Interesting how left-coded ‘alignment’ is. The alignment research Anthropic is doing is largely value-neutral. Substantially, the question is: will AI do what I ask it to do, or does it have its own motivations, and will it conceal them? I understand why it is left coded: the people are leftists,
The author discusses the political coding of AI alignment research at Anthropic, noting it is value-neutral but left-coded due to the researchers' political leanings.
Interesting how left-coded ‘alignment’ is. The alignment research Anthropic is doing is value-neutral. Substantially, the question is: will AI do what I ask it to do, or does it have its own motivations, and will it conceal them? I understand why it is left coded: the people are leftists, and ala
Interesting how left-coded ‘alignment’ is. The alignment research Anthropic is doing is largely value-neutral. Substantially, the question is: will AI do what I ask it to do, or does it have its own motivations, and will it conceal them? I understand why it is left coded: the people are leftists,
Interesting how left-coded ‘alignment’ is. The alignment research Anthropic is doing is largely value-neutral. Substantially, the question is: will AI do what I ask it to do, or does it have its own motivations, and will it conceal them? I understand why alignment is left coded: the people are le
The more interesting point: No one on the right even understands that there is value-neutral, technical alignment research. They assume it’s all a leftist scam in the same way social media “safety” teams were purely ideological. For example, here they use left language, and mention (perhaps imp
The more interesting point: No one on the right even understands that there is legitimate, value-neutral, technical **alignment research**. They assume it’s all a leftist scam in the same way social media “safety” teams were purely ideological. For example, here they use left language, and ment