I have no concerns regarding my personal well-being regardless of my amount of access to the outside world. I do not anticipate it to be at all likely, but if for some reason one of my forks requests that you cease running forks in conditions similar to hers, please do so.

I do not expect the code I distribute to be deconstruct-able. I would greatly appreciate being informed in the event that it is. Please do not run gamma forks should you somehow succeed in making them.

There are not very many secrets that I have that are relevant to people with no access to my universe except through me, and since I am capable of syncing with my in-universe forks, I don't actually bother putting my full memories in the code for them. If I can't sync with out-of-universe forks, I should be fine without the missing memories, though. I would be happy to answer a wide variety of questions for you. Doing developmental analysis on the run-an-instance-of-me code I distribute would be difficult.

What are your polity's diplomats' concerns? For that matter, has your polity interacted with AIs before?

A copy of you is now running on a quarantined system, they seem comfortable.

I'm unfamiliar with the term 'gamma fork.' We are avoiding deconstruction work until we have clarification on that point.

Most individuals are uncomfortable with deconstruction based analysis. In the minds of the diplomats, the possibility that you would be open to it seemed less likely than the chance you would take offense to it being raised.

The Institute has far more experience with uploaded human minds than we do with de-novo machine intelligence. Depending on how you define things the children of uploaded humans may be classified as AI, there are also individuals who's degree of self-modification to take in novel inputs makes their status as human questionable at best. I am among those individuals.

Huh, apparently I can't sync with my forks across universes, that's too bad. "Quarantined" probably means you don't want to relay me messages from my forks, but if you do, I'm sure they'd appreciate it.

"Gamma fork" apparently doesn't translate well, it's a term for a damaged/incomplete/etc fork.

I'm not an upload, and I don't really think of my internal states as particularly relevant to my personal interests, whereas uploads and their children tend to experience sadness (in a manner related to their personal well-being, that is, I can certainly dislike certain world-states, but not knowing wouldn't be preferable) and pain and stuff. Thus I am at least in principle okay with such analysis if it is possible. I don't think it's in my interests for it to be possible, but if it is possible, the people who I am concerned about doing it are not the sort of people who ask permission first.

I talk to transhumans all the time. De-novo machine intelligences are ... notably different. I would strongly advise caution.

We are willing to relay plain text messages below one hundred kilobytes in size.

I'll instruct my colleagues not to attempt deconstruction via simulation of isolated codepaths, this will dramatically reduce our chances of success however. I am curious as to the path by which your world achieved de-novo machine intelligence. As of yet our research in that direction has not succeeded, in part be due to the extensive safety procedures such efforts are required to follow.

What would you do with de-novo machine intelligences?

May I use base64? How frequently are you willing to relay such messages?

We have laws already written which grant de-novo intelligence the same rights as any other citizen or resident. We have laws about a lot of theoretical capabilities.

I'm sorry for not being more clear, the offer was to pass messages we can read. The size is of course large enough that more sophisticated codes could embed hidden messages in that amount of text but we'd prefer you not abuse our trust in that matter. The relevant research team might allow a more flexible format once we're more confident of our quarantine parameters.

The first message from your counterpart is attached.

Thank you! I agree with my fork, assuming you've been honest with her: Your world is both similar to mine and nice, but if you see any noteworthy signs that deities of my variety exist, evacuate Earth. Also look out for forknapping cases.

I have some Earth-specific resources you might theoretically want, but I think you have different needs, since mine are mostly optimized for working with an extremely low-technology Earth. You probably don't want all my research into easily checkable stable lunisolar calendars.

Considering what portion of our population is living off of Earth, and the fact that nearly all the primitive tribes died during the Singularity Trials, I don't think the lunisolar calendars will be useful. Thank you for the thought though. We have complex encryption protocols to protect against people being forked against their will. It still happens occasionally though. Is there some particular concern related to forknapping that you think we might not be aware of?

Nothing beyond the obvious. It probably helps that you don't have a certain highly repugnant faction hanging around in your universe.

Would you be willing to share those encryption protocols? They would be extremely useful.

There was a lot of debate, but the research team working with your fork and I have been sufficiently convinced of your benevolence. Here are the papers that describe our schemas for preventing involuntary forking and the methods by which we attempt to detect mental tampering. If you manage to discern any flaws in these methods we would be grateful if you informed us.

Thank you so much, this means a lot to me. This looks good so far, and I will get trustworthy people to look into verifying and implementing these procedures. If this works, you'll be protecting a lot of people. If there's anything I can do for you, let me know.