Biodiversity is important because a single disease could wipe out everything in a niche if that niche has only one species. This is the concern with having all cows and chickens being the same breed. What does technological husbandry look like?
In the same way that we double down on breeds and species that have the highest productivity, we have a tendency to double down on the technologies that have the highest impact in the current environment. This leads to ideas dying out that might be useful in a new environment (There are many heuretic environments) and not pursuing ideas that are not optimal for the current environment. Ideas that were low impact in a past environment but might be impactful in a current environment are Unexploited Leonardo Ideas.
Seeing that working on a technology that is suboptimal could lead to a better outcome several nodes down the line or in a different environment is one reason why The tech tree model suggests that explicitly drawing out tech trees is important. Often Systems must take performance hits to get out of local optima and viable but unfit ideas can show a path.
You could almost imagine some organizations as technological zoos, trying to keep the technological pandas alive. This analogy isn’t great because the reason to keep pandas around is mostly an aesthetic one. Maybe the Svalbard seed bank is more apt.