Accelerando

"There's a network on the other side of it. We were told it was FTL, instantaneous, but I'm not so sure now. I think it's something more complicated, like a lightspeed network, parts of which are threaded through wormholes that make it look FTL from our perspective. Anyway, Matrioshka brains, the end product of a technological singularity — they're bandwidth-limited. Sooner or later the posthuman descendants evolve Economics 2.0, or 3.0, or something else and it, uh, eats the original conscious instigators. Or uses them as currency or something. The end result we found is a howling wilderness of degenerate data, fractally compressed, postconscious processes running slower and slower as they trade storage space for processing power. We were" — she licks her lips — "lucky to escape with our minds. We only did it because of a friend. It's like the main sequence in stellar evolution; once a G-type star starts burning helium and expands into a red giant, it's 'game over' for life in what used to be its liquid-water zone. Conscious civilizations sooner or later convert all their available mass into computronium, powered by solar output. They don't go interstellar because they want to stay near the core where the bandwidth is high and latency is low, and sooner or later, competition for resources hatches a new level of metacompetition that obsoletes them."
"That sounds plausible," Sirhan says slowly. He puts his glass down and chews distractedly on one knuckle. "I thought it was a low-probability outcome, but ..."
"I've been saying all along, your grandfather's ideas would backfire in the end," Pamela says pointedly.
"But —" Amber shakes her head. "There's more to it than that, isn't there?"
"Probably," Sirhan says, then shuts up.
"So are you going to tell us?" asks Pierre, looking annoyed. "What's the big idea, here?"
"An archive store," Sirhan says, deciding that this is the right time for his pitch. "At the lowest level, you can store back-ups of yourself here. So far so good, eh? But there's a bit more to it than that. I'm planning to offer a bunch of embedded universes — big, running faster than real-time — sized and scoped to let human-equivalent intelligences do what-if modeling on themselves. Like forking off ghosts of yourself, but much more so — give them whole years to diverge, learn new skills, and evaluate them against market requirements, before deciding which version of you is most suited to run in the real world. I mentioned the retraining paradox. Think of this as a solution for level one, human-equivalent, intelligences. But that's just the short-term business model. Long-term, I want to acquire a total lock on the history futures market by having a complete archive of human experiences, from the dawn of the fifth singularity on up. No more unknown extinct species. That should give us something to trade with the next-generation intelligences — the ones who aren't our mind children and barely remember us. At the very least, it gives us a chance to live again, a long way out in deep time. Alternatively, it can be turned into a lifeboat. If we can't compete with our creations, at least we've got somewhere to flee, those of us who want to. I've got agents working on a comet, out in the Oort cloud — we could move the archive to it, turn it into a generation ship with room for billions of evacuees running much slower than real-time in archive space until we find a new world to settle."