#aisb — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #aisb, aggregated by home.social.
-
Speaking of explicit≈consciousness, in another paper I haven't read yet, @[email protected] apparently does the experiment I often describe of training an LLM on strictly 3rd person language. She said at #aisb2026 #AISB it was the stupidest model ever.
www.ingentaconnect.com/content/imp/...
Do Large Language Models Hallu... -
I suggest in q&a that the only thing that makes consciousness "scientifically impenetrable" IS its moral status; we're doing a great job on understanding episodic memory etc. so if Tom succeeds at his project then we'll solve #consciousness quickly and be stuck for decades on valance. #ethics #aisb
-
Tom points out (usefully!) #consciousness doesn't need to be valanced, so doesn't necessarily entail patiency, which I've also been saying since my first #AIEthics paper + 3 consciousness papers, but those ARE more obscure… he calls valanced consciousness #sentience which is a nice clean def. #aisb
-
#AISB @[email protected] attributes "the avoidance strategy" to someone a lot more recent than my 2010 paper that said "We are obliged to build AI we are not obliged to." In an unobscure paper... joanna-bryson.blogspot.com/2015/10/clon...
Clones should NOT be slaves -
In q&a w/ Steve Torrance, John says he assumes my question referred non-phenomenal suffering of twitter. My understanding of #Phenomenology as a concept is sufficiently incoherent I now wond whether it could somehow be applied to a model like the twitter dataset. :-) #philosophy #AISB #AIEthics
-
#AISB @[email protected] concern seems close to one I have with my own 2x2 (tho his skews a bit more towards (anthropomorphic?) concerns of robots' self. Morality is social. I think what he says may better apply to twitter, and any other large model we *can't* really back up. #rightToAutosave
-
Most recent version here (from my #AAMAS keynote on the limits of #autonomy )
#AIWelfare #AISB @[email protected] -
#AISB @[email protected] on AI welfare which he is against on suffering grounds, but introduces a new concern about how self knowledge could create normative vulnerability. IDK if he's seen my 2x2 on this here joanna-bryson.blogspot.com/2017/11/why-... 1/2
Why Robots (and Animals) Never... -
www.ingentaconnect.com/content/imp/... Kristina Šekrst did the experiment I keep recommending of training LLM entirely on third person. Not only did it never say "I", it was "the worst model ever;" she thinks first person is very useful. N=1 but interesting conjecture. #genAI #Consciousness #AISB
Do Large Language Models Hallu... -
Oh /unfin– I forgot. Fisher's fundamental theorem of evolution shows that the rate of evolution is proportional to the amount of diversity available, so this might be the reason complexity in life exists, even if in general it's not a great idea and therefore it's not that much of the biomass. #aisb
-
(eg @[email protected] "intelligence without representation"). I've been saying for years that every intelligence is a mapping of context to action & therefore a model; every intelligence expresses a compressed version of universe. Given your minimalist model of minds (which I like) #aisb …
-
#AISB @[email protected] I get to say this in q&a, he says life isn't evident, I said selection for replicating is, but he says complexity isn't, and I said he's right but Fischer's fundamental theorem (diversity) might get him there.
-
#AISB @[email protected] finishing a great keynote, weirdly claims "no one can explain why the universe represents itself" "no one has predicted minds" what?? All life models the world enough to persist, all intelligence does so more so.
1/2 -
#AISB @[email protected] quotes @[email protected] & Krakauer on #understanding. They might like the work I did 2002-2008 publication that IMO @[email protected] Narayanan & I then validated in 2017 #consciousness #meaning #AIMinds www.joannajbryson.org/s/embodiment...
joannajbryson.org/s/embodiment-v... -
Awesome — younger Hinton gegen present Hinton, via the sadly recently deceased Inman Harvey @[email protected] #aisb #aiethics
-
#aisb Bazil Solomon, Governance Without a Stop Button: Haltability, Ethical Legitimacy, and the Limits of Artificial Intelligence Governance aice-symposium.github.io/Solomon.pdf "Governance cannot wait until consciousness is solved" well that's fortunate. #AIConsciousness
aice-symposium.github.io/Solomon.pdf -
Shouldn’t cut off the #aisb poster “supporting ai research since 1964”
-
I guess I have to go to this, given what I said about #AISB in my Boden Obituaries! Symposium on AI, Consciousness and Ethics 2 July 2026, www.sussex.ac.uk/research/cen... AISB Convention, 1-2 July 2026 (preceded by a day of celebration for Maggie Boden 30 June). aisb.org.uk/aisb-convent...
AI consciousness and ethics: c...
