Welcome! These forums will be deactivated by the end of this year. The conversation continues in a new morph over on Discord! Please join us there for a more active conversation and the occasional opportunity to ask developers questions directly! Go to the PS+ Discord Server.

Improving skills with psychosurgery and time contraction

3 posts / 0 new
Last post
DivineWrath DivineWrath's picture
Improving skills with psychosurgery and time contraction
So I'm looking to optimize improving skills during downtime. I'm trying to mix both the psychosurgery procedure, Deep Learning, and simulspace contraction. When training a skill that is 60+, you may only gain 1 skill point per week. So, to get around that, I want to do time contraction. I'll be using a house rule, where the number of egos you can run also determines how fast you can run infomorphs at. With that in mind: *Rent a private server, 50 infomorph slots, self managed, 1 gp 24 hours * 7 = 168 hours per skill point 168 hours per skill point / 50 = 3.36 hours per skill point So if I wanted 10 more skill points: 10 skill points * 3.36 hours per skill point = 33.6 hours So its looking pretty good so far. But what about deep learning? Because the number of infomorph slots matter, we need to know how many different infomorphs that need to be running. Can an infomorph use psychosurgery on themself (so we only need 1 infomorph)? Does performing psychosurgery on oneself take additional time? Would an infomorph need to fork itself so one could do psychosurgery on the other? Would both forks need to run at the same time contraction speed? In one instance, needing to run 2 infomorphs at once will split the time contraction, resulting in 2 * 25x infomorphs, which would defeat the point of deep learning (you halve the infomorph speed, thus halving the speed one can learn at so that you can double the speed the same one can learn at).
kigmatzomat kigmatzomat's picture
9 women cant have a baby in 1 month
The problem you will run into is linearity. If you need to learn x->y->a then 50 morphs all learn x then learn y then learn z. 60+, imo, is less about acquiring new data/techniques and more about synthesizing the data to create new concepts/techniques. It is the post-PhD stage. If you assume they work on a mini-thesis, can they come up with 50 different, independent thesi to write? Time dilation will work though some forms of training will be limited by the quality of the physics in the simulation. I have had characters use the corporeal form of time dilation (sleep regulator) to get +1/6 days. I am not that well read on deep learning however some other morph needs to do the psychosurgery. My understanding was that psychosurgery was often done with the patient experiencing time dilation, but that was in part so the surgeon could ensure the memories were deeply embedded via repetition.
I'm not rules lawyer, I'm a rules engineer.
DivineWrath DivineWrath's picture
I wasn't even considering
I wasn't even considering parallel forking. If I start with 50 forks, and put them into the same training scenario, I would expect to get 50 identical forks. Thats not much use for me. Looking back, it seems that I may have worded my house rule wrong. The *house rule* I was using was one where you were allowed to trade 1 ego slot for a 1x increase to time contraction. So if I had a server that could run 50 egos at once, I could run 50 at 1x, 10 at 5x, 2 at 25x, or just 1 at 50x. Since I wrote my last post, I've discovered that the target of psychosurgery can't do anything else but the psychosurgery. I suppose they could try to resist if they didn't want it, but thats not covered by the rules. So that rules out doing psychosurgery on oneself. They would have to fork to do that. I'm sure that time dilation is optional. The rules say you suffer a -20 mod if you use it, so I think the default is to not use it.