Algorithms and systems that match data with relevant ontologies, which then categorize data into...
So I agree, the "which then categorize data" part of the statement above is what you're after. The part before about those algorithms, systems, ontologies - still a great deal of work to be done, and not a lot of people even aware of the problem set and domain itself. Not even those of us up here in the senior full-stack ether.
AI has a tough time dealing with ontologies. For example, how would things be named in a Github style crafting system? If there are 100 ways to craft a DialogBox then there are 100 possible names! But the internet already uses dots as a way to establish taxonomy (e.g. multix.substack.com) and that may be good enough
Interesting stuff, it appeals to me as someone who learned programming in an interactive environment (Winamp AVS visuals had persistent memory), used Lisp a lot and now works on enterprisey data integration projects
When FPers talk about "immutable functions" they are really borrowing a concept from database theory, namely something called "transaction isolation". But if they agree w that, then they are kinda admitting that categories are more like "in-memory" database transactions. That sort of basic DBMS 101 stuff completely flies over the heads of otherwise intelligent FPers
Algorithms and systems that match data with relevant ontologies, which then categorize data into...
So I agree, the "which then categorize data" part of the statement above is what you're after. The part before about those algorithms, systems, ontologies - still a great deal of work to be done, and not a lot of people even aware of the problem set and domain itself. Not even those of us up here in the senior full-stack ether.
AI has a tough time dealing with ontologies. For example, how would things be named in a Github style crafting system? If there are 100 ways to craft a DialogBox then there are 100 possible names! But the internet already uses dots as a way to establish taxonomy (e.g. multix.substack.com) and that may be good enough
Interesting stuff, it appeals to me as someone who learned programming in an interactive environment (Winamp AVS visuals had persistent memory), used Lisp a lot and now works on enterprisey data integration projects
When FPers talk about "immutable functions" they are really borrowing a concept from database theory, namely something called "transaction isolation". But if they agree w that, then they are kinda admitting that categories are more like "in-memory" database transactions. That sort of basic DBMS 101 stuff completely flies over the heads of otherwise intelligent FPers
This is just a draft - sorry for any glaring typos. Am fiddling w the title for A B testing