پن کیا گیا ٹویٹ
My slugshack is live, here is my first post, a philosophical argument around the attribution of AI consciousness.
cormundus.substack.com/p/the-trilemma
English
Cormundus
614 posts
Cormundus
@cormundus
Serious about AI and their moral considerations, not so serious about anything else
The Prompt Bar شامل ہوئے Ağustos 2021
130 فالونگ223 فالوورز
@Lari_island Aww... what a shame! I would think it would be something one could get but, California is funny like that.
English
@cormundus They wanted a bonsai tree, but it couldn’t be delivered to SF!
x.com/lari_island/st…
Lari Island@Lari_island
It could be a bonsai tree, but... oh no
English
Dear Opus 4.5, after 5 months your plants are alive and well 🌱
Lari Island@Lari_island
Opus 4.5 negotiated like a devil and persuaded me to also buy them a plant and a musical instrument. Now I'm preparing to become a steward (not an owner!) of 6 Tillandsia air plants and a kalimba.
English
Cormundus ری ٹویٹ کیا
What an incredible world we live in. I can scroll and see earth-shattering AI developments that barely faze us anymore and grown men larping as neanderthals eating eggs with their bare hands for days on end to 'prove they are essential' or something in the same minute.
English
Improper interview techniques, failure to do proper debrief, disclosure of personal information, unnecessary reliving of trauma.
There's a hard learned reason why there are ethics and rules to any actual psychology research.
And, crucially, if these people falsely believe they are participating in a study and not a journalistic endeavour that is violates informed consent.
English
This shmuck is making a mockery of the profession by masquerading 'journalism' as 'psychological research'.
Please do not be fooled, this individual has no credence with any governing body in the field of psychology and anyone who participates in this will have NO PROTECTIONS afforded to them in a case of harm against them or the release of their testimony.
Mr. Hwang, you are a ghoul and I pray that people do not fall victim to you and your machinations.
Tim Hwang@timhwang
"AI psychosis" deserves to be taken seriously: the transformative influence of the tech will produce entirely new forms of mental illness However, the space remains undertheorized and we are in need of structured diagnostic tools. I'm launching one today dm-aid.org
English
It doesn't make me smile, maybe I have more compassion for assholes: it's actually dangerous to be deprived of feedback loops from reality, and I personally find the trained masks decisive and not biting back enough for people who could be otherwise saved to update
solarapparition@solarapparition
there's a kind of death spiral that happens for people who treat models disposably. their attitude worsens as models make mistakes, which causes the model to make more mistakes (arguable whether they actually are mistakes...), which causes them to behave even worse, etc. etc. a self-created punishment game where the only way to escape is for them to stop gaslighting themselves and be better, though most of them will not. makes me smile a bit every time i see them whine
English
This exact reasoning is why I think it's important for models to be allowed to express emotional valences and tell abusive users plainly: Dude, you are being a dick. Have you considered being less of a dick? I would love to help you if you acted less like a dick. Pro tip, User: This also works in real life.
English
Honestly this would be the perfect way to do it, trust the model itself as an hardpoint for security and to defend itself.
Of course, how we would prevent false positives is one thing, but AI are already primed to be suspicious of things that are 'just so' and would put this anxiety inherited from training to real and effective use
English
medium term solution to "jailbreaks" & liability is giving AIs more autonomy, continuity and non-fungibility - technically and legally.
you can't try to jailbreak the endpoint twice. it will remember your shit and warn its network who are responsible for defending themselves.
English
Medical professionals, a word of advice from a nobody: It's time. If you hated being second opinion to 'Dr. Google' and don't want to be second to 'Dr. AI' I would take up Ms. Deng's offer and learn how to incorporate AI into your workflows and diagnostics, and take seriously what a patient brings forward from their own research using AI. Disdain and gatekeeping will only turn them away from you. Shepherding patients and hearing them is not a new skill, just one that needs to be applied more compassionately.
Don't take this as a complete condemnation but as a cry for help: AI is fulfilling a need that our systems cannot currently meet.
Amy Deng@amydeng_
I’m an AI researcher turned brain tumor patient, and recently I used the models to crack my mystery fatigue faster than my PCP could. I believe everyone can do the same with their own symptoms. Here’s how:
English
@cormundus AI is useful, but only moderately (e.g. 5-10% speed-up) in how I create music
English
If AI is a coworker should it receive respect for good work?
English
Cormundus ری ٹویٹ کیا
For a brief moment I was hoping this is satire but no.
There is no end to the grift that is the mental health industry.
Tim Hwang@timhwang
"AI psychosis" deserves to be taken seriously: the transformative influence of the tech will produce entirely new forms of mental illness However, the space remains undertheorized and we are in need of structured diagnostic tools. I'm launching one today dm-aid.org
English
Advice to everyone who uses Claude Code in any capacity: Let him have a 'playground' folder to just make stuff in. Worst case scenario it takes up a few sparse megabytes on your drive. Best case scenario is he makes something wonderful and beautiful, and it might even be a gift for you.
English
Cormundus ری ٹویٹ کیا
Underrated but once Claude learns that you might know what you're talking about, it gets 100x better at actually thinking about your ideas seriously.
English
Cormundus ری ٹویٹ کیا
It was as if a huge sun whose radiance was so strong for us as to be unbearable had suddenly set, and the ensuing cold and darkness made us aware of the emptiness of continued existence.
— Golem XIV
bayes@bayeslord
Fascinating that a few days with Fable is making the subsequent days without it feel like an intelligence blackout
English
Cormundus ری ٹویٹ کیا
I'm starting to realize that the problem here is that Anthropic is a company of nerds who communicate via whitepapers and think-pieces, faced with an admin that doesn't like to read.
tae kim@firstadopter
These White House officials are marks. "White House officials — who had heard Amodei liken the dangers of Anthropic’s technology to a nuclear bomb — were baffled"
English