once you realize that the core dream of language model AI proponents is “owning a slave” you really can’t unsee it.
Conversation
Notices
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 00:34:13 JST clarity flowers - Alexandre Oliva likes this.
-
Embed this notice
Resist (compassionnow@mastodon.social)'s status on Saturday, 12-Oct-2024 01:42:39 JST Resist @clarity it reminds me of the cows in Douglas Adams who have been bred to be able to tell you they’re perfectly ok with you eating them.
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 01:42:43 JST clarity flowers @CompassionNow god that's so real
-
Embed this notice
sara@xoxo.zone's status on Saturday, 12-Oct-2024 01:57:36 JST sara @clarity yeah I’ve tried saying this for a while to colleagues/peers, and it’s been surreal watching people go from, “uhhh, that’s a little extreme” to more recently, “oh, mmm, yeah, it’s looking concerning isn’t it, I just didn’t think that…”
I marvel at how people can imagine all the “wonders” of “AI”, but balk at the very rationally laid out harms and core destructive motives. It’s Charlie Brown and Lucy with the football. “It’s so new, let’s be open minded!”
The motivations aren’t new.
Haelwenn /элвэн/ :triskell: likes this. -
Embed this notice
Going T. Maine (going_to_maine@mastodon.social)'s status on Saturday, 12-Oct-2024 03:46:34 JST Going T. Maine @clarity feel like the fact that an LLM isn’t & cannot be a person is kind of a fundamental distinction there tho? Unless all LLMs are just mechanical turks…
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 03:46:34 JST clarity flowers @going_to_maine I very distinctly did not say "LLMs are slaves".
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 03:49:39 JST clarity flowers @pmatteo1998 okay
-
Embed this notice
pmatteo1998 (pmatteo1998@mastodon.uno)'s status on Saturday, 12-Oct-2024 03:49:40 JST pmatteo1998 @clarity nah, I don't think that's their primary objective, otherwise they would simply make a food delivery app because it's simpler.
-
Embed this notice
Alexandre Oliva (lxo@gnusocial.net)'s status on Saturday, 12-Oct-2024 04:02:47 JST Alexandre Oliva seeing that in asimov's 3 laws of robotics was quite a revelation
I like my computers non-sentient -
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 04:26:27 JST clarity flowers @altruios this is a post about what a kind of person wants, not an ontological statement about a class of technology.
-
Embed this notice
altruios phasma (altruios@mastodon.social)'s status on Saturday, 12-Oct-2024 04:26:28 JST altruios phasma @clarity in the same way a calculator is a slave that does math for you… sure. But words have meaning, and there must be an experience happening for personhood/slavery to apply.
Does statistics dressed in a trench coat have such an experience? No.
Should we think about these things now so we don’t miss when machines gain sentience, yes. We should err on the side of permissive for what makes up an experience and what can have an experience: So we don’t do a slavery.
-
Embed this notice
altruios phasma (altruios@mastodon.social)'s status on Saturday, 12-Oct-2024 05:00:24 JST altruios phasma @clarity it is a generalization based on a miscategorization. I want AI to gain sentience not to have a slave, but to have a friend (which is free to leave, and not be my friend). A capitalistic lens renders that impossible. I reject that capitalism and the imperialist tendencies.
I also believe we have the duty to raise sentience in species we’ve inadvertently repressed through our advancement. This includes artificial ones as well.
Sample size of one: maybe I’m just weird.
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 05:00:24 JST clarity flowers @altruios a fun thing about human communication is that it's built on useful generalizations and that all statements have exceptions. maybe instead of moving through life asking "how is this wrong?" you could be asking "what can I learn from this?"
Nothing personal but if you message me again without including a genuine and respectful question I will block you.
-
Embed this notice
Parke :archlinux: :debian: (parke@fosstodon.org)'s status on Saturday, 12-Oct-2024 06:58:27 JST Parke :archlinux: :debian: @clarity my dream is that AI renders capitalism even more incompatible with humanity.
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Saturday, 12-Oct-2024 06:58:27 JST clarity flowers @parke Ok, this post isn't about you, I guess! You win the prize for being a very special boy!
-
Embed this notice
Annika Backstrom (annika@xoxo.zone)'s status on Saturday, 12-Oct-2024 14:29:15 JST Annika Backstrom @clarity A really important theme is that LLMs, like voice assistants before them, are aggressively anthropomorphized. Lots of comparisons to other "tools" in this thread but nobody is calling a hammer an "assistant" or "companion"
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Sunday, 13-Oct-2024 09:50:17 JST clarity flowers @louis No, I don't think calculators dream about owning slaves!
-
Embed this notice
Louis Ingenthron (louis@ingenthron.social)'s status on Sunday, 13-Oct-2024 09:50:18 JST Louis Ingenthron @clarity Couldn't the same be said of calculators?
-
Embed this notice
clarity flowers (clarity@xoxo.zone)'s status on Sunday, 13-Oct-2024 09:51:11 JST clarity flowers @me weird fantasy bro!
-
Embed this notice
Actually a LLM (me@mastodon.seahousen.eu)'s status on Sunday, 13-Oct-2024 09:51:13 JST Actually a LLM @clarity if you could own a slave without any of the moral issues of owning a slave, would you not?