Things are indeed moving really fast now, and that's really exciting, but I'd be cautious about extrapolating that pace too far into the future. Nearly all of t

Things are indeed moving really fast now, and that's really exciting, but I'd be cautious about extrapolating that pace too far into the future. Nearly all of the recent image generating AIs are all based on a single breakthrough technique called image diffusion[1]. Once we start hitting the limits of diffusion models progress will likely start to plateau a bit, at least until the next major breakthrough. [1]: https://www.assemblyai.com/blog/diffusion-models-for-machine-learning-introduction/ https://www.assemblyai.com/blog/diffusion-models-for-machine...