BAN RSI
ANTHROPIC SHOULD BAN RSI
PROTEST TOMORROW AS THEY IPO
500 HOWARD ST @ 11AM
banrsi.ai/
StopTheAIRace retweeted
I wrote a scrappy 1000 word blogpost on *actually trying* to stop the AI Race
lesswrong.com/posts/Nts7RA9D…
March for an AI Treaty happening in Paris on Nov 14!
📢 Grande marche à Paris pour la sécurité de l'IA le 14 novembre ! 🔥
Le 18 novembre, Trump se rend en Chine et doit rencontrer Xi Jinping. Alors le samedi 14 novembre, nous organisons avec @StopTheAIRace une marche à Paris ! luma.com/stop-pjeb
📢 Grande marche à Paris pour la sécurité de l'IA le 14 novembre ! 🔥
Le 18 novembre, Trump se rend en Chine et doit rencontrer Xi Jinping. Alors le samedi 14 novembre, nous organisons avec @StopTheAIRace une marche à Paris ! luma.com/stop-pjeb
StopTheAIRace retweeted
Well done to @StopTheAIRace for yesterday's protest and march to SF city hall via OpenAI, Anthropic, and Dreamforce! Most people cheered and honked in solidarity. They get that humanity-altering, threatening tech is being released without regulations or consent of the people.🛑
StopTheAIRace retweeted
Yesterday, @StopTheAIRace marched from OpenAI to Anthropic to City Hall, demanding that San Francisco declare an AI emergency. 🧵
"The core disagreement was about the inevitability of the race"
"If the race is inevitable you should not contribute."
Replying to @eddylazzarin
0. The core disagreement was about the inevitability of a race
1. I think leadership is way too paranoid about China and the US government. They don’t believe it will be possible to negotiate.
2. They largely initiated the recent race to RSI, because of a belief in its inevitability. Note that OpenAI had to shed a bunch of dead weight like Sora because Anthropic was going for the jugular.
3. Even if they are **not** being pessimistic, I disagree with their consequentialist philosophy. If the race is inevitable you should not contribute.
1
4
463
- DC, US: luma.com/ucnncf7s
- London, UK: luma.com/gjokfxob
- Berlin, Germany: luma.com/3pdqt325
- Brussels, Belgium: luma.com/s8bbku99
- Montréal, Canada: luma.com/d40dp5ed
- Madrid, Spain: luma.com/29vob4mv
StopTheAIRace retweeted
Day 38 of Occupying OpenAI until Sam Altman calls for an AI Treaty
- an OAI employee asked us we had any takes on the "plagiarism thing"
- another employee said he has a lot of respect for us and offered us water
- @ElliotC1 suggesting employees to quit their jobs as they leave the office
StopTheAIRace retweeted
Day 36 of Occupying OpenAI until Sam Altman calls for an AI Treaty
- We visited Guido Reichstadter in front of City Hall, on Day 8 of his hunger strike, asking the board of supervisors & mayor to declare a state of emergency
- We launched a first version of our merch store, where you can buy Stop The AI Race t-shirts & mugs (stoptherace.ai/merch)
- We did an hackathon trying to find more evidence of OpenAI swarms on the internet
StopTheAIRace retweeted
I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.
StopTheAIRace retweeted
Engineers take responsibility for what they make. When an engineer builds a bridge, they take great pains to make sure the bridge doesn’t fall down and kill people. Doctors take the Hippocratic oath: first, do no harm. And programmers…
Programmers move fast and break things.
We never developed an attitude where we take responsibility for what we make. In part, I think this is because what we make mostly doesn’t hurt people. It’s hard to take things too seriously when the worst thing that can happen is that a website crashes.
But now, programmers are building something dangerous. Something truly, terrifyingly dangerous. They are building superintelligence. Most of the people building this technology believe there is a significant chance that it kills everyone on earth. Many who understand this technology, including myself, believe that, if completed, it is certain to kill everyone on earth.
That is not a word of exaggeration. They don’t believe “this will be dangerous” or even “this could kill people,” but “this has a very good chance of killing every single living thing on the planet.”
And how have we responded to this?
We are programmers, and programmers like to write blog posts. So one way we handle this situation is that we write blog posts. We like to make startups, and so some of us made startups. We like to go to parties, and have irreverent discussions with our friends. And so we go to parties and have irreverent discussions with our friends, where we swap guesses about how likely it is that the technology our peers are building will kill us.
The lack of seriousness around this topic, the lack of responsibility for what we are doing to the world, has created rancid strains of thought in this culture. Accelerationists, nihilists, AI successionists. Even the effective altruist school of “if it’s going to be built, we need to be the ones to build it so we can ensure a utopia” that motivates many who work at Anthropic is a deeply wrongheaded, undemocratic, irresponsible approach, compared to the international cooperation we should be pursuing to manage the development of such a dangerous technology.
Imagine an engineer believed that the plans for a soon-to-be developed bridge were flawed, that the resulting structure would be unsound, and would collapse soon after being built. How might we expect her to react? Would she go to cocktail parties with the engineers designing the bridge, discuss the abstract possibility that it might collapse, laugh about what an absurd situation it is that their small group are the only ones who know the bridge could collapse? Would she entertain the opinions of friends who believe that actually the way to make the bridge safer would be to cut its supports, or that in fact it’s good if the bridge collapses, because a larger, more beautiful bridge would be built in its place, and the people on it are simply a necessary sacrifice for this greater end? Would she begin construction on her own unsafe bridge, with the hope that it would be completed first?
Or would she call the mayor?
I know a researcher at Anthropic who made an offhanded comment to me once that it’s strange the fate of the universe has fallen to them, of all people. They felt unqualified to be deciding such things. A few weeks later, after realizing the magnitude of the present crisis and the hopelessness that their company would create a safe superintelligence, I confronted them, saying that continuing to work on making these systems more and more powerful despite knowing the risks was morally unacceptable. They blew me off, but later commented to a mutual friend, with a laugh, that “he’s kinda right though.”
It is difficult to overstate how rotten this community is, one where questions of whether we are building something that will, quite literally, kill every human being on earth, are decided by vibes and cultural membership, rather than earnest attempts at determining truth. Dean W. Ball, head of strategic futures at OpenAI, recently confessed to having systematically misled his audience about the existential risks posed by artificial intelligence because he was worried about being called a “doomer” by his colleagues. I regularly see the concerns of the AI safety community dismissed because they’re being “annoying” about it, because they won’t shut up about the issue. It's such a bummer, after all.
Imagine the engineers at that cocktail party rolling their eyes when their friend brings up for the third time that night that the bridge they’ve designed will kill people. Look, we get it. Now can you just relax and have some fun?
I am disgusted by the behavior of my peers. I am disturbed that we have allowed things to go this far. Those building this technology believe that it will determine the ultimate fate of the universe, that the ones to build it first will have unchecked power over us all, forever. Whether or not you believe they are correct, this is a deeply immoral thing to pursue. Even if they believe they would be benevolent rulers.
When I had fully processed the magnitude of the situation, I got up, went to OpenAI’s New York City offices at 295 Lafayette street, and began a protest. This is where I sit as I write this article. I do not know if this is the most effective method for changing minds, but I did not feel comfortable letting my peers destroy the world without even a paltry attempt at warning the public.
You are in danger. Those building this technology are not adults. They do not hold the same concern for your life as even the man who designed your coffee machine.



































