Speaking as a Fancy Computer Science Professor at a Fancy Institution of Higher Education who teaches the course on Programming Languages:
-
HTML is a way for humans to express their ideas and their intentions in a form that is unambiguously interpreted by a machine. We express our ideas, then turn them loose. The machine's interpretation may diverge from our human understanding; when it does, our ideas talk back to us and they •surprise• us.
If that’s not programming, I don’t know what is.
2/
-
@kdawson OP may have deleted? or locked account due to harassment? Edited with screenshot of text; may remove username if needed.
-
We might draw a line about Turing completeness or intended purpose. Both completely miss the point.
It is fun to try to find Turing tarpits in HTML and/or CSS! But that’s not what makes programming programming.
The previous post is it. The problems of programming — the things that make it difficult, the things that make it rewarding — all come from that collision of human intent with unambiguous machine interpretation. That’s the game right there.
/end
-
Krzysztof Sakrejdareplied to Paul Cantrell last edited by
no one in particular: ...
me: I could argue HTML is also #AI, we just ask it to lay out documents using simple phrases and through the magic of parameters learned from decades of design experience the HTML interpreter presents a nicely optimized presentation. It's basically magic! -
@inthehands 100%! LaTeX is programming, HTML is programming, Markdown is programming, teaching your dog to do tricks is programming. Not all programming is writing kernel drivers
-
@grwster
I mean, making the mistake of taking your cute remark seriously: the dog isn’t unless it’s a robot dog; that's teaching! One of the things that makes programming programming is the machine and its unambiguous interpretation. Teaching, however…that’s a whole other kettle of fish of a different color! -
Paul Cantrellreplied to Krzysztof Sakrejda last edited by
@wronglang
I mean, yeah, there’s something to the idea that LLMs are not so different as we think from previous attempts to get machiens to follow human instructions!The actual definition of AI is “thing that humans can be good at that computers were recently bad at,” so anything is at most •temporarily• AI once a computer gets good at it!
-
Krzysztof Sakrejdareplied to Paul Cantrell last edited by
@inthehands the definition of AI is disappointingly thin these days
-
Paul Cantrellreplied to Krzysztof Sakrejda last edited by
@wronglang
I stand by mine. It’s the only one that’s held up over my years in the biz. -
@inthehands fair enough, though teaching a dog a trick is about capturing behaviors and putting them on command, and then chaining those commands, which is more like programming than teaching in my book
-
Krzysztof Sakrejdareplied to Paul Cantrell last edited by
@inthehands I have gone through the exercise of trying to talk the chatgpt instance the university of Michigan makes available to its staff into terrible ethical positions and I gotta say it performed better (on, e.g.-Palestine) than a lot of the discourse in the mainstream news media so maybe there's some value to having a machine that produces average text... at least in the context of.. whatever we're experiencing right now...
-
@grwster
I mean, maybe could one say that this kind of training — dogs, horses, soldiers, orchestral musicians, surgical assistants*, anything where you’re trying to teach a living creature to follow instructions with near-mechanical consistency — is teaching the student to be a machine, so that they can be “programmed” for the day’s task?* (no disrespect to any of these roles / jobs, all demanding!)
-
Paul Cantrellreplied to Krzysztof Sakrejda last edited by [email protected]
@wronglang ¯\_(ツ)_/¯ If you’re looking for output that hews to norms, then it’s no surprise that pattern-matched plagiarism from a broad database of human inputs can provide it! And since that’s something that computers were recently bad at doing, then it does qualify as AI. For now.
-
@inthehands orchestral scores / sheet music / guitar tabs, definitely programming… I’m a flawed machine when running those programs, but that’s not the programmer / composer’s fault
-
@grwster
We're wandering now, but heck with it:I have a whole soapbox about how we sometimes misinterpret musical scores as being programming when they’re not. There’s a whole spectrum of •how a score carries meaning• that depends on the types and degree of interpretive liberty the score affords within its musical-cultural context. Some scores are basically MIDI by other means; some are only the loosest framework for improvisation.
-
@grwster The classical world I come from has IMO erred since the mid 20th century much to far toward seeing scores toward the “MIDI by other means” end of the spectrum, particularly with pre-20C music. I’m pretty sure that if we could hear a recording of, say, Chopin or Bach playing their own music, a modern classical musician’s immediate reaction would be “No! That's wrong!!”
-
@inthehands point taken, but circling back to HTML, a certain amount of interpretation by the rendering engine is allowed… so, are programming languages by definition always tightly specified?
-
Oliver Jensenreplied to Paul Cantrell last edited by [email protected]
@inthehands people like to think programming is while loops and if conditionals. I like to remind people of their programmable TV remotes.
"But that's not real programming" my brother you co-opted the word.
-
@grwster
If so, then C is not a programming language! -
@wronglang @inthehands How your computer works under the hood is pretty wild. Branch prediction blew my mind, makes determinism an interesting notion.
