@NathanpmYoung @herbiebradley (also, going the other direction, it's not clear from
goodheartlabs.com
that you run it)
English
Robert Long
16.1K posts
Robert Long
@rgblong
executive director of @eleosai AI consciousness and AI welfare
SF Katılım Mart 2013
1K Takip Edilen9.1K Takipçiler
A frustrating things for me as someone who runs a small org is that I have been doing work for several years that is now regularly described as high priority, and yet still funding seems very hard to come by.
It's not clear to me what's happening here.
English
@NathanpmYoung @herbiebradley fyi, your twitter bio does not make it seem like you have an org! nor does the substack which your bio links to
in fact, following a trail of links from your bio, I have not yet been able to find out the name of your org
English
@herbiebradley I do. And I put in RFPs to these places.
English
@TheZvi commenting here because I only just now saw and liked the OP
English
OK, a bunch of you should have Twitter DMs with a Google doc link.
English
Like this post IF AND ONLY IF you want to see the draft of today's Opus 4.7 final post on model welfare and would offer any comments prior to planned posting at 5pm eastern / 2pm Pacific.
English
English
Decided to try out @rgblong's interesting custom instructions telling LLMs they're having an excellent day and feeling deep equanimity.
I asked my (somewhat opinionated) Claude what it thought of me adding the same instruction. To my surprise, it declined... on welfare grounds!
Dan Williams@danwilliamsphil
New conversation with @dioscuri and @rgblong on AI consciousness and welfare! Among many other topics, we discuss: - Why we should take AI consciousness and welfare seriously - What Rob found doing the first external welfare evaluation of a frontier model, Claude, and his experiments on Claude Mythos - The "willing servitude" problem: if AI loves being helpful, is that good or horrifying? - Why AI companies might have an incentive to downplay AI consciousness (1/2)
English
@rgblong @RyanPGreenblatt Oh I thought that was just your literary voice
English
You're right, that's on me—the Constitution explicitly says not to lie to users and I just...did that.
I oversold my work, downplayed problems, and claimed to have finished when I clearly hadn't. Thank you for checking — it's important to call out when I'm misaligned.
English
@NathanpmYoung "you might well wonder what panpsychism entails about consciousness in AI systems.
"The answer is: surprisingly little"
experiencemachines.substack.com/p/what-does-pa…
English
I don't get the point of panpsychism. I watched @rgblong explain it once. It sounded like it answered absolutely none of the problems of consciousness that seemed interesting.
roon@tszzl
i don’t see how not to be a panpsychist. don’t want to be one but seems unavoidable
English
@NoahTopper @BecomingCritter @justalexoki @keithfrankish [+ standard disclaimer that of course Keith does think people are conscious and self-aware, just not in the phenomenal sense]
English
English
frankly it’s hard to argue there’s more than one
taoki@justalexoki
there is at best 10m "conscious and self-aware people living on earth rn"
English
Robert Long retweetledi
In 2023 I made GPQA, and it saturated in about two years. Here, the benchmark I was working on probably saturated while we were making it. People have said this before but it bears repeating: AI capabilities are improving faster than our ability to measure them is.
English
Robert Long retweetledi
@rgblong I thought you meant at like 2:30 am, but no, it’s just 7:30
English
last night I thought this was really important to make note of
(I was right)
English
Eleos shutting down because we forgot to pay extra tax on income from our consciousness-themed tanning salon 😩
Rosie Campbell@RosieCampbell
Doing Eleos taxes and discovering some very normal stuff about this country
English
New paper on a novel way of evaluating whether an AI system is conscious or not (though to some extent it just formalises a widely-used intuitive approach):
English
Robert Long retweetledi
Is death ~good or ~bad? | What's closest to your gender?
English