2026-08-27
Why you should consider not befriending or dating people working at frontier AI companies in the US
Disclaimer
- Quick Note
- target audience - people who are either already in or considering being in friendships or romantic relationships with people working at frontier AI companies in the US
Disclaimer on where this advice is coming from
- I am not an expert on either friendships or romantic relationships. I am not someone who is in top 10 percentile in terms of being successful at either friendships or romantic relationships. So take anything I say with a grain of salt.
- I literally think your friend/partner/etc working at the US AI companies threatens to kill me and everyone I care about. I am not your friend and I am not on your side, per se. I am not someone who deeply cares about what is in the long-term interest of either you or your friend/partner/etc. I am sworn enemies with your friend/partner/etc and plan to fight them to the death if that is what it takes.
- However, despite being "not on your side", I still think I may have non-zero useful information for you, that you may want to atleast consider once before taking your own independent decision.
The reason is not persuasion
- I am not claiming that if you refuse to befriend/date them, or cut off your existing relationship with them, then they will stop building artificial superintelligence. I am not giving this advice to you as a way to persuade them to stop.
- If you know them personally, you probably have a better model than I do, of how they will react when you state this to them.
- I am sure some people can persuaded this way. But also, a lot of people cannot be persuaded this way. There is a decent chance they will just resent you and continue building artificial superintelligence anyway.
- There is also a decent chance they will be able to find new friendships or romantic relationships even after you cut them off. Society has a lot of moderate people who will not get radicalised in the same direction as me or you, and hence this person can probably still find new people.
The reason is conflict of interest
- By befriending or dating this person, you are setting up bad incentives for your self. You are making it harder for yourself to think clearly about the problem of risks from artificial superintelligence. You are making it easier for yourself to agree to reasons why accelerating AI is fine actually.
- By befriending or dating this person, you are making it harder for yourself to accept that you are in the social circles of the most dangerous mass murderers in human history. I know comparisons to Nazis are cliche but I genuinely believe that in a hundred years from now, if homo sapiens is still around, there is a decent chance that the people building ASI today will be looked at as far worse than the Nazis or the Stalinists.
- By befriending or dating this person, you are increasing the likelihood that you end up tying your source of motivation, your source of validation and approval, even your very sense of self and identity, with this person. All this makes it harder to think clearly about the problem of risks from artificial superintelligence.
- By befriending or dating this person, you are making it harder for yourself to take actions that hurt this person's interests. I support a pause on further AI research. This requires shutting down the AI companies.
- By befriending or dating this person, you make it harder for yourself to agree that the AI companies should be shut down, and these people's salary and associated high status title should be taken away. If they have invested their sense of purpose, curiosity, motivation, identity, etc into their job, you make it harder for yourself to agree that all this should be taken away from them.
- By choosing to stay silent about this problem, you are likely still contributing to chilling effects and distorting everyone else's perception of how good or bad it is to work at the AI companies. Maybe consider being honest and not deceptive with all the other people in your social circle, if you genuinely believe artificial superintelligence is dangerous to build.
- As a concrete example of this, most people living in the US are also facing chilling effects when talking about how the US executive branch and intelligence agencies are now involved in the race to build artificial superintelligence. I am able to openly talk about this because I don't live in the US.
- I am not an expert politician, but my guess is that shutting down the AI companies is also not going to be so easy, morally or legally speaking. Probably it is going to require humiliating these people in public, and imprisoning atleast a few of these people, as an example to set to everyone else. You can read about the Nuremberg trials and Germany's attempts at denazification following WW2 for a (admittedly cliche) example of this.
Subscribe
Enter email or phone number to subscribe. You will receive atmost one update per month.
Alternatively, you can subscribe via RSS, or a third-party service like blogtrottr
Comment
Enter comment