At the time of the announcement, the company, apparently, was not sure that the technology would work normally.

Almost all modern popular games support DLSS super -selection: with its help you can increase the frequency of personnel, renders in low resolution and increasing its quality thanks to AI. Initially, NVIDIA did not plan such a technology.
The story of the creation of DLSS It is described in The Nvidia Way Jensen Huang and the Making of A Tech Giant, built in an interview with company employees. It turned out that the idea of technology was born just a couple of weeks before the announcement.
At first, NVIDIA wanted to talk about DLAA technology – this is AI -based smoothing. The head of the company Juan Jensyun found that this was not enough: he noted that a better picture would not help increase the sale of video cards.

Having thought out, the top manager suggested trying another approach: to use the tensor nuclei of new video cards to transform the image with low resolution to the picture with high resolution. NVIDIA Research employees promised to think.
A week later, Juan reported that the first experiments show promising results. The head of NVIDIA arranged this: he ordered to add the mention of the DLSS on slides prepared for the NVIDIA presentation at Sigggraph, where the technology was announced a week later.
Apparently, at the time of the announcement in Nvidia, they were not sure that DLSS would be able to work as it should – experiments were carried out only a couple of weeks. But in the end, everything worked out and the company was able to create a technology now used by millions of players.
Read more