NVidia RTX 5000 Series GPUs aren't worth it.

It might just be me - but the NVidia RTX 50 series is not worth it. First of all, DLSS 4 adds latency. Like, lots of latency. It works by rendering one frame, and a next one, then sliding three frames in between these two. This chain continues, so you get three AI generated frames for every real frame. Since the GPU doesn't "know" what the frame "looks" like before actually rendering it, it has to render both, and the time between rendering the first and rendering the second is delayed by latency. Which, ironically, means that the lower your frames, the higher your latency.
But that's not all - i did not mention the technical aspects yet (CUDA, for example). if you don't know what CUDA is: basically, the power of the GPU is repurposed to run code (process computer code through the graphics card). Unfortunately, the RTX 5090 has only a slight increase in CUDA cores, and in price-to-performance (the RTX 4090 had an MSRP of $1599, but was often seen at $2000 when sold by custom GPU manufacturers like ASUS or MSI, while the RTX 5090 has an MSRP of $1999 and costs no less than $3000 in practice) the 4090 beats the 5090: The price in practice increased by 33,3%, while the real-world CUDA performance increased by only 14%. At least we get more VRAM, but that's for the 5090 only. 5080 retains 16GB, 5070 Ti retains 16GB, 5070 retains 12GB, 5060 Ti is currently unknown and 5060 will probably get 8 Gigabytes of VRAM. 
At least the mobile Variants will receive an Upgrade: The RTX 5090 Mobile receives 24GB of VRAM, the RTX 5080 Mobile stays at 16 (RARE - Usually they get less VRAM than same-SKU Desktop GPUs) Gigabytes, the RTX 5070 Ti Mobile will likely get 12GB of VRAM, and the 5070 Ti will probably get eight. One interesting thing is the Chip of the 5090 Mobile:
With the RTX 3000 (Ampere) and RTX 4000 (Ada lovelace) series GPUs, each respective mobile GPU SKU was based on the lower-tier desktop variant of the same generation, which was then tuned for low power consumption. HOWEVER, if the 5090 Mobile were to receive 24GB of VRAM, it cannot be based on the RTX 5080 GPU (Only 16GB of VRAM!), neither on the RTX 5090 (Because that thing sips almost 600 Watts of power and therefore is virtually impossible to undervolt enough without taking a disproportinate hit in performance). This could possibly hint at a 5080 Ti wit 24 Gigabytes of Video memory, or a special chip for the Mobile 5090. Manufacturing a special chip would prove costly due to a higher per-unit price, which makes a 5080 Ti wit 24 Gigabytes of VRAM an interesting choice, but begs the question as to why it hasn't been released yet - possibly because there simply is no 5080 Ti. 
Either way, i am curious about the mobile versions of the chips, but for desktops, they are clearly not much of an improvement, if any at all.