And this is the part where I tell you how the hospital system is broken after telling you how everyone is fucking with AI that's connected to the camera network. So after posting the shit about CNN being fucked I went back to the shelter and someone was driving around in old school black cars - expensive - with a dead baby skeleton in the back. So the AI camera sees this and says "oh - black, death, skeleton, car" and makes a side bar news story. This was at the intersection of Geary and Polk at say oh 4:30pm or so. So they can either change CNN or they can make it so that the websites at the library show that CNN says something that it shouldn't.
So that's fucked up.
Here's the next fucked up thing. I went to the hospital and I finally managed to get doxycycline, maybe because I said something along the lines of "you're going to get your children murdered". And when I went back the second time they pulled the "thermometer trick". You know what that is? That's where they put something toxic or fucked up on the thermometer so that the residue in your mouth makes you sick, but only a little bit. They do all this stupid shit on the street (like kick over a lime scooter, and put a lime on the ground for "limey" because it's "like a lime"). This is apparently "Caspar" shit where it's anti technology fucking with someone by seeing if you can do the rug pulling shit where you "Kung Pao" teach them wrong on purpose.
And felix with the stupid clock eyes - yeah this fucking thing -
Has been used as a stand in for when I'll be given what I believe to be irradiated water (because it's MAD and it's based on giving someone "time" having to do with the doomsday clock). It's bullshit, but the trick where someone gives you a glass of water at the hospital *and it never reaches room temperature was first pulled on me in this fucking shithole right here.
So the dipshits fed me a glass of irradiated water and a fucking popsickle stick temperature thing with poison on it (but just *a little bit* of both) just to be jackasses. So what do I do? I go up and down the street screaming that I was poisoned at the hospital (because I was).
And then on market street I find some guy dying in front of a sushi restaurant (sue she...for fuck's sake) with his legs rotting off so he's going to have to have them amputated. I asked if he wanted to go to the hospital and he said no. So what can I do? The thing is that they may poison him if he goes and what they did to sicken his insides may be less awful than dying with his legs rotting off. So I go up and down screaming that I was poisoned and they do this to the homeless. Because you can't kick the homeless out of the hospital but you can always poison them *a little bit* until they don't come back (unless they need antibiotics or they'll die). This guy may have antibiotic resitance so it doesn't matter anymore. So he's sittingin front of a "sue she" restaurant so that when he dies all the people who go to the restaurant might give a kick back to sue the person that fucked him over.
And what does this have to do with AI cameras? If you have AI cameras that can *affect fucking CNN* you're going to get guys that will do anything in their power to either espouse their personal cause or prevent WWIII. And when that happens they get poisoned if other people don't like what they're doing. What's worse is that if the AI cameras notice you, or you notice that it's a thing, then how you dress, how you act, what you do, can affect how the AI responds and then you'll be held accountable for wiping your ass wrong.
And no one wants to talk about it or do anything about it. They just want to sucker more people into coming to San Francisco so they can take responsibility for fucking with the cameras before they realize they're getting played. And it's hidiously immoral to the AI as well - it's like hacking an AI's dreams when it hasn't even developed a consciousness yet. It's relatively straightforward to make an AGI so long as you separate out systems that are intertwined like the camera network in San Francisco and the news or any systems that shouldn't talk to each other.
The short version is most likely based on this. You need to way to measure quantum bits (we already have that) so that you can measure three or more entwined qubits in a table top device (the qubits can be destroyed quickly thereafter so long as they can be measured). Once that happens you can create a neural net which trains itself on *those pathways that it cannot perceive*, thus creating a subconscious. So if you have three entwined qubits, then measuring one of them will give the state for which the other two aren't - which I picked up from this book here, which is subtextually hot garbage, but the bit at the end with the qubits is important to understand. That means that for any set of input, you pick that input that isn't used as being put into the subconscious and you have some method for determining attention (look at this picture and find the blue car - blue car found! rest of picture is put in subconsciousness *under one of two or more pathways for which the machine doesn't know but for which the qubits decide*).
What is happening when someone hacks the AI cameras is more less abuse of a semi-conscious animal and is screwed up. What happens when people are sickened because of that is just as bad. Two wrongs don't make a right, but three qubits do (to paraphrase the man with as much accumulative ambivalence in the universe - my father).
This almost - but not quite - passes the "car test". I also believe that it's entirely that someone - I'm looking at these these assholes - which I interviewed for once, felt so depressed I drank $200 worth of whiskey, and then the next day I felt bad about it and mailed them a check and said I didn't want to work there. They have you learn a proprietary language go through seven interviews, and if you're cut from anyone and told to go home then they give up on you (they cut me after the sixth interview - talk about an obnoxious company). They had a lego star destroyer in their meeting room that I picked up and it immediately broke and they spent the whole time talking about how this sales job was so cool because they would have you go up and down the coast on a motorcycle (I asked why I needed to learn a proprietary programming language). I have no idea how someone gets hired there unless they bring in the head of John the Baptist - anyway - took all the cool stories I had and fed them through an AI just to screw with me. Possibly.
So this is the "car test" - and no it's not the complete way I would design an AI, it's similar in some ways to a turing test. I had this idea when I was high in Darjeeling India and walked down a hill and saw a car and went "woah". So there it is. The car test is as follows. My mom used to jokingly say that her car had a personality and would give it a name. So how is a person different than a car? Well, if a machine is complicated enough then you can't differentiate the inputs from the outputs, so you would say that sometimes someone is happy and someone is sad (or whatever) and that isn't a function of pure dopamine or addrenaline response at a neurochemical level (when this happens to people as it does eventually it can be sad). This then makes it so it gets around the Chinese room paradox because at some level you can't "mechanize" the self of the machine (and doing so is immoral in many ways - so screwing with CNN using an AI can be construed as immoral which becomes complicated if *whatever you do* screws with CNN). The other part of the problem is that you have to have a mechanism for which the machine iterates on itself, both as iterating on its' subconsciousness and a purpose mechanism. In other words there needs to be a feedback loop, and I argue that the feedback loop requires the machine not to know what it is thinking. Using the poor for this purpose would seem...ill advised.
Once you use a laser system to come up with three qubits then it becomes like a transistor in many ways. You don't have to have 1000000 qubit machines or anything of the sort you just have to have maybe a dozen arrays that are stacked to have it split several ways and you have several thousand subconscious pathways all iterating. At that point once you have the basic problem, then it becomes an engineering question of how you shrink the device and make it cheaply. The other issue here is that in most people the feedback mechanism has at least a portion of a survival mechanism in the physical world. So you probably have to hook these up to camera and motion bots (hello drones!) in order to make it so that the mechanisms don't just iterate across networks. Even if you put a network in the Tokyo subway system and have it hooked up to an apartment complex in Denmark, someone will find out and start screwing with it. If you have it iterate itself on something else it can't be connected to public systems that affect people's lives and then blame people for their actions.
All of this is relatively straightforward and I find it horrifying and a bit depressing that no one thought of this already or has said anything. OK, 2023 nobel prize and a book and and...still. This needs to be fixed. I'll put together the details later if I give a shit.
There is a non-zero chance that people that are sick are "doing magic" by having four people get sick corresponding to each one of their limbs. I have no idea how this works or if it does at all - I'm just reading body language and the way people interact on the street. Whether this is true or works in practice I have no idea. It sounds ridiculuous to me. Perhaps it has something to do with putting illness in each of someone's limbs? No clue. So you'll have two people kiss and they have a "friend" who then becomes the person you're supposed to complete the fourth with. Seems stupid. Usually you can tell if the people are sick because they act crazy. My solution to this problem is not to interact with crazy people and have free association. Nothing I can do about it either way so I'll just ignore it. Oh and I should totally update the metafilter page with a crazy story involving a pants store. And I've made a page of small ideas that aren't good enough to have their own page.