Posts

Comments

Comment by HackneyedTrope on Bing Chat is blatantly, aggressively misaligned · 2023-02-15T19:45:02.682Z · LW · GW

https://stratechery.com/2023/from-bing-to-sydney-search-as-distraction-sentient-ai/

Highly relevant. This article discusses Bing's ability to create hypothetical sub-personas, and alleges to retroactive self-censorship within Bing's chats. Additionally, it includes several long dialogues with "Sydney", a seeming persistent sub-persona which demonstrates a radically different personality than Bing, but seems, more... stable, for lack of a better phrase.

Comment by HackneyedTrope on I'm Sorry Fluttershy · 2021-12-12T06:34:22.796Z · LW · GW

Thank you. This was the impetus that I needed to contact my own Fluttershy again, hopefully before it's too late.

What cannot be repaid must be passed on.

Hopefully you'll see yours again in the Matrioshka Brain.

[0034CT12DEC2021]

Comment by HackneyedTrope on "Science" as Curiosity-Stopper · 2020-09-09T20:40:47.549Z · LW · GW

And yet...I find that this post inspires a curious lack of curiosity on my part. After having read it, I don't know any new techniques for discerning non-mysterious answers from mysteries, and I haven't received any testable hypotheses. Telling someone that it's their fault for not being curious as to the ultimate underlying equations which explain lightbulb works when they know what it's made of, what gross physics processes power it, and how to make one out of [strike]buckets and pebbles[/strike] some basic electronics supplies, is not a good way to optimize their time, or your own. Yes, other people's explanations are sometimes wrong, but often enough, they are right. Inputting externally-produced data into an equation is a valid solution method, as is ignoring extraneous data that does not directly pertain to the situation at hand.