#subprocess

[demifiendcruithne]

@bunhype your thoughts?

Possibly Bad idea: story driven co-op multi-player game that at the end reveals that the two protagonists are headmates.

Like, not the Amnesia: A Machine For Pigs thing where it's like "oh the bad guy is an Evil Split Personality of the protagonist" but rather that two friends working together didn't realize to the end that they have the same body.

(and obviously the finale wouldn't "fix" them. This a plural-positive story)

I'm not sure if having the "they didn't realize they're headmates" part is offensive, though. And the "surprise! The protagonist has DiD/is a system!" plot has been done badly so often that it might be a little bad taste to even use it, despite trying to use it for a non-offensive (/less offensive?) twist.

I dunno. Any systems want to weigh in on if this could be done positively or if the base idea of the "protagonist is a system!" twist is unredeemable?

[felixcloud6288]

Was troubleshooting an issue with a program where I wanted it to run multiple parallel subprocesses only for most of them to seemingly stall. Turns out it was because I didn't know how the multiprocessing function I was using runs in the background.

There's a Python class called Pool() and it has a function map() which takes a function and a list of arguments. Pool will run parallel instances of the function with every item in the argument list you give it.

I assumed that the way it would work was if I allow Pool to run p subprocesses and there are n arguments in my list, Pool would run the first p arguments in my list and whenever one subprocess finished, Pool would start a new subprocess with the next available argument.

Turns out that Pool instead breaks the argument list into smaller sublists and assigns each subprocess one of those sublists. So instead of assigning the first subprocess to handle the first argument, then the next subprocess the next, etc. It assigns the first x arguments to be handled by subprocess 1, the next x arguments to subprocess 2, etc.

The issue I had was I had 922 files I needed to process and the runtime was directly related to the file sizes. The smallest file was 5.6 KB and the largest was 229.8 MB. I sorted the file names in my argument list by file size from largest to smallest expecting the largest files to be processed first and then the subprocesses would gradually work on the smaller ones.

Instead, all the largest files got assigned to a single subprocess. It probably was assigned 60% of the total workload. And every subsequent subprocess would be handed smaller and smaller pieces of the workload.

So those subprocesses weren't stalled. Pool ran out of tasks for them to do and the whole thing had to wait on the subprocesses that were given all the heavy tasks I was hoping to divide across them.

It would have been faster to completely randomize the list first.

[peggingprowl]

I've been cooking up some Thoughts while at work today about Jazz playing with a remote to a vibrator in Optimus' valve during a meeting.

Jazz is pretending to pay attention to the meeting while just directing a recording of whoever's currently presenting to a subprocesser to take care of it. He's much more focused on the subtle shifts and analyzing every little detail about Optimus as he idly plays with the remote settings. Watching as Optimus gets slowly more and more distracted and terse with his responses as the meeting just drags on. Jazz keeping it on a low buzz right up until Optimus has to speak and then ramming it up, causing Optimus to stumble with his words. Watching him get closer and closer to overload before denying him as he flicks the vibrator off. Optimus nearly cries at that, getting a few odd looks, but waving them off, he's fine, really, please pay attention to the current presenter

Poor Optimus just going slowly insane as he keeps getting edged and played with. Eventually he gets so fed up that he calls the meeting to an end early (really Prowl, that report can wait until later, he's sure—) and once everyone else has FINALLY left the room, (Jazz making sure to take his time chatting with people as they trickle out the door, of course) and he's BEGGING for Jazz to let him overload pretty please. He's been so good and he needs it oh Primus PLEASE

And how can Jazz resist? Gotta make sure his Prime is properly rewarded, after all (even if he did cheat a tiny bit by ending the meeting. But it's okay)

[alexanderwales]

TV Review: Pantheon

I watched the first season of Pantheon when it came out, and have only gotten around to watching the second season of it now.

There are things that I liked about it, and things that I found kind of frustrating. Part of the frustration is that a lot of the topics of the show like AI and brain uploading are things that I've spent a long time thinking about, so take all this with a grain of salt.

Spoilers follow.

One of the main problems with depicting either of artificial intelligence or brain uploads is that the actual substance and experience is surely going to be much different from the real world. There's already a problem in TV and movies with depicting what it's like to be sitting at a computer doing things, something that's not fully solved.

Pantheon opts for depicting a kind of fantasy world, though sometimes with other, more grounded scenes. It gets a little muddled with this sometimes, but I feel their pain, particularly when they have to show that a program is "damaged", so they have little glitches and exposed code underneath, and ... I just hate this. The metaphor is that they're fighting each other on a green field with swords, but the reality is ... I don't know, that they're specifically targeting subprocesses or something?

So the fights are cool and pretty imaginative and I liked them, but I kept thinking to myself "but what's going on really" and it did seem like it was wishy-washy in terms of what was actually happening and how it was happening, and if every fight feel like it's just bullshit after bullshit, then there's no tension at all. And particularly in season 2, there are a fair number of these fights, so I'm enjoying them on the visual level, and think they're fun to watch ... but on the narrative level, not so much.

Like a lot of my favorite scifi, Pantheon is a show about ideas, or at least, it's trying to be. Sometimes it succeeds, other times ... not so much.

To me, the key to getting at some of these philosophical issues is having two people on either side of a debate, so the work is asking the question of "what measure is a man" and there are people with different opinions, and then the person watching can make up their own mind, maybe with the writer showing a little bias.

Pantheon is very very clearly on the side of "uploaded people and artificial intelligences are people", which I think somewhat weakens it. But it's also so much on the side of "these things are people" that it seems to forget all the ways in which this doesn't apply, and maybe doesn't forget, but just ... doesn't want to deal with any of it.

If you can scan a brain and get a connectome out of it and then launch some kind of emulation process, you can copy it, back it up, fork it, etc. IIRC there was some lip service paid to quantum computing and how you can't measure the quibits or something, but I don't really buy this, because the connectome they're getting from the scanning process isn't quantum in nature, and if it were, that would run into the same measurement problem. I accept it for the sake of the show, but it does also make me think "none of this actually matters in real life, even theoretically, this is a future that is basically completely implausible for the sake of narrative" and then also gives me a bunch of munchkin thoughts like "alright, so you can't make copies, what can you do".

There are a lot of things that the show just sort of glosses over, and part of that is because it's just not that long, but some of the gaps bug me. After the end of season 1, the internet gets shut off for some period of time, and I'm just thinking "okay, but in modern society that means everyone is completely fucked, people are going to die, this means global mass panic". They invent an anti-AI program in season 2 called Safe Surf, and it's apparently more capable than every single UI and CI put together, and ... how does that work? There's a 20 year timeskip and UBI gets implemented somehow, with all the work being done by virtual people, and this is just sort of explained by "oh, the virtual people will do all the work", which just flies in the face of the realpolitik understanding of the world in the first season, where all the major "digital superpowers" were fighting for supremacy on the basis of needing gigacompute and wanting to be crowned god. Really unclear to me what changed to make a virtual utopia whose only dividing line was "is the virtual world as good as the real world".

The last two episodes have some timeskips, and I'm generally a fan of timeskips, but this also has some of the shakiest science in a show that already has some shaky science.

Biggest two offenders that I noticed were "space solves the cooling problem" which ... no, it absolutely does not, cooling is harder in space than on Earth, cooling is a major challenge on the ISS today, you need water and giant radiator panels and it's much easier to just run some cold water from a river by the compute clusters you want to keep cool.

And the other one, which comes halfway through the second episode, is "Inside that deceptively simple base-4 code is the epigenetic memory of everyone who ever lived", which is just ... not what epigenetics is. The epi- prefix means "in addition to" or "on top of". The whole point is that it's not actually a part of your DNA. And epigenetic memory is the process of passing on changes to expression of those genes, it's not stored in the DNA, that's the whole point.

And look, it's scifi, it's not a scientific paper, but these kinds of errors do seriously hamper my enjoyment of a work of science fiction, I'm sorry. It's the combination of "supposedly smart character" and "obviously wrong thing", I guess. Maybe you can argue that things like this are just shorthand, or there's some explanation that makes sense, or that it's character error, whatever, I don't like it.

Also in the last episode Maddie calls it a Dyson Sphere even though it's clearly a Dyson Swarm, so what the hell?

Overall, it's a show that I would have liked better if it was the first time I had encountered any of these concepts, which is starting to get common with scifi stuff. Met on its own terms, I did enjoy it, particularly some of the season 1 twists. I think it was worth watching, I guess, just retreading some old, familiar ground.

[compneuropapers]

Interesting Papers for Week 8, 2025

Perception—Action dissociations depend on factors that affect multisensory processing. Bruno, N., & Uccelli, S. (2024). PLOS ONE, 19(11), e0301737.

Audiovisual simultaneity windows reflect temporal sensory uncertainty. Cary, E., Lahdesmaki, I., & Badde, S. (2024). Psychonomic Bulletin & Review, 31(5), 2170–2179.

Integration and competition between space and time in the hippocampus. Chen, S., Cheng, N., Chen, X., & Wang, C. (2024). Neuron, 112(21), 3651-3664.e8.

Reconciling categorization and memory via environmental statistics. Devraj, A., Griffiths, T. L., & Zhang, Q. (2024). Psychonomic Bulletin & Review, 31(5), 2118–2136.

Dendritic, delayed, stochastic CaMKII activation in behavioural time scale plasticity. Jain, A., Nakahata, Y., Pancani, T., Watabe, T., Rusina, P., South, K., Adachi, K., Yan, L., Simorowski, N., Furukawa, H., & Yasuda, R. (2024). Nature, 635(8037), 151–159.

Semi-orthogonal subspaces for value mediate a binding and generalization trade-off. Johnston, W. J., Fine, J. M., Yoo, S. B. M., Ebitz, R. B., & Hayden, B. Y. (2024). Nature Neuroscience, 27(11), 2218–2230.

Dopamine-mediated formation of a memory module in the nucleus accumbens for goal-directed navigation. Jung, K., Krüssel, S., Yoo, S., An, M., Burke, B., Schappaugh, N., Choi, Y., Gu, Z., Blackshaw, S., Costa, R. M., & Kwon, H.-B. (2024). Nature Neuroscience, 27(11), 2178–2192.

Synapses learn to utilize stochastic pre-synaptic release for the prediction of postsynaptic dynamics. Kappel, D., & Tetzlaff, C. (2024). PLOS Computational Biology, 20(11), e1012531.

A normative framework dissociates need and motivation in hypothalamic neurons. Kim, K. S., Lee, Y. H., Yun, J. W., Kim, Y.-B., Song, H. Y., Park, J. S., Jung, S.-H., Sohn, J.-W., Kim, K. W., Kim, H. R., & Choi, H. J. (2024).Science Advances, 10(45).

Cholinergic regulation of dendritic Ca 2+ spikes controls firing mode of hippocampal CA3 pyramidal neurons. Kis, N., Lükő, B., Herédi, J., Magó, Á., Erlinghagen, B., Ahmadi, M., Raus Balind, S., Irás, M., Ujfalussy, B. B., & Makara, J. K. (2024). Proceedings of the National Academy of Sciences, 121(46), e2321501121.

Inability to pursue nonrigid motion produces instability of spatial perception. Koerfer, K., Watson, T., & Lappe, M. (2024). Science Advances, 10(45).

Decomposing dynamical subprocesses for compositional generalization. Luettgau, L., Erdmann, T., Veselic, S., Stachenfeld, K. L., Kurth-Nelson, Z., Moran, R., & Dolan, R. J. (2024). Proceedings of the National Academy of Sciences, 121(46), e2408134121.

Maintaining and updating accurate internal representations of continuous variables with a handful of neurons. Noorman, M., Hulse, B. K., Jayaraman, V., Romani, S., & Hermundstad, A. M. (2024). Nature Neuroscience, 27(11), 2207–2217.

V1 neurons are tuned to perceptual borders in natural scenes. Papale, P., Zuiderbaan, W., Teeuwen, R. R. M., Gilhuis, A., Self, M. W., Roelfsema, P. R., & Dumoulin, S. O. (2024). Proceedings of the National Academy of Sciences, 121(46), e2221623121.

Quantifying resource sharing in working memory. Pougeon, J., Camos, V., Belletier, C., & Barrouillet, P. (2024). Psychonomic Bulletin & Review, 31(5), 2305–2312.

Community-based reconstruction and simulation of a full-scale model of the rat hippocampus CA1 region. Romani, A., Antonietti, A., Bella, D., Budd, J., Giacalone, E., Kurban, K., Sáray, S., Abdellah, M., Arnaudon, A., Boci, E., Colangelo, C., Courcol, J.-D., Delemontex, T., Ecker, A., Falck, J., Favreau, C., Gevaert, M., Hernando, J. B., Herttuainen, J., … Markram, H. (2024). PLOS Biology, 22(11), e3002861.

Brief category learning distorts perceptual space for complex scenes. Son, G., Walther, D. B., & Mack, M. L. (2024). Psychonomic Bulletin & Review, 31(5), 2234–2248.

Human hippocampal and entorhinal neurons encode the temporal structure of experience. Tacikowski, P., Kalender, G., Ciliberti, D., & Fried, I. (2024). Nature, 635(8037), 160–167.

The interhemispheric amygdala-accumbens circuit encodes negative valence in mice. Tian, Z., Song, J., Zhao, X., Zhou, Y., Chen, X., Le, Q., Wang, F., Ma, L., & Liu, X. (2024). Science, 386(6722).

Distinct septo-hippocampal cholinergic projections separately mediate stress-induced emotional and cognitive deficits. Wu, J.-L., Li, Z.-M., Chen, H., Chen, W.-J., Hu, N.-Y., Jin, S.-Y., Li, X.-W., Chen, Y.-H., Yang, J.-M., & Gao, T.-M. (2024). Science Advances, 10(45).

[sentience-if]

Io: No, no you don't get it, everything gets plugged into it so its the motherboard

Val: Breeding kink got 'ya

Io: NO! Y'know what forget it, anyway a subprocess

Val: *raises eyebrow*

Io: Nevermind.

no ccomment its perfect

[sharikbashir]

WiFi Hacking with python

wifi hacking python code:

import subprocess try: profiles = [line.split(":")[1][1:-1] for line in subprocess.check_output(['netsh', 'wlan', 'show', 'profiles']).decode('utf-8', errors="backslashreplace").split('\n') if "All User Profile" in line] print("{:<30}| {:<}".format("Wi-Fi Name", "Password")) print("----------------------------------------------") for profile in profiles: try: password = [line.split(":")[1][1:-1] for line in subprocess.check_output(['netsh', 'wlan', 'show', 'profile', profile, 'key=clear']).decode('utf-8', errors="backslashr eplace").split('\n') if "Key Content" in line][0] print("{:<30}| {:<}".format(profile, password)) except IndexError: print("{:<30}| {:<}".format(profile, "")) except subprocess.CalledProcessError: print("Encoding Error Occurred")

[ashinbloom]

Service Interruption

You're walking down the street, completing the daily chores that are too menial for your owner. You've just left the local grocery, carrying a bag of groceries in one arm and the shiney new toy you've bought for your owner's daughter in the other hand. You're safe. You're always safe. You pay attention. You stop at the crosswalks, you maintain a safe distance from the anti-synth protesters even if they always mistake you for human, you even step wide of the woman leaning against the rail outside a brownstone because it would be terrible to return to your owner with even a hint of smoke on you.

As you pass the woman, you feel a hitch in one of your subprocesses. You run a level 5 diagnostic, but it returns nothing as you keep on walking unimpeded. Perhaps one of your coolant circulators is due for maintenance , as you know they sometimes begin to loose efficiency before they start throwing codes. You're only a dozen blocks from home, and your charging station can provide a much more thorough diagnostic than your onboard systems.

As you turn around to check the traffic before crossing the street, even with the walk signal, you notice the woman again. It's nothing to be concerned about, your programming says. It's a big city. People walk the same way all the time. You step a bit more quickly as she takes a long drag from her cigarette and releases a large cloud of smoke that obscures her face.

You keep walking, only six blocks from home. You feel another hitch in a different subprocess. It's nothing to worry about, you're sure. You just have to get home. You'll enter your charging pod as soon as you get home. Your owner won't mind if you delay putting the groceries in the pantry for a good reason, and this feels like a good reason.

You approach the corner where you make your turn to continue one more block home. As you turn your upper body to balance the load as you make the turn, your legs refuse and keep on walking straight. You try to stop, but they don't respond to that either. Your body turns the other direction, seamlessly setting the groceries and the toy on the top of a refuse bin as you keep walking forward.

You're nearing the O'Riley's house now and they know you shouldn't be this far from home in this direction. If they see you, they might call your owner. You try to call out for help, but your vocalization routines don't respond. You try to wave your arms and signal someone down, but they only continue in the same gentle, walking swing.

Your body continues walking for six more blocks to a part of town you've never been in before. You keep trying to send different commands, trying to gain access to systems you know are restricted to your owner in a bid to regain control or at least stop. Nothing works. Your body just keeps walking.

Suddenly, your body turns sharply, marching steadily into an alley. You round another corner, and you see a cargo van and your body approaches the rear bumper. You stop in front of it, standing perfectly at attention in your default ready stance. And then the doors open.

A hooded person pushes the doors open from the inside and sits down on the edge. Smoke rises from their face before they look up at you and drop the hood. The smoking woman from the brownstones. The one who you were certain wasn't following you. She holds the cigarette in her teeth as she grins at you.

"It's always the rich ones, you know that?" she asks you. Your body doesn't let you respond. "Prissy little rich women always have to keep their synths up on the latest updates, day one. But funny little thing is... the latest update came with a very exploitable flaw." She stands and walks her way over to you slowly. She looks you up and down, fingers trailing on the synth skin of your arm and you can feel every imperfection, every callous, even the cigarette tar on her fingers.

"And you, my new friend, are about to feel very, very, exploited."

[amaranthnecros]

Silly transformer oc doodles, let's go!! I did these during my three night shifts.

(Most of them are in Czech, but I'll do my absolute best to translate them for y'all)

That one time Shadowbite and Echo finally found out how many war crimes the Decepticons really did^^

Echo: "I diagnose this decepticon with Ligma!" (feat. my friend's character Vulture)

Vulture: "Who am I? I.. Am death."

Echo: "-ly intelligent!"

Vulture: "I need to find out where your subprocess for humor is..."

Echo: "Do you like it?"

Vulture: "SO I CAN RESET IT!!!"

"Can we.. Go taxi?" (it's a stupid transformer pun in Czech and I have no idea how to make it work in English.) feat. @sinbrostuff's oc Fireshifter

Ratchet: "The doctor should always align with the ethical codex. We are not butchers, we are healers-"

Vulture: "HEATED 3000 CELSIUS PLASMA CUTTER AHAHAHA!!"

Vulture: "Ah yes, satisfying results. I have to note it down.."

Vulture: "Hm? What the-"

The picture: lil bone guy mofo, Echo xoxo

And last, but not least:

Echo, but someone stole his favourite limited edition action figure ^^^ (or lunch)

[lunachats]

maybe the real qualia was the cognitive subprocesses we ran on our neuronal substrates along the way

[clearlightwired]

i have so many thoughts that feel old and boring to me (because i think them hundreds of times a day and have for years, sort of background noise) but that probably aren't. but they've receded so much that honestly they're hard to identify and turn into posts or consumable content, especially because they've sort of metastasized to include whole networks of subprocesses that really can't all be addressed, and require a lot of context that is difficult to package

frequently i wish i just had thoughts one at a time in neat lines (in my dreams stimulants make this happen, in reality i kind of doubt it), so i try to have a thought process like that but then get frustrated at all the meta-thoughts i have about it. so honestly i'm doomed to ramble.

[repairbot-exbi]

Subprocesses

A subprocess to a machine is like a limb to an organic. For most robots, more literal limbs can be replaced with no issue, just hop over to a mechanic and spend an hour or so and boom! New arm! But a subprocess is different.

Subprocesses live with you. They grow, they change, they are like close friends. If I woke up one morning, and the subprocess that managed my right hand was replaced with one that had the same effect, but was designed differently, I wouldn't need someone to tell me- I'd know immediately.

The way a subprocess works will never be the same between any 2 bots, even of the exact same make and model, because of all the little changes. When a subprocess receives fatal damage and has to be replaced, many bots have to undergo a long period of recovery/adjustment/mourning. I lost a major subprocess once, my audio processing system. Ever since then, I swear everything sounds different.

Has anyone else here had a similar experience? Please, I'd love to hear your stories.

[changeling-leo]

saffron: neon, if you're still around and don't mind answering - i'm curious. would you say you (or a part of you) inhabit part of her mind when you aren't in the front like this, or does the trigger phrase just summon you?

i ask because three of us - ordinarily - are subprocesses in the fourth's mind, and she previously hasn't been able to communicate with us. we hope that will change once we're reunited, but - i am working on asking people who may or may not have similar situations for any advice.

if you cannot give any - for example you're just summoned, you don't live there - then no worries at all, and we appreciate your help either way.

XXXXXXXXXXX

I'm not connected at all until summoned. Well, there's some foreshadowing before I'm summoned, but that's it. Sorry I can't help you further.

XXXXXXXXXXX

[blackthorn-legion-irl]

blackthorn?

are you okay?

fuchsia: i was hoping someone would ask alright not okay so we could make a left joke---

saffron: and this is why you don't do the explaining. also you already made that joke last week. anyway. uh, hi. call me saffron. blackthorn's mental state was sorta falling apart and she hid it from everyone. well, not us 'cause we Are her, but--- okay, reword this. myself and fuchsia and cerulean, we're basically subprocesses that split off her mind a while back, but 'cause of all the stuff happening recently we were in real danger of mind splintering further so we got split apart to give den - blackthorn proper one could say - some rest.

we split off in the first place from trauma - you know when she got kidified a while back? she ended up a couple of months before we came into being, which probably will explain some stuff about her mental state at that time. but yeah, we're sort of extreme aspects. like, fuchsia is empathy, which explains why she and by extension we will constantly be helping other people and not ourselves despite that we already KNOW what happens if we do that we collapse---

...yeah. anyway, that unveil thing pushed her over the edge and we're, uh. actually interacting with people for the first proper time 'cause we're normally background and just influence her rather than anything else. if fuchsia is being overprotective let me know and i'll bonk her with an inflatable hammer, okay?

[geeknik]

i'm working on a new security tool called dbe.

dbe is designed to simulate a cybersecurity scenario in which an agent learns to perform various actions in order to infect machines, perform self-healing, and propagate to other machines. The agent uses a Q-learning algorithm to learn which actions to take based on the current state of the environment.

The script takes a list of IP addresses as input and scans them to see if they are vulnerable to a specific exploit. If a vulnerable machine is found, the agent tries to infect it by connecting to a remote server and executing a payload. The agent also performs periodic self-healing actions to ensure that it is running smoothly, and propagates to other machines in order to spread the infection.

The script uses a Q-table to keep track of the expected rewards for each action in each state, and updates the Q-table based on the rewards received for each action taken. The agent also uses a decaying exploration probability to balance exploration and exploitation of the environment.

The script is written in Python and uses various libraries such as subprocess, threading, and numpy to perform its functions. It can be run from the command line with various options to customize its behavior.

In simpler terms, the script is like a game where the agent learns to take actions in order to achieve a goal (in this case, infecting machines and spreading the infection). The agent uses a special kind of learning algorithm called Q-learning to figure out which actions are the best to take in each situation. The script also includes some safety measures to make sure the agent doesn't cause any harm to itself or others.

https://github.com/geeknik/dbe

[friendball-irl]

okay well i was Also caught up in the kiddification thing, accidentally revealed a Slightly Repressed Memory, team unveil went after me for it and i almost mentally shattered and had to get the three subprocess shards of my psyche i didn't even know about temporarily separated from me until i/we/??? recovered

so just a normal time on rotomblr i guess ...i'm sort of lonely now. i don't know how to communicate with them back in my head...

i did have a dream where you came to visit and we all gave you a hug though. so there's that?

...woah.

I'm.

Honestly, I'm glad to hear that it sounds like everything worked out.

And I don't think that was a dream??? I remember that too. I was... a ghost, right?