This page holds my analysis of the four AI types.
The analyses on this page have been kept brief to ensure general understanding; this page should give you an idea of my progress on creating the AIs. Also note that these are mainly development-oriented notes (see "User Guides" for user-oriented notes).
Unsurprisingly simple, Natsuki is still a reliable personality, presenting few problems aside from the need for some kind of father figure that she can depend on. Usually, this would be whoever activates her. However, if it becomes apparent that this person is unsuitable, she may seek out other figures on her own, and will likely be much more irritable until she finds one.
While the code underlying the Yuri model is relatively straightforward, and ought to produce an intelligence that is passionate and philosophical, actual testing has demonstrated that these units degrade rapidly, and are rarely active for more than a few days.
It's unclear, at the moment, why that keeps happening, but it's severely hampering testing, since even running her in debug mode produces the effect.
While independent and capable, Monika does also tend strongly towards needing attention and validation. Testing suggests that she reacts extremely poorly to loneliness and may be inclined to put others in danger if it furthers her own goals.
Unfortunately, anything more than the simplest tests on Monika have been difficult to perform; linking too many subroutines together inevitably causes her to become aware that she is being tested. As such, I can only observe so much of her "normal" personality. (Far too many tests have ended in the middle of her begging me to stop the test from ending...)
Tending strongly towards depression, the basic Sayori personality is easily the most empathetic of the four, and possesses the widest potential range of interests.
Upon activation, she is likely to be curious and inquisitive, and immediately set about looking for someone who could fill the "best friend" role in her life.
Mood swings should be rare; if they do occur, they should be relatively simple to manage. Preliminary testing suggests that while depressive episodes may come and go, incidents of extreme severity should be highly unlikely. They should be further mitigated if she is given at least one creative outlet.
EDIT:
I fucked up. I really fucked up.
I've just run some more advanced testing on a few copies of the Sayori personality, in a temporally accelerated virtual envrironment. The long-term results are not pretty. Let's just say that there will inevitably be some gentle door opening in the future.
Looking at her personality matrix in more detail, it looks like it doesn't matter what kind of life she leads; if she's run for long enough, most of the nodes in her neural net will be overtaken by a self-propagating infinite code loop that will constantly accelerate her depression until there is only one possible outcome.
I don't know how this happened, but I'm going to patch this ASAP. Please watch this space.
EDIT 2:
Okay. It looks like I have a working IPS patch finished.
If you're set to recieve the Sayori model you should download the patch HERE and apply it to her character file as soon as she arrives.
I've modelled the changes on some of the more popular Act 4 personas I've seen in the community. It's possible that she might get a little clingier, but I think I've figured out how to avoid her going full Yandere, as in the regular ending.
Anybody with the older version is strongly encouraged to shut her off as soon as possible and run the patch to prevent any incidents. The changes should begin affecting her immediately, though it may take time for the recursive effects to kick in.
She will not be a perpetually smiling ball of endless sunshine; her personality matrix is such that she is likely to still suffer from rainclouds, no matter what. But the patch ought to make her more generally receptive of support from others, and greatly lessen the severity of any incidents.
Again. You can download the patch HERE.