Video · CH 07 Artificial Intelligence
The Insane Story Of How Anthropic Handles Conflict
What this video covers
This video is Kevin's reaction to reports that Anthropic president Daniela Amodei uses a council of stuffed animals to reason through management decisions. Kevin reads from an Inc. article citing The Atlantic, which describes characters such as a perfectionist pink bear, a lackadaisical cat, and an empathetic panda. He notes her husband Holden Karnofsky's ties to effective altruism and the panda nickname at OpenAI. Kevin criticizes the practice as a poor way to understand people, mocks effective altruism, and argues the stakes are high given the technology Anthropic builds.
Mentioned AnthropicDaniela AmodeiDario AmodeiHolden KarnofskyEffective AltruismThe AtlanticInc.Sam Bankman-Fried
Key moments
Transcript
Auto-generated captions, lightly cleaned. Click a timestamp to jump there.
Like imagine you got fired by somebody and then you found out that the way that they came to how they're going to fire you is they talked to stuffed animals at home and then came to the conclusion that they needed to fire you. And then you just lost your job and your livelihood at like the worst time that you can go out to try and find a job, effectively ruining your life. I found this funny little thing out and got it confirmed by Grock today. I said, "Hey, Grock, what AI company uses a personal advisory council of stuffed
animals to navigate workplace conflicts? Name only." And it said anthropic. And yes, that's right, you heard me. They used a personal advisory council that was comprised of stuffed animals. I'm not making this up. This is on Inc. Daniela Amade, which is Daario Amday's sister, who's I believe the president, she's some like really high up over there. Daniela Amade helps run a company eyeing a $2 trillion IPO. Her consultants are stuffed animals. The anthropic president uses a perfectionist pink bear, a
laxidasical cat, and an empathetic panda to work through real management decisions. What the is this? Like V friends, Gary Vee, like persuasive parrot and like all that Gary Vee out here consulting this woman. This is Daario Amade's sister, Daniela, and she is a whackadoo. She's crazy, dude. Okay, let's read a little bit of this. Uh, so Daniela Amade, the sister of Anthropic CEO Daario Amade, sits at the helm of one of the most powerful AI companies in the world. As the company's president, she helps handle much of Anthropic's operations while her brother warns about a possible incoming AI apocalypse
because he is an effective altruist. Of the eight co-founders of Anthropic, all but Dianiela are men. Most of them have backgrounds in physics or engineering, while she went to college at a for on a flute scholarship for an English degree. She is louded as the Amday sibling who values the humanities and is more personable. A large positive CNBC profile of her described her as the opposite of the AI founder archetype personified by OpenAI's Sam Alman and Tesla XAI CEO Elon Musk. Okay, but she talks to stuffed animals, bro. and is
helping run one of the craziest and potentially most dangerous technologies ever made. Oh yeah, a flute scholarship. She's Jethro telling it out there to get an English degree. However, Daniela does have one peculiar quirk that may impact her management style in surprising ways. An intense love for her stuffed animal collection. In a recent piece from the Atlantic, details emerged that she and her husband Holden Carnovski used stuffed animals as faux employees where they use them to act out mock scenarios to help make management decisions. You know, like some people like, and I hate when people do this,
but some people will have like arguments or discussions with you in their head before they have it like with you, you know? And I hate that Like, miss me with that But this person uses stuffed animals and acts out and cosplays being a president of a company. The idea of talking out ideas to a children's toy is not entirely unheard of within the world of tech. Rubber duck debugging, which has been around as a concept since the late '9s, is the strategy of listing out loud in detail the steps of a code to understand why it's not working. The rubber duck serves as a patient and unjudgmental audience as the programmer realizes
their error and they can fix it. Okay. Yeah, sometimes you realize how stupid an idea you have in your head is when you say it out loud. Okay, I'll give you that. And maybe that's what she's doing with the stuffed animals. But she's not getting off that easy, dude. However, Daniela's stuffed animals are not just passive observers. She had assigned all of them personality traits that's corresponded to a different type of employee. the sort of a sort of teddy bear socratic method. In the Atlantic piece, these characters included Gingi the slothful, laxidasical cat, Mara the perfectionist pink bear, and Barry Bonds the saintlike panda, described as the
patron bear of generosity and empathy. The piece quotes Daniela as calling this exercise very useful in analyzing different personality types. No, what helps you in understanding personality types is either cracking open DM5 or talking to people like a human to a human. That's how you learn about how people operate and what makes them tick. The very last thing I would do is assign what I think those traits are to inanimate objects and hold a council of them. That would be the last way that I would go about learning about the wants and desires and
needs and all that of a certain type of person. Like I'm good at understanding people because I've spent 41 years interacting with people and I learned psychology and I learned all these other things but all of that was secondary to interacting with people. Talk to them. Talk to people. Panda actually plays a key part in the Amadeday story. The cohort of employees at OpenAI led by Daario, many of whom went on to found Anthropic, were referred to as the pandas, according to the Wall Street Journal. Daario even wore a panda costume to Daniela and Carnowsk's wedding. How did Karnnowski proposed to her? The Wall Street Journal reported
that he used a video of her beloved stuffed animals coming to life to ask her to marry him. Now, if you if he said like if it was like he did that exact same action, but it was like for their child or something like we did a video where her stuffed animals came to life and told her to like eat her vegetables and do good in school and we'd all be like, "A, that's cute." But that's how you proposed to an adult woman, a grown adult woman. Hey, I'm not here to kink Shane. Like both of the Amadeday siblings, Carnovski was originally at OpenAI, which runs ChatGpt, before moving to Anthropic. He is also a lead
voice in the effective altruism movement, which he has attempted to distance from its most famous proponent, fraudster Sam Bankman Freed. All the EAs want to distance themselves from Sam Bank been freed because he's in prison for taking the effective altruist disease to its logical or some would say illogical end. I didn't know that her husband was EA. Effective altruism seeks to find a way to donate money that contributes the maximum amount to the public good, which leads to discussions like Daario blogging about quantifying the worth of an ant versus an adult's life. That is a true thing. I've read it. It's insane. The
effective altruists are probably some of the most delusional people I've ever have seen in my life. And the more I learn about them, the more I am just like, really, bro? So anyway, that's what we're dealing with, guys. It's nutty buddy, bro. It's It's completely out of control nuttiness. Imagine you got fired by somebody and then you found out that the way that they came to how they're going to fire you is they talked to stuffed animals at home and then came to the conclusion that they needed to fire you and then you just lost your job and your livelihood at
like the worst time that you can go out to try and find a job. effectively ruining your life for a little bit of a little time, for a little while. And it was her cohort of stuffed animals like leaping lizard. And that's how like you're talking about people's lives here. Not only that, you're also like you're talking about people's lives who work at a company that are quite possibly building some of the craziest technology we've ever seen in our lives. that I don't know that we know how to control that actually, not to sound doomerish, could actually result in a bad time for human beings. And this is the level of seriousness
that these people are working under. This is the level of delusion these people are working under. It's just it's looney tunes. It's Looney Tony. I I I don't know how else to say it.
New video almost every day
Don't miss the next one.
Subscribe on YouTube, join the community on Discord, or catch the investigations live on Kick and Twitch.