The reasonable upper bound of benefit when designing a new OS or a new information-sharing medium is, a priori, much higher than the upper bound of a new measuring cup or a new dessert. So, this means someone entering those quests can reasonably state that they are tackling something big, i.e. worthwhile.
I see your point. While breakthroughs (such as Penicillin) can still be a result of unrelated research, it is true that focused experimentation is more likely to make an impact.
The penicillin story is not what most people think.
Fleming discovered the mold and its anti-bacterial effect in 1928 by happenstance, but this knowledge languished until the outbreak of WW2 in Europe in 1939. At that point, the Oxford group realized the war was going to cause enormous numbers of gas gangrene cases (truly a terrible affliction), and they very deliberately set out to find a treatment. Their research was, therefore, incredibly directed, and, indeed, most of their work centered on culture and chemical preparation.
Fire, wheels, writing, and industrial production of penicillin -- greatest inventions ever.
