Looks like it’s getting close to The Ex time again. We managed to infiltrate the fair via alternative means two years in a row but last year was a bust. It’s a bit too early for any tactical planning but we’re looking to try again this August. Fingers crossed!
If you’ve been following along with /sectionb you may have noted that, along with a couple of demonstrations, Dmitri described how an artificial “neural network” is used by the Section to get operational intel from the agency.
Since ChatGPT isn’t the novelty that it once was, let me leapfrog over it to suggest the existence of an artificial intelligence, specifically a neural network, that’s trained to recognize the correlations between a person’s psychic perceptions (whatever the m.o.), and verifiable real-world targets.
To put it another way, instead of being trained to answer queries based on natural language input, the neural net is trained on the input observed by an agent under controlled conditions favourable to psychic experiences. I like to think of it as something like “The Minority Report” but with significantly fewer enslaved mutants and far less glowing ooze.
The neural net then “translates” subsequent input to produce the most likely meaning of the experience based on previous training. In the /sectionb approach one size doesn’t fit all, hence the need for every agent to train their own net, but the individual approach is the same.
I may have riffed on the idea a bit but I’m certain that I didn’t come up with the concept. For the life of me I can’t remember where I’d first read about using a neural network in this way but I’ll be sure to post an update if/when I locate the source(s). Nevertheless, I thought it might be fun to demonstrate how such a technology could be not only possible but actually practical for use “in the field”.
So I slapped together AONN, a GPT-2 based, generically psi-spy-oriented generative neural network. It lives at the bottom of the /sectionb website (below the dossiers), and uses an apropos pseudo-terminal interface.
You run it by entering “/aonn” at the “>” prompt but keep in mind that the neural net and all associated code are nearly 140 megabytes in total so the initial startup (download) is a bit of a wait. The components will be cached by your browser for subsequent sessions but still …
I make no claims regarding the relevance or accuracy of the responses but despite a lot of seemingly coherent gibberish I’ve occasionally found the text to be eerily specific and lucid. If you suspect that you may have micro-PK abilities then this may be a good chance to test them. After all, the neural network is literally running exclusively within the machine in front of you (i.e. there is no remote server that generates the responses).
Considering that something generic like this can run in a shoddy terminal emulator inside the JavaScript virtual machine within a browser, it’s safe to say that a neural net that generates only two to three words would have no problem when used with something like the Section’s pod.
The portability of the pod’s rotary subwoofer (not to mention the power supply), is a bit more questionable but considering the mostly DYI nature of the technology I like to think that Dmitri overcame these obstacles. Maybe an AI suggested something.
I believe in continuing the tradition of cyberpunk in that fiction should be plausible. Since we’re now more-or-less living in the era of that ageing (and occasionally prophetic) genre, why not throw some psi in there and call it psiberpunk?
Regardless, I hope it makes for good reading while simultaneously suggesting that, to paraphrase Dmitri, there may yet be plenty that separates us from the machines.
Colour me nostalgic but wasn’t it was around this time last year that the cops were looking for William, 31?
I don’t know if the fuzz got their man (?) but this time around they’re looking for 66-year-old Chester who is 5’9″ with stocky build, bald, and walks with a limp. He was last seen wearing a white shirt, blue jeans and blue and white shoes. Seem familiar? Maybe the following image will help:
For example … ye ole’ Rob Ford coke story. TCL broke that shit 15 months before it was ever mentioned in the mainstream:
Also there was the “lost” statue at University of Toronto, except that TCL managed to somehow capture Vickie a full 7 years before an official from Buckingham Palace even noticed that she’d gone “missing”:
Sure was a long time ago. Wonder what else has been “uncovered” since then … š
… that the fuzz is out in force today. Beware, Toronto!
Also, happy Canada Day or whatever.
... with pictures
... with sounds
... with videos
... with opinions
... that's shorter
... that's longer
