Yes - also the future of Paul McAuley's Cowboy Angels, which I suspect was an important influence on The Peripheral (Gibson mentions McAuley in the acknowledgments, though not for this book]

basically unavailable in the US right now, which is an enormous shame.

[also, while we're on the topic of older SF that is unexpectedly on-the-nose, yesterday finished a re-read of @unlikelyworlds.bsky.social Cowboy Angels, about an America that semi-colonizes alternative versions of itself, including a former Nazi US once run by a "Dear Leader" with three idiot sons.]

en.wikipedia.org/wiki/68_Mill....

[eaten by owls]

My favourite example (formatting rendered wonky by time) is the old advice to use LaTeX rather than word in working papers to make it look as if you're intelligent goodauthority.org/news/the-gar...

probably much more than you wanted - but the Campante et al. paper is likely useful in thinking through the broader problem and how journalism could respond. [recipes are different - the Bloomberg piece doesn't really talk about copyright problems or existing pre-AI weird SEO padding etc]

LLMs make it much more difficult than in the past to distinguish sincere hard effort from automated boilerplate, so that it is harder to tell a sincere boss committed to helping you work better from one who is lazy and couldn't care less about you. There are similar dynamics for other situations.

The intuition is crudely that -say, if a boss gives you a negative annual review, you will treat it differently if you think the boss spent a lot of time and sweat working on the report to make it thoughtful and constructive, than if you think the boss is just throwing boilerplate at you.

Completely separately, there is a very interesting Wojtowicz and DeDeo paper - arxiv.org/pdf/2407.14452 - arguing that LLMs make it harder to signal sincere willingness to cooperate across a variety of social situations, by making it cheaper to send previously costly signals.

There is some interesting recent research from @filipecampante.bsky.social et al. suggesting that the market for news may resemble a separating equilibrium to some limited extent filipecampante.org/wp-content/u...

distinguish themselves from slop providers, leading to a general degradation of information. A separating equilibrium would be one in which reliable sources could send costly signals, and in which news consumers would select reliable providers, creating a market for reliable news.

If the setup means that honest types will send costly signals and bad types will not, you end up in a separating equilibrium where you can tell one kind of player from another. So in this context, a pooling equiibrium would be one where reliable sources don't have means/incentive to ...

depending on setup, the other players may send costly signals. If all types of players have incentives to send the same kind of signals, whether they are diligent or feckless, you end up in a pooling equilibrium - the signals don't have useful information that allow you to distinguish good from bad

sorry - game theoretic jargon. In signaling games, you might want to distinguish between 'types' of players - for example, are workers lazy or hard-working? Or are news-slingers unreliable or diligent? You don't know the type of other players, though the players themselves do and can communicate.