I’ve been telling my friends that 5-10 years from now, the only things that you’ll be able to ~100% trust is what happens in front of your eyes, in that very moment. You can elect to trust reliable news organizations to vet things for you, but even if you do, due to polarization a huge subset of the world will think you’ve been got and discard everything as fake.
Look at stuff like Sora, or all the new voice models coming out. Just a few days ago there was a high school athletic coach (!) arrested for cloning the principal’s voice using that to say vile stuff. He only got caught because he used his own e-mail.
Now combine that with the fact that Microsoft’s new Phi-mini model approaches GPT-3.5 performance using 3.8 billion parameters, whilst GPT-3.5 uses 175 billion. And we’re only ~5 years of optimization into this tech.
Good point. When I was first exploring the Bible, I'd get bored with the detailed lists of people who saw things happen. Took me years to realize that the lists were people who were still alive at the time and could provide verification.
We already know you cannot trust what you see with your eyes (check any "compare eye-witness reports with trusted video recordings" or watch Penn & Teller).
Ghost in the shell anticipated so much. The premise of the first movie is that a model gains sentience (creates its ghost out of thin air), claims asylum, and the model merges itself with the ghost within the major.
Everything in that series depicted is stuff that either exists today or will exist within 30 years. Everything.
You think that the gap between "a meat robot" and "a meat robot of a specific shape" is that big?
(If you're thinking of the robots' brains in the TV show, sure, that's a totally different challenge; but you don't need them for a replacement-takeover scenario, there's a lot of other ways to do that).
AR has better real world capabilities and use cases than VR. AR is what you use to enhance reality or make it easier. VR is where you go when you want fiction.
Aside from entertainment, I don't understand this VR push when AR is arguably better for many of the non-entertainment uses
Wouldn't AR be as susceptible to a sort "real world" injection attacks as well? Vision Pro uses cameras to record the outside world and display it back to the user, it should be relatively easy to add/remove objects in the rendered feed.
"I’ve been telling my friends that 5-10 years from now, the only things that you’ll be able to ~100% trust is what happens in front of your eyes, in that very moment. You can elect to trust reliable news organizations to vet things for you,"
The time is now. Even the mainstream news organizations are probably high 90s% reliable as they've been caught selectively editing, not vetting sources or facts, and displaying biases.
I’ve been telling my friends that 5-10 years from now, the only things that you’ll be able to ~100% trust is what happens in front of your eyes, in that very moment. You can elect to trust reliable news organizations to vet things for you, but even if you do, due to polarization a huge subset of the world will think you’ve been got and discard everything as fake.
Look at stuff like Sora, or all the new voice models coming out. Just a few days ago there was a high school athletic coach (!) arrested for cloning the principal’s voice using that to say vile stuff. He only got caught because he used his own e-mail.
Now combine that with the fact that Microsoft’s new Phi-mini model approaches GPT-3.5 performance using 3.8 billion parameters, whilst GPT-3.5 uses 175 billion. And we’re only ~5 years of optimization into this tech.
I want to get off Mr Bones’ wild ride.