The kakistocracy’s Crime Syndicate arm, among other dirty tricksters, now has at its disposal new cutting-edge technology for creating fake personas. Nvidia researchers posted details of the method for producing completely imaginary fake faces with stunning, almost eerie realism. Here’s the paper. Coupled with computer generated images (CGI), one might as well largely disregard almost all that emerges from the Lugenpresse and Brave New World sources.
It’s called a “generative adversarial network,” or GAN. GANs employ neural networks to train a computer to learn the nature of a data set well enough to generate convincing fakes. When applied to images, this provides a way to generate often highly realistic fakery. Nvidia’s promo video demonstrates the new technology.
Researchers took inspiration from a technique known as style transfer to build their GAN in a fundamentally different way. This allowed their algorithm to identify different elements of a face that “researchers” could then control. The approach can also be manipulated and remixed using varied elements, such as age, race, gender or even freckles. Some results are shown below.
Adobe also has a project that uses GANs to improve the realism of images after they have been manipulated. GANs can also be used to sharpen up degraded images or older video. Of course, this could be used to ever-so-slightly alter to appearance of real “crisis actors” in future visual presentations, or eliminate the need for them altogether.
Read “Did Rochester Institute of Technology Class of ’07 Grads Provide Fake Personas for the Vegas Shooting?”
Psychologically, this advancement should be a slam dunk for pulling off hoaxes, blackmail, compromising people, assassinations, false flags and black ops. “Kayfabe” can now be enhanced. This is a term associated with pro wrestling, meaning presenting staged performances as authentic or real. Fake drama, fake tension, fake feuds. Jeff the pro wrestling fan explains it.
It’s fertile ground as several generations have already been brain warped through video gaming and “reality” TV to the point that they can scarcely discern CGI fiction, or what Winter Watch calls the “Cartoon World.” And most of the population either are unable to apply critical thinking or are too naive to process even glaring examples of skulduggery.
Motherboard discovered a Reddit editor (redditor) named “deepfakes”. Just a random dude who happened to enjoy playing the swap game with celebrity faces with the help of publicly available porn videos, a machine learning algorithm and his trusty home computer, of course. He made several convincing porn videos of celebrities, including Gal Gadot, Maisie Williams, Taylor Swift, Scarlett Johansson, Maisie Williams, Daisy Ridley and Aubrey Plaza.
This amateur was quoted as saying, “I want to improve it to the point where prospective users can simply select a video on their computer, download a neural network correlated to a certain face from a publicly available library and swap the video with a different face with the press of one button.”
False evidence needed to drum up star chamber crime charges? How much easier has blackmail and compromise become — as in “Hello, we taped your private intimacies and we know what you did.”
Who needs elaborate old-school compromising operations like Jimmy Savile, Craig Spence and Jeffrey Epstein when you have this tool? And we already know who’s minding the hen house.