@BrendanNyhan@toddntucker Wanting to retain a craft element (also mentioned in the tweet) doesn't amount to parochial career assessment concerns, though, and can be important to the long-term health of a field.
@yuntiandeng Do you think that if novelists used AI and only prompted revisions rather than writing anything themselves, the result would be better novels?
@harryjwang "Whether the words were typed by a human matters less than the substance behind them"? AI didn't just type the words, though, as if it were taking dictation. Similarly, when AI writes code, it doesn't just do the typing. It makes decisions about how ideas should be expressed.
'The Landing in Summer.' Mary Dawson Elwell emerged on the British art scene as a painter of interiors, or more accurately, 'portrait interiors,' of her Yorkshire home in the 1930s. Her watchwords are stillness, silence, containment, privacy, meditation, illumination.
@artinsociety Thanks for this Philip - I think the โreflectionโ of the couple may be in a companion piece to this work also by Spencelayh
Both fascinating in their detail and content.
@erikphoel In any case, such systems don't understand anything. It at least used to be possible to see that fairly easily just looking at the text they produced. For a while, the system would respond as if it understood, and then it would say something that showed it clearly didn't.
@erikphoel Although I don't think he explained things very well at that point, the Word doc paragraph was connected to a larger argument and was specifically about docs "containing a conversational transcript".
@DoloresGMorris@DouthatNYT As DVDs and digital photos show, a sunset's colours and shapes can be represented by 1s and 0s in considerable detail. So why would an AI using a camera as its 'eyes' not count as looking at a sunset?
@DoloresGMorris@DouthatNYT I read it with interest. I agree with your overall point about Claude describing things it hasn't seen. I'm not sure how your argument about cameras vs eyes is meant to work or quite what the conclusion is meant to be.
@JennyAscendant@akuareindorf@OakesMarianne They wanted parts of it withdrawn, and said which parts and why, which is more than most of the politicians currently criticising the code seem willing to do.
@academic_la@aziz0nomics What do you think will happen the next time they're attacked if they loose their ability to wage war? Or do you think they'll still be able to wage brief wars, just not an endless one?
@AriSchulman If LLMs have the power of speech, I'm happy to say it doesn't require conscious thought, just as playing chess extremely well turned out not to require it. None of that means conscious thought isn't involved when humans do those things.
@HannesThurnherr@AdrienneLaF We know from out own experience that our consciousness isn't epiphenomenal. If I'm consciously imagining a cat, that plays a causal role in my answering "a cat" when asked "what are you thinking about?"
@cacophonicadent@samzliu The reason is that we understand how computers work well enough to know that an AI's behaviour is due to the mechanical operations of the machine on the program and data it's given. Their consciousness (if they have any) can't change that, so it would be epiphenomenal.