I put Rayuela down, probably for good. I’m…sort of bored with it, and my language skills are probably not where they need to be to pick up on the subtleties. I started re-reading a collection of Philip K. Dick’s short stories which included a letter from him to another author explaining what he thinks science fiction is:
We have a fictitious world; that is the first step: it is a society that does not in fact exist, but is predicated on our known society; that is our known society acts as a jumping-off point for it; the society advances out of our own in some way, perhaps orthogonally, as with the alternate world story or novel. It is our world dislocated by some kind of mental effort on the part of the author, our world transformed into that which is not or not yet. This world must differ from the given in at least one way, and this way must be sufficient to give rise to events that could not occur in our society - or in any known society present or past
For PKD, setting a story in the future isn’t enough to make it science fiction, and this had me mulling over the dislocation he talks about, which gave rise to more woolgathering, until I landed on a what-if scenario:
What if Eternal September had never happened, and moreover, never could have happened? That is, what if the internet as we knew it in 1993 had remained the exclusive domain of academia, government, and large enterprises?
Remember that at that time, there were bulletin boards, many joined to message-passing federations like FidoNet, and larger private online services like CompuServe. If those had remained the only options for connectivity, what would the world look like in AD 2025?
Just for the fun of it - and fully conscious of the irony - I put the question to both ChatGPT and Claude and got back answers which, to be honest, were a little depressing in that ‘homesick for a place which doesn’t exist’ sort of way. Herewith some of their observations, and it’s pointing out that without the internet corpus, it seems pretty unlikely that LLMs would have emerged at all. AI research would likely have remained focused on things like genetic algorithms and neural networks.
Without internetworking between commercial services and the academic internet, we’d have multiple incompatible digital ecosystems. CompuServe might have evolved into a corporate-focused network with robust business applications, while AOL could have remained a consumer entertainment platform. GEnie, with its origins in General Electric’s timesharing systems, might have specialized in technical and engineering communities. Each would have developed proprietary protocols, incompatible file formats, and distinct cultures.
Local bulletin board systems would have become the primary bridge between these isolated networks. By 2025, BBS networks might have evolved into sophisticated federated systems - imagine FidoNet scaled up with modern hardware but retaining its store-and-forward architecture. Local sysops would function as information brokers, manually transferring messages and files between incompatible commercial services.
This fragmented world might have been more sustainable - each network optimized for its specific community rather than trying to be everything to everyone - but it would have severely limited global knowledge sharing and collaborative innovation.
In short:
- No Eternal September → slower, more elitist Internet, delayed mass adoption.
- BBS-first world → more fragmented, localized, community-governed networks; less centralization; a digital culture shaped more by sysops than by Silicon Valley.
By 2025, you’d probably have a digital world that’s:
- Less commercialized, more community-run.
- Slower and more text-heavy.
- Rich in local cultures, but without a single meme-flooded global commons. Basically, closer to Ham radio + local newspapers than Twitter + TikTok.
I graduated from college in 1993 and used the text-only internet (newsgroups and email) a fair amount while I was a student. I remember seeing one of the guys in the computer lab show us a new program called Mosaic, but there wasn’t much you could use it for at the time. Archie and Veronica were way more useful. Right before graduation, I got a part-time job helping a market research company run consumer tests for a CATV set-top box, which offered electronic messaging, interactive menus, and so on. The big idea at the time was “convergence,” where your telephone and television would all come on a single wire, and the fight everyone was preparing for was whose wire: The telephone company? Or local broadband CATV franchisee?
Good grief, it sounds like a discussion of steam vs. horse now. Still, if the walled garden online services had remained isolated and things like FidoNet continued to develop as a sort of backbone, we would have lost the immediate access to (nearly) unlimited information, but we’d also be spared the global bathroom wall of opinion and invective. As for the internet, it might have remained a B2B sort of thing, or a place for enterprise applications. Websites might be limited to catalog storefronts for large retailers, but maybe they would have cut deals with AOL and GEnie instead. Social media would not have emerged, but there’d still be localized enclaves and cliques. So memes, maybe, but slow-moving and regional at best. Newspapers would probably still be around, and maybe CATV/minitel-like services emerged for the non-computerized households.