Thoughts are somewhat like electric current: prefers lower resistance paths to the ground.
Twitter succeeded because it is the least resistance path to an audience. Blogs in comparison are kiloohms of resistance.
Twitter succeeded because it is the least resistance path to an audience. Blogs in comparison are kiloohms of resistance.
Infact we can generalize this to N ideas cases.
Take two ideas. A and B. Give weight ∆ and 1-∆ to them. Multiple the positions of those ideas with the weights. This gives you a new interpolated position. Now repeat this with new position and idea C. And so on. You will quickly cover the area of convex hull of these ideas.
I think with future advanced LLMs the above scenario is achievable.