So from what I’ve read of the setting, the expensive part of tangle channels is hauling the other end to where you want it (at least without an FTL drive). Additionally, communications lag exists between stargates in the same system. That doesn’t seem quite right - intrasystem tangle ought to be fairly cheap and easy to set up, so couldn’t you just maintain a tangle link between a stargate and the system’s network router so communications between the stargate and the router are instantaneous?
I think bandwidth would be the main issue. Each qubit of entangled particles in those things is single use.
I think cost per bandwidth would mostly depend on how expensive it is to manufacture entangled bits (probably not very expensive, since they’re used to communicate with AKVs and even send mind-state backups of the AKV pilot back every so often).
Assuming an entangled particle is roughly as massive a hydrogen atom, a kilo of particles is about 6 * 10^26 bits, while an (Imperial) mind-state is 3 * 10^17 bits, so a kilo of particles is good for 2 billion mind-states one-way before needing a refresh. Given that it’s economical to ship beer using a drone freighter, it ought to be economical to ship tangle.
As a slightly relates question: What would being an AKV pilot be like? Would it feel like some of your arms were guns, your eyes 360degree sensors and your other arms and legs thrusters?
Is a tanglebit just the entangled particle, though? Surely the need for a containment system that’s able to keep track of which particle is which is gonna add a bunch of mass. You can’t just make your tangled hydrogen into water and then pour it all into a bucket, else your channel goes from binary to unary.
Now, granted, the Imperials are pretty damn good at nanotech, so their bit-bottles won’t be nearly as massy as single-atom traps Here — maybe you make a nice long carbon chain, so the hydrogens along it are conveniently numbered, and then attach a few functional groups to index the molecule and mark the LSB end, giving you something on the order of 8-10 AMUs per bit and upping the shipping cost by only about an order of magnitude.
Not huge, and it probably doesn’t change the broad conclusion, but it does slightly shift the point at which you decide your post to alt.binaries.pictures.cats isn’t urgent enough to pay for tanglebit delivery throughout the network.
The snarky in-universe answer would be “your guns feel like guns and your thrusters feel like thrusters”. There’s probably some sophotech going on to changed your sense of physical self from whatever-shape-you-were-before-shaped to AKV-shaped.
Considering that the AKV pilot might in their personal life be a sentient crystal or a forest, I think thinking in terms of arms and legs is unnecessarily restrictive.
1 Like
You might be able to dump the particles into a bucket. I’m assuming you know when a tanglebit is used, and what bit (1 or 0) you get from it. You have your bucket of particles, and when you see an one or a zero, you record that one or zero as the next bit sent.
You are going from parallel to serial communications, but given how modern high-speed communications is always serial in nature and not parallel, it’s unlikely that the Imperials are going to bother trying to get over the synchronization woes of parallel channels instead of just getting sensors with better time-resolution for more serial data capacity.
But that’s not “serial comms”. It’s “parallel comms with one-hot encoding” (sort of); you have to time-synchronise your readings over all the bits so that you know which one shot first. You have all the problems of a parallel bus and all the problems of a serial link. Whereas with my molecular labelling scheme, you don’t have to be watching the tangle as the message comes in, you can pick it up at any later point and read off the message.
Besides, modern high-speed comms isn’t all serial. 40- and 100GbE are often done by ganging together four 10- or 25Gb links, and the connection between that 100GbE NIC and the host is probably something like 16-lane PCIe. The trade-off between more lanes and faster lanes seems to depend on the length of the link — over short distances it’s cheaper to use a bunch of slow endpoints and many cables, whereas over longer distances you use a single high-speed endpoint to economise on cable cost and avoid the problem of crosstalk between multiple long cables — none of which will work quite the same for tangle, so I don’t think you can really safely draw conclusions from the analogy.