There is probably Nobody right now can say where the current gpt approach saturates and what potential limits it has due fundamental limitations in either gradient descent based technologies, or GPT architecture.
Therefore it's impossible to extrapolate what gpt-x with (x>4) might be able to do.
Despite the immense progress amd many use cases we are currently in a booming industry and that means wild marketing claims, exaggerated expectations and grifters.
If you have any more data I'm looking forward to be corrected on this.
the first iPhone was impressive, it showed what the future would be. Then iPhone 3G was also a massive leap forward, it brought us the App Store. iPhone 4 was pretty big, FaceTime.
After that nothing has really changed (I'm on a 7 and the camera actually beats the 14's in a side by side comparison, at least in some cases). I imagine GPT will be similar.
Though I don’t anticipate any doom. I think it will force a return to trusting only what you can personally verify. That’s a damn good thing. It’s only very recently that what a random person across the world proports to be true is instantly subjected to the internet rage machine hype cycle of all of humanity. It’s pretty clear now that was a bad idea, made even worse now that everything is so easily fabricated.
As an aside, It’s crazy how recently lately I’ve been told not to trust my own observations and instead must believe “the science” that XYZ media conglomerate is pushing. Hopefully those days are ending too.
