Note Chat GTP etc cannot think, it is a LLM it has no independence, it is only a database some of which is curated and works on a very complex predictive text model. It would have been very interesting for him to look at models like Stable Diffusion and the detailed creation of images through complex model and lora interaction. The difference between image generation which is approached quasi scientifically and something like Midjourney which is purely prompt based and less predictable.
So the collective intelligence is just moving faster now, it has found a way to progagate faster through networks and collective memory aka data. Same old intelligence we had before, just less localised.
Note Chat GTP etc cannot think, it is a LLM it has no independence, it is only a database some of which is curated and works on a very complex predictive text model. It would have been very interesting for him to look at models like Stable Diffusion and the detailed creation of images through complex model and lora interaction. The difference between image generation which is approached quasi scientifically and something like Midjourney which is purely prompt based and less predictable.
So the collective intelligence is just moving faster now, it has found a way to progagate faster through networks and collective memory aka data. Same old intelligence we had before, just less localised.
Thanks for getting us Graham again, Erfan! :)
Really interesting talk! And interesting questions.