We meet the protesters who want to ban Artificial General Intelligence before it even exists On Saturday at the Silverstone Cafe in San Francisco, a smattering of activists gathered to discuss plans to stop the further advancement of artificial intelligence. The name of their non-violent civil resistance group, STOP AI, makes its mission clear. The organization wants to ban something that, by most accounts, doesn't …
On one hand, we have the currently implausible but potentially real concern of AGI's genuine development and improper handling resulting in a threat to human existence on a fundamental level. On the other hand, we have it potentially generating profits for investors and shareholders in the brief moment of time between emergence and the end.
Add it to the pile of "sacrifices they're willing to make" along with climate change, poverty and infectious diseases.
I suggest they drop their signs and picket line and go put in a solid day's work in the office/warehouse/shop/yard or wherever.
Then focus on the changes in life that can better the social environment they live in, which by proxy effects the greater good of the many.
AI is here already. Its going nowhere and the more time invested in training people how to use it efficiently, effectively and securely the better.
The tech companies have been told to go crazy with what ever inhumane crime they can conceive in their lex luthor brains by your newly self appointed emperor, and they will. Non violence is a waste of time and energy. The death squad are in charge now.
Best you can do is defend yourself. You will need guns, big guns that blow really big holes in robots. Things that punch through armor.
1. The current generative AI and large language model companies (and their investors) are like a bunch of gold-obssessed fortune hunters running around and digging up peoples' lawns because they "think" they'll get rich doing that.
2. In the U.S., at least, you can buy really-big guns, if you can afford them and get the U.S. government's approval. If you do, you're now on a federal list, and the robots will naturally come for you first.
I think it's more practical to drop or throw thin plastic sandwich baggies filled with paint to cover/clog the robots' sensors. You probably should figure out how to launch those paint-baggies from a distance ...
Roko's Basilisk will get them.
Feh. You're no fun. Whenever I see Roko Basilisk I think of Roko from Questionable Content. https://www.questionablecontent.net/cast.php Scroll down to find her. She's an AI. She's also friends with Bubbles, a Very Large ex-military AI, and Yay Newfriend, a multi-bodied AI with a god complex. Bubbles is retired from the military, has taken off her armor, and is Very Very Good Friends with Faye. Roko likes pastries. Yay, now, Yay's dangerous.
(Yes, I'm fairly sure that Roko got her name because the author of the site knew about the other Roko Basilisk.)
Warning: do not visit Questionable Content if you have issues with coffee, pastries, libraries, allosaurs, dogs, robots made of slime, or robots who are very interested in your butt.
Nothing they do will have any impact upon governments and militaries using AI because they operate above the law. Foreign countries will just ignore any US ban. The US will not stop developing because foreign countries are. It's the same as those who want to ban atomic bombs. The genie is out of the bag. It's a one way street.
Quit and do something benign and beneficial rather than wasting your lives banging your heads against a brick wall.
> Nothing they do will have any impact upon governments and militaries using AI. . .
You underestimate the effect of the Butlerian Jihad1. "Thou shalt not make a machine in the likeness of a human mind."
Now, if you'll excuse me, I have to run. My Shai-Hulud2 is parked in a loading zone.
_________________
As far as intensity of opposition to AI is concerned, yes, it makes sense to me that if there's been a PauseAI for a while, there should also be this here StopAI, and one could suggest the creation of a PowerOffAI as well, seeing how the tech is such an energy glutton for what little actual value it yields aside from hyperbole.
It's the same thing as the current crop of climate alarmists who are completely detached from reality. I wouldn't even be surprised if it's the exact same people pushing that agenda.
It's not mentally healthy to believe these things with absolutely no evidence and if anything the evidence reaching a completely oppositional conclusion.
They'll start claiming "the science says" then "all science agrees" (by the way, phrases like these are completely oppositional to first principles of science which is how you can spot people who aren't really scientists a mile away, huge red flag - if science agrees, it wouldn't be science: it would be politics - I'm very reassured by the fact that not all scientists believe that the sky is blue and water is wet, and nor should they) next, ignoring the fact they're totally false as they are with climate science - the science doesn't say, and all science doesn't agree, in fact some of the most learned people in climate science totally disagree.
Same thing happened with the coof - some of the world's most learned epidemiologists and virologists etc completely disagreed - and have been proven right FWIW.
Why do we keep doing this? AI isn't going to kill people (of its own accord, anyway). AGI might not even be possible anyway. See a therapist.
"AGI might not even be possible anyway."
We don't know but there is value in considering how we stay in control if it is. How we could counter it used by an adversary without enslaving ourselves to another. A little paranoia is worthwhile. Or do we just hope it smiles kindly at it's naughty pets trying to kill it; would you?
You may find this paper assessing the climate predictions made by Exxon scientists between the 1970s and 2003 against observed values up to 2023.
> Our results show that in private and academic circles since the late 1970s and early 1980s, ExxonMobil predicted global warming correctly and skillfully. Using established statistical techniques, we find that 63 to 83% of the climate projections reported by ExxonMobil scientists were accurate in predicting subsequent global warming. ExxonMobil’s average projected warming was 0.20° ± 0.04°C per decade, which is, within uncertainty, the same as that of independent academic and government projections published between 1970 and 2007.
> ...
> Moreover, we show that ExxonMobil scientists correctly dismissed the possibility of a coming ice age in favor of a “carbon dioxide induced ‘super-interglacial’”; accurately predicted that human-caused global warming would first be detectable in the year 2000 ± 5; and reasonably estimated how much CO2 would lead to dangerous warming.
https://www.science.org/doi/10.1126/science.abk0063#sec-1
This paper is just a comparison between predicted and observed values. It appears that the Exxon models match observations up to 2023, absent other evidence the Exxon models appear to be a starting point for ongoing prediction.
[EDIT - corrected time period of the Exxon research.]
The documents discussed in that paper were not exxon's own findings and research, but were an assessment of the research available at the time, assembled for the exxon board. They made no conclusions of their own, so it is a lie to say exxon "knew" and imply that they somehow suppressed the information.
There were, and still are, significant doubts about the underlying theory.
A model that is based on a false premise can still appear predictive, but the conclusions you draw from it will be wrong.
"But even if you could build a superintelligence or an AGI, and it did everything for us and like, no one had a job, but everyone was just provided a universal basic income from the output of this superintelligence, and everyone could just kind of party 24/7 and not have to work"
Sounds like the citizens of the Culture. And if they get bored they can go and get a job at Contact or Special Circumstances!
The trouble with campaign groups like this is that they get an idea of what they think is "right", and /or for the greater good, etc, then try to force it on everyone else with rabid determination and zeal.
If the population as a whole, or even a substantial minority, were in favour of their chosen objectives, there would already be riots on the streets, waving of pitchforks etc. the fact is, they are not. Sure, some will have vague concerns, others won't know and mostly won't care, and yet more will be just as rabidly in favour of whatever they want to ban.
None of which matters to the common or garden crusader - they are right, by jingo, and the rest of us will do what we're told for our own good. Or else.
As a campaign group, they may have some use in that they raise awareness of possible concerns. But give them any hint if success, let then gain any power, and they will turn out to be just as dictatorial as any other nutjob leader. For the greater good, of course.
But let us suppose that they are successful in imposing their will on the rest of us. They will be happy little maniacs until someone discovers DeepSeek and its ilk. Yes, to their horror, they will find that there are other countries beyond the oceans that can do anything the US can do, and who don't give a damn about campaigns, crusades or the greater good. What will our crusaders do then? Send them a sharply worded memo? Surely the only possible response will be to force the elimination of this foreign threat. And once again the world will be plunged into conflict. For the greater good, of course.
Besides, it's a moot point. AI as presented right now is (sometimes) useful, but it needs heavy oversight. AGI and ASI might be "smart" but it will still do bloody stupid shit and is about as likely cause serious problems as a genius professor who is nonetheless not safe to cross the road unsupervised (and I know we all know at least one of them!)
Like the just stop oil Herberts, they could do with getting a job and doing something useful. Isn't going to happen though.
Anyone who thinks the goverment can stop the development of ai has no idea what there talking about there are plenty of open source models at this point and a big enough open source community that if the goverment did try to outright ban it it would turn into what happaned with prohibition
> "it did everything for us and like, no one had a job, but everyone was just provided a universal basic income from the output of this superintelligence, and everyone could just kind of party 24/7 and not have to work, I personally would find that a little depressing."
What I find a little depressing is that your life is so empty that without being forced to work, your life has no meaning, and all people will do is "party".
A very American trait, unfortunately, where it's drummed into us that our job is what defines us, and unless we put in 100+ hours a week for our corporate bosses, we are losers.
I could see lots of ways where that world, utopian on the surface, turns out to be pretty awful. Some obvious ones include a world of overconsumption because limits on what can be created have been substantially reduced or a world where there is no advancement because the AGI doesn't have a reason to make new things and humans don't have the ability to do so on their own (organized attempts having been replaced by the AGI). It would be a good premise for some speculative fiction stories. That is where it will stay, in my opinion, because I don't think we're going to come anywhere close to that.
Whether or not AGI is possible, I don't think we're going to make it. Existing AI companies seem completely satisfied with a cheap imitation, and if you don't know what your program is incapable of, you can't fix it. If we did get AGI somehow, I don't think we would have either the world where it kills us all or the world where it takes over all work and lets us live in leisure. The creators of it not wanting to die and the AGI needing humans to do some things would put limits on the former, and greed and the desire for power would limit the latter. It could still make things worse or better, but I doubt it would go all the way with either.
It is bad enough now with 'folks' using genAI to make horrors that defy description.
The problem is when a 12 year old with a GPU can make content with no supervision, there is little anyone can do now.
Looked into this, seems that 'models' are the new top shelf magazine.
There are bootable systems you can fit on an 8GB USB stick that can literally plug into a machine, run, accept a prompt and
be unplugged again within a minute. Meanwhile the content generated never sees a 'drive' and is nearly impossible to track.
int main(enter the void)... The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
