Quincey Forder wrote:

okay! thanks! there goes my idea for a story, though a muse that was left behind for years, and turned bitter. some scavenger finds the morph she's inside, and sells it, unaware of the stowaway AI Not quite sure where I was going with that one, but since they go with their owner, that idea is scrapped

Actually, there's a way that can pan out. Try this out for size. The subject was a mild bioconservative who didn't particularly accept the usage of cortical stacks. He took a bullet to the brain, which destroyed the ego but left the body (barely) alive. The body is found and presumed "empty", so the finder heals it, cleans it up, attaches a cortical stack and sells it, completely unaware that the original owners muse lays dormant in the mesh inserts.

In general, I expect that the standard egocasting and egotransfer protocols work a bit like this:

• There is a core ego, the neural network/software that is run to produce the person.
• There is a set of ego metadata that describes who the ego is, citizenship and ID, various neurological information.
• There is a user-owned and -defined "exoself" that contains core virtual possessions: various certificates, personal files, and of course the muse. This may also include information stored in various implants and external objects.

When an ego transfer occurs all three are copied from the origin morph into an egobridge. The egobridge is special in that it is one of the few devices that ever interact with people's unencrypted ego data. It can download the whole ego, metadata and exoself into a new morph, or send it to an egocaster. In the case of egocasting it encrypts most of this information (various parts may be accessible using different keys) and sends it to an egocaster server that transmits the bundle to its destination. At the destination it is sent to an arrival server or a designated resleeving facility. The keys for unpacking the ego bundle has been sent separately over an encrypted channel (it is just expensive to quantum encrypt a whole ego if a short shared one-time-pad or quantum encrypted channel is enough for the key). Various customs procedures might occur (for example, checking identity or whether the exoself carries any contraband). If resleeving is successful an OK message is sent back by the ego and/or the muse with a personal key, and the copy at the origin gets deleted. Otherwise there might be a timeout at the origin, where the ego is awakened and asked what to do (better that than being beamed to /dev/null). Note however that there might be hiccups in a lot of places here. The muse might for example not be deactivated when the ego is moved out of the morph. Sloppy resetting of implants (especially exotic ones) may leave muses and other information behind. A muse might get activated on the destination side to deal with a problem in re-sleeving, and then get into its own adventures. And of course, a hacked egobridge might allow somebody to get bootlegs of both the ego and their muse. (There can be a "testament script" produced before egocasting announcing preferred morph, legal contracts, what accounts to debit, what to do in problem situations etc. This could in some situations be a small smart contract that could even hold some decryption keys, making it hard or impossible to get the ego running unless certain legal conditions are upheld. E.g. "I have been instructed by my client not to release the ego decryption key unless the morph is female Sylph, as certified by a financial or notary institution on this attached list." The further complications left by this are left as an exercise for the devious GM.)

Arenamontanus wrote:

In general, I expect that the standard egocasting and egotransfer protocols work a bit like this:

• There is a core ego, the neural network/software that is run to produce the person.
• There is a set of ego metadata that describes who the ego is, citizenship and ID, various neurological information.
• There is a user-owned and -defined "exoself" that contains core virtual possessions: various certificates, personal files, and of course the muse. This may also include information stored in various implants and external objects.

When an ego transfer occurs all three are copied from the origin morph into an egobridge. The egobridge is special in that it is one of the few devices that ever interact with people's unencrypted ego data. It can download the whole ego, metadata and exoself into a new morph, or send it to an egocaster. In the case of egocasting it encrypts most of this information (various parts may be accessible using different keys) and sends it to an egocaster server that transmits the bundle to its destination. At the destination it is sent to an arrival server or a designated resleeving facility. The keys for unpacking the ego bundle has been sent separately over an encrypted channel (it is just expensive to quantum encrypt a whole ego if a short shared one-time-pad or quantum encrypted channel is enough for the key). Various customs procedures might occur (for example, checking identity or whether the exoself carries any contraband). If resleeving is successful an OK message is sent back by the ego and/or the muse with a personal key, and the copy at the origin gets deleted. Otherwise there might be a timeout at the origin, where the ego is awakened and asked what to do (better that than being beamed to /dev/null).

Actually, I think that other than during transmission or for backup security purposes, ego encryption isn't all that common. Biomorphs obviously cannot use an encrypted ego, and a synthmorph would need the brain unencrypted in order to run it (or at least its brain emulation software would need to be capable of reading the encryption, which means it can be reverse-engineered to decrypt it, and thus again render encryption worthless).

Arenamontanus wrote:

Note however that there might be hiccups in a lot of places here. The muse might for example not be deactivated when the ego is moved out of the morph. Sloppy resetting of implants (especially exotic ones) may leave muses and other information behind. A muse might get activated on the destination side to deal with a problem in re-sleeving, and then get into its own adventures. And of course, a hacked egobridge might allow somebody to get bootlegs of both the ego and their muse. (There can be a "testament script" produced before egocasting announcing preferred morph, legal contracts, what accounts to debit, what to do in problem situations etc. This could in some situations be a small smart contract that could even hold some decryption keys, making it hard or impossible to get the ego running unless certain legal conditions are upheld. E.g. "I have been instructed by my client not to release the ego decryption key unless the morph is female Sylph, as certified by a financial or notary institution on this attached list." The further complications left by this are left as an exercise for the devious GM.)

Actually, I think that there is likely confirmation data passed between places in order to ensure that things are in place prior to transmission. For instance, a biomorph's DNA hash, or a synthmorphs identification number, is likely transmitted from the receiving bridge to the transmitting bridge as a means of informing the transmitter which body they'll be in when they get there. Perhaps there is even a "scouting AI" that is sent prior to the ego to make sure that the other end is set and the receiving location is keeping their end of the bargain. I'd imagine that other in the most shady of situations (like darkcasts), there are far too many precautions taken for too many problems to occur. Of course, if you pay Nine Lives to send you across the system, you're probably asking for it. :D

Arenamontanus wrote:

Yes, a normal biobrain or cyberbrain likely runs thing in an unencrypted form (it is a program running on a computer or a neural network topology in a brain). However, it seems entirely feasible to use something like the encryption used in some current embedded systems, where the data in memory is kept encrypted and decrypted only inside the processor. One can imagine something similar implemented in altered biology. So it seems possible (for a hefty increase in price and incompatibility with standard implants) to make an encrypted brain.

The problem with that is that unencrypting the ego becomes a simple issue of reverse-engineering the processor (which has to contain the means to decrypt by necessity). It will slow down intruders at best, but cost you a lot in implant compatibility and potential problems (wouldn't it suck if someone who hacked your cyberbrain changed the decryption key and essentially turned you into a worthless pile of scrap?). There is one field that I can see this having feasibility, however. Indentured servants likely have their cyberbrains encrypted, and installed with tamper-proof rigging. This way, the clanker has no way of egocasting to freedom and is ensured to stick around for the remainder of his term.

Arenamontanus wrote:

I completely agree. Sensible people plan their trips days in advance and set up a nice protective channel of checks, doublechecks, contracts, AI, trusted third parties and encryption. But player characters tend to have to make trips on far shorter timescales...

Player characters also have access to the resources of Firewall, though. Also, I think it fair to note that in the standard campaign, Firewall usually picks sentinels for jobs in their region. Why pick a sentinel on Eris for a job on Mercury, when it'll save you time to pick a sentinel on Mercury instead? The only exception would be sentinels with unique skillsets, who might need to be called to a region because there is no one else capable of doing what they do (probably what happens with Firewall asyncs).

Quincey Forder wrote:

that drafting by local convenience is fairly well used in the scenario "Mind the WMD", where the PC are "activated" because they were on a Scum barge at the time the mission needed to be done, and the farcast to where they needed because they were already in the know To recenter the debate on the Muses, how sentient can a Muse be? Could an infugee with according skills or an AGI "work" as one? that would be an interesting twist, to have a player doing the sentinel and another playing the muse (kinda an hi-tech version of Bloodlust, where one played the warrior and the other played the god-sword)

Muses can be very sentient, and some people actually do run infomorphs as a muse on their inserts (yes, you can run an infomorph on your mesh inserts, and you don't need a ghostrider module to do it; no you can't run a muse AND an infomorph on your mesh inserts at the same time). In fact, without going the route of PC muse, my character's muse was specifically designed to tackle the issue of "muses as beings". When you think about it in context, muses are very complex AIs, even if they are not complex enough to be AGIs in themselves. They are capable of empathy, a limited degree of learning, and catering themselves to their owner's wishes. At what point do we consider someone to be sapient? If sapience is defined by a person's capability to learn, then does that mean that people with certain long-term memory disabilities are not sapient? If they are, couldn't muses be? What about emotions and love? Sure, AIs are only imitating such emotions, but aren't AGIs as well? More to the point, if emotions in biological beings is just a series of chemical reactions in their mind, how is that any more or less real than the emotion algorithms that muses have?

Quincey Forder wrote:

so, in some ways, GITS' tachikomas could be considered a kind of Muses given physical form while doing what Muse do (I don't remember exactly in which arc they were helping the major hacking into a military system. I think it was the 11 Inviduals) muses can't be AGI (though nearly so) but can AGI become a Muse? that would be a cool way to adapt a classic of cyberpunk sci-fi: Count Zero. but that brings interresting question; the AI voodoo pantheons be TITANs themselves? TITANs spawns? and would the possession of such kind of Muse be considered a form of Insurgent viral contamination?

No; a muse is a specific type of AI designed to be a companion to people. You can get an AGI in your mesh inserts and have him/her handle your everyday needs, but he/she will still be significantly more intelligent and capable than any muse will.

Quincey Forder wrote:

Do you think that the muse appear in AR in your view range? or is he/she purely a voice over?

The way I see it, people could customize things so that their muse has as much or as little access to their systems as they want. Some people certainly could have their muse as a constant AR presence sitting in the corner of their vision, and it seems like the sort of thing that would make them more interesting to interact with for players, so I'm all for it. I suppose some enthusiasts could go far enough as to get ectos with speakers in them or something so their muse can interact with the outside world if necessary, though I imagine that their limited AI would make that problematic.