Has anything actually been invented since 1970?
Relevant:
2) Breakthroughs are simply very impactful innovations -- of course earlier innovations are going to have an advantage: it's like saying "The total number of descendents per capta has been dwindling since the dawn of ages. We're doomed!". Claiming the rate of breakthroughs should remain constant is ridiculous.
3) Additionally, inventions are information. They're cumulative and knowledge instensive. We can't expect them to occur at a constant rate simply because it will get harder and harder to learn eveything required to invent some really new technology/scientific discovery. I like to think of human knowledge as a chunk of information that we compress and pass along. There's only so much we can compress, and our brain hasn't really changed significantly to accomodate and absorb more information faster. The result of course is specialization, and again specialized innovations by definition have limited scope.
4) To claim we had no specialized breakthroughs recently is even more ridiculous. Take the computing aspect of engineering alone (and math) and you probably get more small breakthroughs than there were inventions at all in the 18th century. Telecommunications also developed completely in the last few decades.
The WWW is actually a 1960s idea. Watch Engelbart's demo.
Sometimes I wonder if the problem is the massive growth of free market ideology (market fundamentalism) since 1970: Ayn Rand, the Reagan Revolution, libertarianism, etc. The free market is heavily biased toward extracting pre-existing value and things with short time horizons. It rarely spends money on very speculative R&D, and the culture that it begets is a hustler culture of quick-to-market, quick-to-exit, etc. Things like the Engelbart demo are the product of very smart people with very rough, flexible deadlines sitting around and thinking a lot. There's no way to monetize that. A related problem is that invention itself is hard to monetize except via patents... the innovator often loses in the market to fast followers.
In machine learning terminology we could say that the free market is a greedy hill climbing algorithm prone to converge on local maxima and unlikely to cross fitness valleys.