>>14128422> computer programs improve existing formula's and algorithms by being able to account for a level of detail that would take too long for someone to work out mechanically. Its basically optimization - and dont get me wrong, its useful.Not really. While every operation in a computer is reducible to arithmetic, the mathematics that makes it work isn't just "tedious" - it is literally intractable without the use of computers. Modern engineering involves using computers to solve engineering problems in ways that doesn't just involve speeding up how quickly the pen meets the paper using digital means.
>However, improvements to the computer programs are derived from fundamental science....no, if you go on arXiv, it's usually a mix of people in computer science, chemistry, physics, and statistics. The body of discrete differential geometry used in simulation everywhere was jointly created and studied by the CS and math departments of the world.
>Not the pool of programmers that write these computer programs.writing a computer program and creating the method are two separate activities. Most people who make these methods are a mix of CS or some other scientists. Your mistake is mixing up computer scientists with the programmers...when really, everyone knows how to program.
>But all the AI programmers I know of are scientists first and programmers second.You're talking out of your ass.
Nobody, unless they're a bona fide codemonkey, is a programmer first. This is like the engineer saying they're a wrenchmonkey first. Programming is the most convenient tool you use to communicate or build your ideas.
The fact that you don't understand modern computational science and call it "basically optimization" (SDP and solvers are used...but are hardly the crutch of the problem when you're doing something like, idk, computational electrodynamics to study the E&M relevant to high speed IC design) tells me you don't know a lot.
