![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Mar. 3rd, 2026 06:09 am
stream of consciousness
To be discussed later - realizing that my current favorite writers are all Asian American men - Ed Yong, Ted Chiang, Ken Liu, with very different sensibilities, but all with a deep thoughtfulness to them.
I just finished reading Ken Liu's translation of the Dao De Jing (Tao Te Ching), which takes a less traditional approach, and sprinkles throughout the chapters anecdotes and a little history and context. I really liked it.
Came across his Substack, and this entry in which he talks about his frustrating experience of trying to get his AI assistant to do what he wants it to do and failing. Made me reflect on how even the smartest people I come across, especially if they are pro-AI, lapse into attributing qualities to AI that are human and not machine. Like...the AI is not lazy. The AI is incapable of processing the request you keep trying to make it do, and is also incapable of telling you when it cannot do something. Like, it just sounds exactly like this Meta exec who went viral when her "agentic AI" went ahead and tried to delete her entire inbox, ignoring instructions to stop.
In other AI news, apparently there was a mistranslation of a Korean analysis of the war, where the original text said that the US "could" be moving military assets out of South Korea, when the AI English translation said the US was actually "considering" moving it, and the mistranslation went viral.
I can't tell if this is the administration trying to get out of taking the blame or they're just stupid (why not both?), but I guess Rubio et al are going with "we bombed Iran because Israel was going to do it anyway and we needed to defend Israel first", which is now prompting outrage from...Gavin Newsom and Sen Ruben Gallego...of all people.
I just finished reading Ken Liu's translation of the Dao De Jing (Tao Te Ching), which takes a less traditional approach, and sprinkles throughout the chapters anecdotes and a little history and context. I really liked it.
Came across his Substack, and this entry in which he talks about his frustrating experience of trying to get his AI assistant to do what he wants it to do and failing. Made me reflect on how even the smartest people I come across, especially if they are pro-AI, lapse into attributing qualities to AI that are human and not machine. Like...the AI is not lazy. The AI is incapable of processing the request you keep trying to make it do, and is also incapable of telling you when it cannot do something. Like, it just sounds exactly like this Meta exec who went viral when her "agentic AI" went ahead and tried to delete her entire inbox, ignoring instructions to stop.
In other AI news, apparently there was a mistranslation of a Korean analysis of the war, where the original text said that the US "could" be moving military assets out of South Korea, when the AI English translation said the US was actually "considering" moving it, and the mistranslation went viral.
I can't tell if this is the administration trying to get out of taking the blame or they're just stupid (why not both?), but I guess Rubio et al are going with "we bombed Iran because Israel was going to do it anyway and we needed to defend Israel first", which is now prompting outrage from...Gavin Newsom and Sen Ruben Gallego...of all people.