much smaller and more useful place to look.
0 comments