Ages? I think it was about 2016 when people started credibly saying it’s over. So, that’s 8 years ago, which is about how long i think it took for the high end to trickle down to affordable. The graphics difference between a ps4 and a ps5 are… noticeable, but not shocking, and they were released in 2013 and 2020 respectively.
Graphics between an end-of-life and a new next-gen console are a bad comparison. Late PS3 and early PS4 games also saw only small improvements. But compare early PS3 to late PS4 titles and the differences are staggering. The PS5 has about an order of magnitude more processing power than the base PS4 and Moore's law states that computing power doubles every two years, which means it gains an order of magnitude every 6 or 7 years - which is precisely the time frame between the PS4 and PS5 release. Moore's law definitely isn't dead and the new console generation proves it.
> The PS5 has about an order of magnitude more processing power than the base PS4 and Moore's law states that computing power doubles every two years, which means it gains an order of magnitude every 6 or 7 years
That's not what moore's law is. Moore's law says that TRANSISTOR COUNT will double every 2 years.
> Moore's law is the observation that the number of transistors in a dense integrated circuit (IC) doubles about every two years
First off, your calculation is wrong because these chips are not the same size. Secondly, if you actually go with the definition of Moore's law, my argument still holds - and it's trivial to prove. From your linked article:
>Moore's law [states] the number of transistors that could be housed in a dense integrated circuit [doubles every X]
(emphasis mine)
The PS4 GPU was a 28nm chip and PS5's is a 7nm one. A factor four reduction in length between transistor terminals equates to a factor of 16 increase in transistors that could be housed in one unit of area. That's precisely what you'd have expected from your definition of Moore's law. In reality there's a bit more to it, like efficiency and power consumption (Moore included that in his original estimate), which slightly lessen whole deal of increasing compute power. But as stated in my first comment, it still holds as an OOM increase when you look at raw FLOPs in the end, which is what we expected.
It's not only about calculation speed; the cost of building an AAA class game has exploded over the last 10 years. Maybe you can render 10x as much stuff on screen, but somebody has to actually draw and integrate that in the game. Also, the cost of the needed artwork seems to put game design on the back burner since taking risk is no longer an option. (just like modern Hollywood movies seem to be bland)
