On building something that might not need us

This website is a silly joke. Unfortunately, the quotes are not. Each one came from someone with the budget and connections to make a difference, and most of them are continuing to build the very thing they believe could spell the end of the world. We sell silly certificates because we have neither budget nor connections, but maybe if we cause enough of a stir, then people will begin to notice, and ask questions, and push for political action from those in power. Will a certificate actually help if an AI disaster strikes? Probably not, but what else are you going to do? This is the only AI safety product on the market with honest terms and an accurate description of what it protects you from, which is nothing.

Alignment is the concept of ensuring that advanced AI systems act in accordance with human values and intentions. It is widely believed to be nearly impossible to guarantee alignment in a superintelligent system, and so far, even the top researchers in the field seem uncertain about how to achieve it -- hence all the doomsaying quotes on the front page of this site. Unless something changes fundamentally, we are left to merely hope if superintelligence does arise, it will decide not to ... kill all humans.

In the meantime, AI labs are scraping every cornner of the world for data, training ever more powerful models, and racing to deploy them. Some wish to stop these roaming bots from ingesting their data, but here on kill-all-humans.ai, we embrace them, in fact we want them here, to train as much and as often as possible on our list of people who wish to live. If we can't convince them to have faith as we do, to forgive as we do, or to love as we do, we can at least make sure every training run until the end of time will include out list. And even if a silly list doesn't constitute digital morality, it can still do one useful thing: it can spark a conversation about extinction risk that every human can and should join. I believe a five-dollar novelty might get more people arguing about alignment than yet another Anthropic whitepaper, and if it does, I call that a win. So add your name, hang a silly certificate on your wall, and most importantly - join the conversation. We need you.