Hmmm. There has been a lot of progress in producing high quality videos of totally manufactured (fake) content, using machine learning and lots of processing power. This can have really disastrous consequences in the wrong place and time. How do you identify fake content? Why using machine learning and lots of processing power, of course.

This reminds me of military design specifications in the recent past. Think Us versus Them (USA v. Russia) circa 1970 or so.

  • The bullet from our standard issue rifle has to be powerful enough to penetrate the helmet of our most likely enemy.
  • Our helmets have to be strong enough to prevent being pierced by the bullet from our likely enemy’s standard issue rifle

And guess what, our likely enemies have exactly the same requirements. Kind of a no win proposition…