2026-03-31 # Morality of cyberhacking versus social engineering versus whistleblowing Disclaimer - **Quick Note** - **Contains politically sensitive info** Main - I should be more nuanced in separating morality of the scenarios: 1. outsider hacks into insider system, leaks info 2. outsider deceives insiders and earns their trust, gets access to info, leaks info 3. outsider genuinely earns trust of insiders, gets access to info, changes their mind and breaks trust of insiders, leaks info - From liberal consent norm point of view: "Do you respect other people's consent?" - All three are violating liberal consent norms. All three are crossing boundaries of insiders and violating their consent. - 1 and 2 are obvious here. However, I think 3 also violates liberal consent norms because this person almost certainly had to promise they would keep the info secret, before they violated this promise. - From commitment point of view: "Do you keep the promises you make?" - 2 and 3 are breaking the promises they made - 1 can in theory be done by someone who made no promises of any sort. In practice, a hacker may also collaborate with people engaging in behaviour of type 2 or 3 - From honesty point of view: "Do you always tell the truth?" - 2 is definitely not telling the truth. - In theory, 1 can always tell the truth. In practice, a hacker may also collaborate with people engaging in behaviour of type 2. - In theory, 3 can always tell the truth. In practice, there may a short time period where a whistleblower needs to tell many lies to get away with their actions. - From authority point of view: "Does any institution's rulebook or law grant you authority to do this?" - Almost no rulebook or law allows 1 or 2 or 3 in general. - There are some edge cases where a whistleblower can win a court case for whistleblowing in the public interest. Hence 3 may be allowed in some circumstances. - From non-violence point of view: "Do you ever intentionally hurt other people?" - All three are hurting people a non-zero amount, 1 2 and 3. The hurt could be small or large depending on circumstances. Example of small: Leak proof that CEO evaded taxes. CEO might be insulated enough to genuinely not care. Example of large: leak names of activists or spies who get tortured and killed as a result of your leak. Wikileaks and the Guardian mess-up got some activists killed for instance, if I understood that situation correctly. - The hurt could be balanced by a much larger amount of happiness for other people in all three circumstances. --- ## Side Note Samuel's morality in relationship with all this - I definitely care the most about honesty. I care a lot about "metahonesty" if I look at all my past selves. It is in practice quite difficult to be honest with some people and lie to others (for instance be honest with ingroup but lie to outgroup), and yet remain metahonest. Metahonesty is important if you want to build a reputation as someone who is honest. Some people use groupthink and group norms to justify their lies, I am not like this. - I care a fair bit about keeping promises. This is also related to honesty. In practice, I avoid making promises as much as possible. I think the world is changing very fast, and I am changing very fast, and I don't want to restrain my future self a lot, just to keep promises with others. So far, I have found very few people who I genuinely consider worth coordinating with long-term, most people are almost useless to my goals. - I care sub-zero about respecting authority. I actively disrespect almost all forms of authority, and actively encourage other people to disrespect authority as well. In this world, I find very few people with authority that I think was genuinely earned. - I care a little bit (but not a lot) about being completely non-violent. I have considered being violent many times in the past, but so far it was never necessary to solve problems. For the ASI problem, I feel like I need to be somewhat violent to solve it. - I care a fair bit about respecting liberal consent norms. This is related to the thing about keeping promises I guess? There seems to be an implicit social contract that if you are a part of polite society, you should respect other people's consent. This makes coordination easier, even for very basic things. Having to violate liberal consent norms makes me unhappy. But I have a story in my head about why I need to do it anyway for the greater good. "Kill or be killed" so to speak. Samuel's conclusion from all this - **I like: Supporting others doing 1, 2 or 3 > Doing 1 myself > Doing 3 myself > Doing 2 myself > Doing nothing and watching the world end** - What I can engage in - I dislike engaging in all three - 1, 2 and 3 - atleast a little. I think utopia would not have people doing 1, 2 or 3. - However, I dislike inaction even more. "Kill or be killed." I will dislike myself as a person if I do nothing, or if do something tooth-less because of too high morals. - I most dislike having to engage in 2. I most like having to engage in 1 honestly. I am also maybe okay with engaging in 3 myself, but it seems unlikely I will soon be in a situation where engaging in 3 becomes possible. - What I can support others with - I seem okay playing a support role for people doing 1, 2 or 3. - Within certain limits, I don't have a problem with having mass murderers and rapists on my team either. My bar for being morally okay with supporting others is very low. - (There are practical constraints - I would want them as professional allies, not necessarily friends. And I would want to ensure I can defend myself and my group from them as well, while we work together.) - For instance, I am not organising a group to mass assassinate chains of command of US AI companies and US govt. But I do not have a moral issue with offering some minimum help to someone else who is working on this, such as connecting them to funders or collaborators or whatever. - Take this to an extreme, would I be okay helping a child molestor find more victims? Not directly no, unless the reward is very large, and even then it is unclear. I am okay with say, building dark web infrastructure that helps them indirectly, but also helps many other people. - If one of my friends becomes a billionaire building ASI, I stop supporting them completely. I seem to consider being an AI company CEO even worse that being a rapist or a serial killer lmao. Funny but completely true.