Its never 1, Its always Two
Ironically, one number was written numerically and the other verbally.
Lately i have been performing extensive research work, and its always at some edge that science did not investigate. These edges were not investigated because of their complexity to say, but rather due to their low perceived ROI. There were not many use cases if such boundaries were explored and pushed, they somehow been perceived as human-technology limits. Meaning we cannot gain any value from them unless we discover new technologies that can draw out value from their results.
Saying all that, you must want to know where i have been digging my feet. In which muddy sands...
Education, i have been trying to realize the unexplored value AI creates in the educational universe. Its very weird, other industries have direct use cases, but education seems hard to crack. Every effort leads to the same LLM as wrappers with no realized value that can be truly captured.
Looking at it from afar, the normal layman will shout use llms to teach. However, this is only a foreground visible use case that when implemented yields limited value. The true value lies deeper, its a democratization and time compression event that happened. We can generated high pedagogical value knowledge if we construct the correct pragmatic foundation which can only happen if we have a large work force of pedagogical workers... LLMs. This goes backwards, if we have a large working force with no foundation, then our outputs would be worthless.
Its not the same, studying with a stochastic bird is different than having that bird navigate a per-determined map of the closed set of all realistic possibilities. The set can only be constructed through a scientific process of synthesis, observation, finally curation.
Two can be found, in fact i saw it!
Whenever i built my hypothesis and reach a point where i must decide on a construct to proceed with, all the constructs had restrictions and limits which is normal. However, its only normal if we look at such decisions as irreversible and costly. Irreversible meaning if we choose it then the rest of the hypothesis must be studied in a specific way, and costly meaning if we try to change it later it will affect all downstream processes. This is exactly the thought process that i had to change. I had to start thinking not as a human with limited time and effort, but rather as a humanoid, clone, entity, whatever you want to call it... that can perform all possibilities and test them. In fact, why test each one when you can merge all together. My hypothesis can contain many merged decisions that compensate each other's negatives.
This way of thinking went further, i was able to accommodate many structures together, many process together, etc...
I realized i don't have to think of ONE, when i can have 2.