The problem terrain insights generated by many "failures" are what make resolving interesting trivial, silly and unlikely questions so helpful. They generate novel knowledge and new ways of thinking about things. They often point the way to useful but previously not envisioned work.
Edison and the long line of "failed" lightbulbs is a cliche, but still rich wisdom.
But 1000 Edisons working on 1000 highly different "light bulb" problems, sharing the seemingly random insights they each learn along the way, are going to make even faster progress -- often not in anticipated directions.
PS: veritassium is just amazing. Yesterday I learned that conservation of energy is a local phenomenon and a geometric consequence, not a law of the universe at all. I am 36 with an engineering background and conservation laws were close to sacred laws of the universe. It turns out, not really and it has to do with the universe expanding. Veritassium just drops it like that, with a nice story.
Like speech, ideas require an open field with a lot of garbage to hit many home runs.