4 ms·

Yeah but more and more project would get written in another language (Go,Rust etc) that would dominate next 20-50 years when this language would become "legacy"

• AtlasBarfed 1y ago
  Llms should be ushering in an era where writing any halfway decent code in a mainstream language should then enable porting that code to virtually all other languages while maximizing the standard library or language features. Alas chathippityi can't even port something as enduring, test exampled, source available, and well documented as SED. There is a little bit of depth to something like that, but it should be well in the wheelhouse of the claim capabilities of llms. Alas, I get nothing but the most superficial of flag capabilities and sed capabilities. Effective porting should be a prime capability, because in general I would consider that to be breadth-based knowledge Rather than involve deep language knowledge. There was a wiki style sit, I believe it's called Rosetta or Rosetta Stone that attempts to do this before the age of llms. And really that should be the low-hanging fruit for some stage of llm maturity that it can do that and renders that site obsolete. I'd love to hear opinions to the contrary, but it appears we are pretty far from that