While this blog has so far been mostly timeless, I'm going to comment here on a current event: an article arrived today in my mailbox, titled "The End of Programming"* and subtitled, "The end of classical computer science is coming, and most of us are dinosaurs waiting for the meteor to hit," written by Matt Welsh.
I'm arguing that the use of PicoStack to develop small web applications could continue, long after most computer applications are developed using artificial intelligence (AI) techniques, and code is being written routinely by AI agents instead of humans.
Agree
First, I agree with Welsh that computer science (CS) as a discipline changes over time. Thirty years ago, he points out, "CS [was] taught as a discipline with data structures, algorithms, and programming at its core" and still is today, but he doubts that "in 30 years, or even 10 years, we are still approaching CS in this way."
As an undergraduate twenty years earlier than him, so fifty years ago, I remember that those three topics were the core, but we also studied the use of integrated circuits to build devices, the writing of compilers, and the internal workings of operating systems. The latter three topics were considered important then by the department but are gone now. So, yes, the discipline does change.
Few of today's students would know (or care) about the design of digital devices, how to write a compiler, or exactly how the operating systems they use work. Most do not even host their own domain and websites‡.
So yes, many of today's CS students will likely switch over from writing code to training deep learning AI systems during the course of their career.
and yet also disagree
This is still a human world, and things just don't change that quickly. As a former close colleague (now passed) used to say, "the world is a low-pass filter"§.
things change slowly
The world of business still depends on COBOL programs, many written over fifty years ago. Perhaps these will someday be maintained by AI coding systems, or replaced, gradually, by training deep learning systems. In the meantime, there will continue to be a need for people who can read and write code.
there is a long tail effect
While modern, large, popular web applications may come to be written in new ways, and that rather quickly, there is a long tail of web applications that will continue to be written as they are today (perhaps using PicoStack) or even written as they were thirty years ago (when PHP was first developed; it is still in large-scale use (just recently noticed that even facebook uses it in some places❡)).
For every highly successful web application there are tens of thousands of smaller ones. That is the long tail, and PicoStack can play a role therein, for a long time to come.
there are hobbyists
In the previous post, Managing channels, mention was made in a footnote, of a story from 10 years ago, highlighting a solution to the problem of Forgetting to Water the Christmas Tree? How Personal Clouds Saved Christmas using techniques similar to PicoStack.
The protagonist "already had an Electric Imp connected to a Vegetronix Water Sensor using a breakout board. Further, he was able to sense the water level in the stand. But he wanted to receive a text when the water level was below a certain threshold." Of interest is that the three devices mentioned (and linked to from the original story) are still commercially available today. The remainder of the solution involved using a "personal cloud" (today we'd call it a pico) to provide the notification. That part, the personal cloud, has not remained available, but can be provided using PicoStack.
Conclusions
Yes, things are going to change. But they will change slowly, and opportunities are rife for using PicoStack to develop niche web applications using a bit of coding, and many people, including readers of this blog, will use PicoStack and other venerable technologies to do just that. For years to come.
Notes
* Matt Welsh is a former professor of computer science at Harvard University. The article appears in Communications of the ACM†, January, 2023, page 34. In a spirit of full disclosure, Welsh recently founded Fixie.ai, a "startup developing AI capabilities to support software development teams," and so has a vested interest in the change he describes in his article.
In the same issue, on page 27, labelled "Coming Next Month," an article "The Premature Obituary of Programming" is promised, so this is clearly a debatable question.
† ACM is the "Association for Computing Machinery, Advancing Computing as a Science & Profession".
‡ an informal survey of a few hundred current CS students showed only about 3 percent operate a personal website.
§ a low-pass filter allows slower changes while actively resisting quick changes (in Scott's analogy).
❡ last week, my sister emailed a link to an obituary, copied from her facebook feed, and it begins https://l.facebook.com/l.php?u=… followed by a URI encoded destination URL and hundreds of characters of tracking tokens.
No comments:
Post a Comment