Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

Публикация

Предыдущая публикация

(Well, *might* be a problem, depending on what amounts to a sniff test dressed in formality. Which isn’t to say it’s a bad standard.) Anyway, that’s a bit of a tangent. I wish we had more detail on the core story here. One question: how is the author supposed to “substantiate” non-use of AI?

2282

I’m probably further down that road than most - I’ve basically put together a non-Microsoft, non-Apple, non-Google software environment to write my next book in, in part so I can say clearly that I’ve routed around AI, and I couldn’t *prove* non-use.

Обсуждение

Ответы

I did that by using Linux. All free software. Lots of support for writers.

Ответ для jacques124

Indeed. I decided I wanted the somewhat unusual Qubes distro, which is an adventure in itself. But while this allows us to avoid AI, it doesn’t allow us to *prove* that we’re not using it. That is borderline impossible.

Oooh can you share what you used?

Ответ для Aliette de Bodard

Yes please

Ответ для Aliette de Bodard

Cheap mini pc from eBay plus Linux Qubes (super-secure, supposedly finicky, though I rather like it - but other distros would do) plus Zettlr and LibreOffice. Today’s task: export Ulysses docs to Zettlr, set up backup and syncing.

Write your manuscript with a quill pen on hand made parchment paper

Ответ для Ian McQue

I hope my fountain pens and notebooks will be sufficiently antiquated? I don't think I've got enough spare time to get into making paper, and I already have a shelf of blank notebooks to work through...

Ответ для Ian McQue

I seriously considered an IBM Selectric. I learned to type on one of those; how hard could it be? (Very. Total pain in the arse, no backups, manual page numbering, physical cut n paste…)

I’m not about to document my own keystrokes or record my work sessions with a video camera, and even if I did, that wouldn’t get it done.

Ответ для Nick Harkaway

Putting the burden of proof on the other side is equally hopeless. AI detectors don’t work to a certainty. It’s a recursive mess. Sidebar: AI models rank their own output higher than that of other models and humans. (That‘s tautologous, surely, if not intuitive.) arxiv.org/pdf/2509.00462

arxiv.org