empthought

How was it "against their consent?" The Siri AI / Apple Intelligence feature required a user to explicitly enable it. It didn't just get enabled for you.

I know this because I just explicitly enabled it, and now I have the new Spotlight search. I didn't have it before I enabled it.

That's silly. You may as well say Santiago Calatrava isn't an architect any more because he's delegating the vast majority of the architecture work for his projects.

You're not including the physical supply chain energy consumption of distributing video game equipment in this analysis. Nobody ships LLMs to big box stores and tries to sell them to consumers.

You are delegating work and reviewing delegated work, but not managing. No performance coaching, no career development, no mentoring, no conflict resolution, no influencing opinions, and on and on.

I think you’ve forgotten about the “software architect” role.

If all you do is write requests and then have the LLMs spit out the code, then you’re not a programmer, you’re a software architect.

Middle managers have to supervise people, otherwise they are not managers.

The statutory name of this department is the Department of Defense.

I read your book when I was ten years old! I was later stoked to find our computer system at school was a Unix machine, so I could try some of the commands you explained in the story, like ps and sendmail.

“That’s real, and it’s interesting.”

Stop with the slop.

Yeah, NeXTSTEP is notoriously a locked-down, untweakable OS. It's not like anyone was able to use it to program anything of consequence.

We don’t need aggregators like HN pitching AI-authored content, no matter how pedagogically sound. We can direct our own AIs to generate our own AI-authored content.

Stop posting slop.

How about

  for wmv in Path(sys.argv[1]).rglob("\*.wmv"):
        print(wmv, end=" ")
        r = subprocess.run(
            ["ffmpeg", "-i", wmv, wmv.with_suffix(".mpg")],
            stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
        )
        lines = [l for l in r.stdout.decode().splitlines() if "kb/s:" in l]
        print("\n".join(lines) if lines else f"ERROR {r.returncode}")

?

If you go outside stdlib you can use the sh library instead of subprocess.run.