I don’t agree. Aircraft had their growth phase earlier, first flight in 1903 to landing on the moon in 1969. Airliners look like they did in the 60s because there’s nothing left to improve on with regards to efficiency. How many orders of magnitude are there between the wright flier and a 747?
We’re reaching the diminishing returns portion of the logistic curve for computing with regards to both things like hardware and compiling, compression, etc. It’s just hard to compare industries which are done growing with ones that still feel like they might be exponential forever.
> It’s just hard to compare industries which are done growing with ones that still feel like they might be exponential forever.
Some industries never evolved too much. I don't see a big functional difference between cars a 100 years ago and cars now. And there were electric cars in 1900s.
A car is constraint by lots of things just like GP comment affirmed. Putting more gadgets in a car won't change what a car is or how it works at a fundamental level.
From the very first microprocessors to today's there is a big leap. If we apply that leap to cars they should be using nuclear fusion and take us to moon and back in seconds. And it's not like trillions of dollars weren't invested in car R&D in more than a century.
Software growth seems exponential only because of hardware and the Moore's Law. Which will come to an end soon.
Thanks to Moore's Law, we programmers, afforded to be careless, not think about performance or use unnecessary abstractions just out of esthetical or ideologic considerations. It allowed businesses to say that time to market trumps everything and customers should simply buy more powerful boxes or more boxes. Now this will come to an end soon and we will have to think about efficiency, speed and performance.
Or maybe chips will move from silicon to some alternative processes where growth is still possible.
You are making my point stronger rather than weaker: Aircraft have only become more efficient in the last 50 years or so and that's precisely the equivalent of better compilers and compression.
And the diminishing returns for computing are for the most part driven by the lack of investment in new ways of computing because we got so much for free. But once the free ride is over you can expect the research into efficiency to be picked up again because that will be the only field where real progress can be made. There have been a few half-assed (apologies to those involved, it was no doubt a ton of work) attempts at making computing fabrics and clockless machinery. And that's because by the time you have something working the scaling advantages have already overtaken your work and absolutely nobody is going to give up a working architecture for something unproven if it doesn't given an immediate return. Some pretty good ideas died like that.
Yes, for digital computing. But computing is on the cusp of becoming more non-determistic, which will allow analog computing to absolutely demolish digital computing in performance.
We’re reaching the diminishing returns portion of the logistic curve for computing with regards to both things like hardware and compiling, compression, etc. It’s just hard to compare industries which are done growing with ones that still feel like they might be exponential forever.