Short answer that doesn't involve a video: DLSS (Deep Learning SuperSampling) is a technique created by Nvidia, that uses a specially created AI model accelerated by specialized hardware called Tensor, to improve the render resolution in games with no loss of performance. Or, to gain performance for the same render resolution.
Nvidia are the current hardware partners to Nintendo for creating the Switch consoles, and are most well-known for creating powerful graphics cards for desktop PCs, and efficient Tegra system-on-chip solutions for mobile devices like the Switch.
So the practical upshot is that Nvidia, the makers of the hardware powering the current and the next Switch consoles, are in possession of technology that can radically improve either performance, or resolution, of games running on the next Switch console series, and there is a distinct possibility that this technology, that is already known to work with mobile Tegra hardware, will be present in the successor to the current Switch console. Allowing, theoretically, to have it match the effective performance of the current-generation high-power consoles like the PS4 Pro and Xbox One X - not through raw power, but efficient use of DLSS technology.
The actual likelihood of DLSS being present in the next Switch console is, admittedly, somewhat low - Nintendo is not known for jumping onto relatively unproven tech bandwagons like this. But given how much of a cost-to-performance benefit it is, its eventual inclusion in some next iteration of the console is almost guaranteed.