Conlang thread

Have any of you tried learning a constructed language? I remember being autistic enough to try lojban at uni, I actually had a friend who did the same and we would have a circlejerk.

>>1
Lojban is not perfect, an AI(neural network) would design a better language without
any preconceived notions: syntax-optimized easy-as-esperanto as-precise-as-lojban as-dense-as-ithkuil,etc.
Maybe even add some viral ideas(making words sounds reflect semantic values more precisely) to insert utility loanwords into English
to allow it naturally replace complex slang.
English is quite ripe for neologistic hostile takeover, i'd say such a
language can be subverted(like Newspeak) and even 'upgraded' with
a sufficiently smart AI trained to devise vocabulary with a basis of more
effective language(not Snow Crash fan, but biologically we're monkeys trying to devise vocalizations of complex mental machinery and monkeys are not pinnacle of technology - we're not introspecting our brain but evolving means to do so..an AI brain mapper would allow precise semantic 1:1 map to concept-word dictionary foundation to create a hyper-precise language referencing the exact brain mental process/network activated by the word)

>>7
I don't give out any concrete ideas, just hints at what design
i'd like to see in future: software-defined languages,
logical brain-computer interfaces that map exactly without 'interpretation'
and language evolution guided by AI models not hobby linguists
spending 10 hours to think of a precise meaning of 'set'.
Instead of programming AI natural language processing,
upgrading the language to a level of precision inherent in programming languages structures where AST of text
can be modeled as an exact graph with 0 ambiguous content(no garden path sentence, no alternate meaning, context-free structure).
Such language will be superior within human-to-human communications
too by eliminating dependencies on context and reducing semantic space of meanings to exact concept references.