An essay mourns the last people who understand machines. Every generation has said that
The blog post argues the loss isn't competence but acquaintance: the intimate knowledge you only get from a machine that pushes back. In 1995 a beige PC wouldn't run a game until you'd rearranged its bits by hand.
A programmer's essay mourns a vanishing kind of knowing: not skill, but the acquaintance you earn from a machine that resists you. "You only ever know the things you can lose to." AI can recite every manual flawlessly, he grants, yet that isn't the hands-on understanding built from struggle.
The thread turned on him. A programmer in Paris: back in 1990, classmates already called someone else to fix their PC. Another mapped the same nostalgia onto cars, telephones, electricity, then shrugged: "We could do this forever."
"You only ever know the things you can lose to"
The essay is a lament for a specific kind of knowledge. Not competence, the author insists, but acquaintance: the intimate understanding you only earn from a machine that fights back. In 1995 a beige PC wouldn't run a game until you'd edited autoexec.bat, freed conventional memory, set the right interrupt address for the sound card, and listened to the modem shriek through its handshake. The friction was the whole medium. "A machine that cannot challenge you is a thing you cannot know."
His worry is not that skills die. It's the double bind he names in three words: "More dependent than ever. Less acquainted than ever. Both at once." AI models, he concedes, have read every manual no human reads and will recite exactly how the machines work, flawlessly and forever. But reciting is not knowing, and the machines no longer need us to understand them. The generation coming up will never miss what it never had. The burden of noticing the loss falls on the last people who knew.
"We always were the only people who ever knew how it worked"
The thread did not accept the premise. The sharpest pushback came from someone who was there before the golden age supposedly ended:
"In 1990 people fellow students called me to fix their computer, they had absolutely no idea how any of this worked… their reaction wasn't to fight it, or understand it. It was to call someone to do it in their stead." — a programmer in Paris, France
Others picked at the founding example. A commenter who grew up on Sierra games recalled the whole ritual being a single install disk: "Most I had to deal with was installing the drivers for my SoundBlaster card." Another noted the essay's nostalgia is really DOS-specific: getting a game running on a C64, Spectrum, Amiga or Apple II "didn't involve understanding the computer" at all.
The most economical dismissal reframed the whole genre of complaint. Swapping computers for cars, one voice showed the argument runs forever:
"It is already shocking to a certain degree seeing very young people not being able to use a vehicle in the narrow sense because all they ever learned were the mechanical controls of the so-called automobile." We could do this forever. — PaulDavisThe1st, maintainer of the Ardour audio workstation
That historical shrug got its fullest form from a high-karma old-timer who ran the tape back a century:
"This is sort of the story of the telephone system of the 1950s-1970s, or electricity in the early 1900s, or cars from 1950-1980s… As the technology matures, it gets hidden away to become increasingly invisible to the end user." — nostrademons
Where the essay's optimism about AI cracked
On one point the room split hard. The essay's line that "the knowledge is not in danger, in fact, it has never been safer" because AI has read every manual drew a flat contradiction:
"LLMs do not, and cannot, actually know a single thing. They are a statistical model, not knowledge… you need someone with actual knowledge to be able to recognize the BS and correct it. But we are losing the knowledge." — an engineer warning of a machine-god future
A counter-voice called that FUD: "LLMs do hold knowledge empirically. They sometimes give out wrong information, like humans… Things have always been like this. It's called specialisation." And one person answered the whole debate with a Wikipedia link to Asimov's novella Profession, where people are educated by having knowledge tapes fed directly into their brains and originality becomes the rarest skill of all.
The cynical read arrived too. An engineer sketched a near future of people "goading AI to right direction by repeatedly changing what they ask slightly," memorising which incantation worked without understanding any of it: not knowledge, but cargo-cult prompting. Against all of it, the maintainer of a Lisp compiler offered the only real reply to inevitability: "you individually can choose whether to lose control or not." Programming languages, UNIX, debuggers aren't going anywhere. The acquaintance is still there for anyone stubborn enough to want it.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
This is the last story in the July 1, 2026 issue.
Go to the issue archive