Roses are red...and orange, and blue, and
=========================================
We went to the "international rose test garden" today that's downtown
and then on a long walk through the giant forest to see the big stand of
redwoods.
I may be, as people say, a Professional Computer Toucher but I never
feel as at home as when I'm in the woods.
I don't feel like these are contradictory. Computation---at its
best---is an extension of mind and body, a way to extend our ability to
think and communicate and observe the world around us.
It's not weird to feel at home in the loud-quiet of a forest and also
use those tools to think any more than it'd be a contradiction to use
pens and paper while hiking.
I want to find ways to convey some of these thoughts when I'm teaching
computer science. My job, strictly speaking, is to prepare students for
very particular curriculum so they can transfer to 4-year schools
well-prepared.
But I'd be lying if I didn't think the unspoken part of my job is to
give burgeoning programmers and computer scientists a healthier view of
what computers are good for, especially as the current trend is "turn
over more control and autonomy to janky machine learning interfaces".
If I have to read one more "is programming obsolete now that we have
LLMs!" thinkpiece I'm going to lose it.
What's frustrating is that I don't even think things like LLMs are
innately bad. They're just not good for the things that capital wants to
use them for.
I have a project on the backburner about using local, smaller, LLMs with
students and showing them what they're good for and what they aren't.
The tl;dr is "don't ask qquestions, use it to help process large
quantities of information for a first pass". They can actually be
fantastic learning tools that way, but you wouldn't know it from the
hype.
The more students I can convince to use their knowledge for convivial
means and not to be the arms manufacturers of modern capital the better.