Playing on Jevons Paradox's Field

Nvidia and Jevons Paradox

Jevons paradox is a phenomenon in economics that says when the efficiency and productivity of using a resource increase due to technological progress, instead of leading to a decrease in its consumption, it actually causes an increase in the consumption of that resource. It's as if efficiency and sufficiency never align; the higher the efficiency, the lower the sufficiency.

For a few days now, the discussion has resurfaced that by optimizing algorithms, the same performance can be achieved even with weaker hardware. In a sense, in the competition between algorithm and hardware, it is the algorithm that wins. This is a view that I also believe in; but I think there is a point that is better to pay attention to.

We have been seeing over these years that algorithms have become more advanced and optimized, and hardware has moved forward alongside them. I think Jevons paradox is perfectly observable in this case too: if we consider algorithms as those technological advancements and hardware as resources, the more our technologies advance, the greater our exploitation of hardware becomes.

Why is this interesting to me?! With this DeepSeek R1 situation that occurred and showed that choosing a more appropriate algorithm reduces the need for hardware, there is an ongoing discussion that surely in a few years, Nvidia and other microchip manufacturing companies will be driven to petty theft. But I think this event, although it is a reminder of the repeated victory of the algorithm in this game, is considered another jumping point in hardware advancement, and there will be no news of the industry's sufficiency in hardware.