This is an interesting crystallization that parallels a lot of thoughts I've been having, and it's particularly hopeful that it seeks to discard the "hacker" moniker and instead specifically describe the subjects as programmers. Looking back, I was only becoming terminally online circa 1997, and back then it seemed like there was an across-the-spectrum effort to reclaim the term "hacker" into a positive connotation after the federal prosecutions of the early 90s. People from aspirant-executive types like Paul Graham to dirty hippies like RMS were insistent that being a "hacker" was a good thing, maybe the best possible thing. This was, of course, a dead letter as soon as Facebook set up at "One Hacker Way" in Menlo Park, but I'd say it's definitely for the best to finally put a solid tombstone on top of that cultural impulse.

As well, because my understanding of the defining activity of the positive-good "hacker" is that it's all too close to Zuckerberg's "move fast and break things," and I think Jared White would probably agree with me. Paul Graham was willing to embrace the term because he was used to the interactive development style of Lisp environments, but the mainstream tools have only fitfully evolved in that direction at best. When "hacking," the "hacker" makes a series of short, small iterations with a mostly nebulous goal in mind, and the bulk of the effort may actually be what's invested in the minimum viable product. The self-conception inherits from geek culture a slumped posture of almost permanent insufficiency, perhaps hiding a Straussian victimhood complex to justify maintaining one's own otherness.

In mentioning Jobs, the piece gestures towards the important cultural distinction that I still think is underexamined. If we're going to reclaim and rehabilitate even homeopathic amounts of Jobs' reputation, the thesis we're trying to get at is that his conception of computers as human tools is directly at odds with the AI promoters' (and, more broadly, most cloud vendors') conception of computers as separate entities. The development of generative AI is only loosely connected with the sanitized smiley-face conception of "hacking." The sheer amount of resources and time spent on training foreclose the possibility of a rapid development loop, and you're still not guaranteed viable output at the end. Your "hacks" can devolve into a complete mess, and at eye-watering expense.

I went and skimmed Graham's Hackers and Painters again to see if I could find any choice quotes along these lines, since he spends that entire essay overdosing on the virtuosity of the "hacker." And hoo boy:

Measuring what hackers are actually trying to do, designing beautiful software, would be much more difficult. You need a good sense of design to judge good design. And there is no correlation, except possibly a negative one, between people's ability to recognize good design and their confidence that they can.

You think Graham will ever realize that we're culminating a generation of his precious "hackers" who ultimately failed at all this?