Nice!

2 comments:

-- Agree strongly with LLMs not being the only "AI" in the mix (remote, local, whatever). In fact, I think there are tried and true methods to implement decisioning using optimizers and linear algebra that can help as well... LLMs are NOT the only hammers for so-called AI nails (and we should not label all things "AI”).

-- Going full-on autonomous... (a) requires a decisioning engine into which we pour investments to understand what decisions can be automated in some way, and where the balance of humans in the mix is needed to connect dots that statistics won't; and (b) full-on autonomy ... the question itself ... is it possible to achieve it? Do we want it?

I believe that the automations made possible by rear-view-mirror AI will unlock potential to do both old things in new ways, and to do new things we couldn’t have done previously. Some but not all of this is drone work (where autonomy fits), and we won’t have to do it anymore — we’ll have more time for new things that matter.

I believe that the human imagination is able to consider combinations and compositions differently than AI, which will lead to assembling and creating as with legos at one’s whim and wish.

I believe that software development will change -- to say otherwise would protest the obvious. BUT I don't believe there is a finite set of problems in the world that will be solved once and for all by an ever-maturing statistical means to connect dots. Further, I think we're on the cusp of democratization of imagining and creating with tools that help us get beyond the limits of our skills sets as they currently are, and past the siloed methods that we've used to date that will — if not already have — become outdated.

I choose to be optimistic … with my eyes as wide shut as I can keep them … and I hope to work with people having curious minds (like yours) to see just what can be automated and what can be used to amplify us as humans ... somewhere in the middle will be a beautiful balance ...