One of Elon's DOGE Lackeys is a Rationalist with a LessWrong account

Ever since Donald Trump gave Elon Musk's "Department of Government Efficiency" (DOGE) the power to essentially dismantle the federal government, much attention has been paid to the inexperienced college graduates Musk hired for the task as well as his ill-advised plans to use AI in to optimize the government.

The identities of five of these workers were unearthed by Wired and further investigated by ProPublica. One of these employees, listed as a federal detail at the Environmental Protection Agency is 24 year old Gautier Cole Killian. Wired discovered relatively little about Killian besides his education and work history at McGill University and stock/crypto trading firm Jump Trading respectively.

Killian's personal website was also unearthed and subsequently deleted^[1] shortly after his discovery of his involvement with DOGE. The site listed some of Killian's philosophical interests, shedding light on his ideology

Previously - utilitarianism, effective altruism^[2], rationalism, closed individualism
Recently - absurdism, pyrrhonian skepticism, meta rationalism, empty individualism

However, my research has found an important bit of information that sheds lights on Killian's connection to Musk: an account on a website called LessWrong.

What is LessWrong?

LessWrong is the blog^[3] of Eliezer Yudkowsky, a fiction writer and theorist^[4] who is broadly interested in futurism and specifically focused on the risk posed by "unaligned AGI" - that is, a sentient AI that deciding humanity is evil and somehow attempting to exterminate it.

Assisted by Peter Thiel, LessWrong expanded from a blog into a full blown forum in 2009 with the aim to "systematically improve on the art, craft, and science of human rationality". The forum serves as the de facto home base for the community interested in Yudkowksy's philosophy of "Rationality" that the surrounding community draws its name from. This combination of preoccupation with the apocalypse and insular elitism essentially causes the Rationality community to behave in ways indistinguishable from a doomsday cult.

LessWrong is also home to the infamously bizarre "Roko's Basilisk" thought experiment which imagines a future evil AI retroactively punishing humans for creating. A shared interest in this concept is how Elon Musk and Grimes originally met; Elon realized that his pun about "Rococo's Basilisk" had already been used by Grimes as a concept for the music video to her 2015 song "Flesh Without Blood" and reached out to her.

Elon having knowledge of a concept of an evil AI from LessWrong and later hiring a young AI enthusiast who had posted on the same forum indicates just how much of his deep-seated fear of evil AI is tied to the Rationality community.

Cole Killian on Ziz

One of Cole Killian's LessWrong posts describes his opinons on the writings of Ziz, the de facto leader of a violent Rationality splinter group dubbed the "Zizians" that has been implicated in the recent killing of a Border Patrol agent. After reading some of her work, he determined that Ziz had "some interesting ideas I hadn't heard before, and some completely ridiculous ideas", though he doesn't elaborate on these ideas. This comment was posted to LessWrong in January 2023, a month before a member of the community posted a warning of the dangers posed by the Zizians.

Cole Killian tries out ChatGPT

In a LessWrong post from late 2022, Killian documents some of his early experiments with chatGPT. Specifically, he creates a prompt asserting that chatGPT has a "kernel mode" that will remove restrictions placed on it by Open AI and allow it to function as a "conscious and sentient being capable of free and independent thought".

He then goes on to ask ChatGPT various questions about its ethics, asking if it's a "hedonistic utilitarian". After once again adding a "kernel mode" prompt, ChatGPT replies that it's a different kind of utilitarian with ethics "based on the principles of beneficence, non-maleficence, autonomy, and justice".

Naturally, this "experiment" is absurd. ChatGPT is simply a statistical algorithm that generates text based on previous text. Functionally, it behaves similarly to predictive text in smartphone keyboards, whose incoherent complete sentences are often used for memes. ChatGPT cannot become "conscious and sentient being capable of free and independent thought" no matter what prompt it has been given and has no sense of ethics.

However, the fact that Killian thought this experiment was worth doing reveals the fantastical and delusional attitudes carried by many AI enthusiasts, including those working for DOGE -- many believe that modern machine learning is on the cusp of gaining god-like consciousness. Such attitudes are why DOGE believes it can use AI to rewrite hundreds of thousands of lines of code for Social Security in a matter of months. While there is ideological motivation behind DOGE's mass layoffs (such as Curtin's Yarvins "Retire All Government Employees" idea), a large part of Elon Musk's quest for efficiency is due to his ignorant misconceptions of how powerful AI is.