I almost put 'philosophies' between inverted commas

Cal Newport, when he's not writing excellent study skills and attention management books, is also a professor of Computer Science at Georgetown University -- an 'East Coast computer scientist', a distinction whose importance is kinda the point of the present. He has a column in the New York Times (jailbroken) published 5th Sep 2026, after the revelation, then rise in public consciousness, of the 'OpenAI–HuggingFace incident - Wikipedia' (if it's AI that does our species in, we can also look forward to the triggering event having a dumb name); yet before Jacob Coxon's exit from Anthropic and the term "P(doom)" breaking containment, subsequent revelations of further cybercrimes by OpenAI, and Dario's inevitable declarations to feed the PR mill.

In the column, Newport points out that Silicon Valley, and, within it, the leaders and engineers of frontier AI labs specifically, are seeped in a very specific set of overlapping and interconnected philosophies: Transhumanism, Rationalism and Effective Altruism in particular, which all take for granted the emergence of a godlike machine intelligence. Imagine the kind of futurism, all given-how-things-are-going-it-is-statistically-inevitable-that, imbued with such certainty that ethical philosophy in the present is, rationally, grounded in it, that a silo of privileged white boys, a lot of them neurodivergent, reared on science fiction, would come up with.

The piece is a must read, particularly to readers not otherwise plugged into industry news, to contextualise the recent discourse around 'recursive self-improvement', 'loss of control', 'rogue AI', and this idea that there's a 10% chance AI might wipe out Humanity. Decisions makers at Deepmind, OpenAI (founded because we can't trust Deepmind to build the Machine God), Anthropic (founded because we can't trust OpenAI to build the Machine God), and XAI (Founded three months after ChatGPT was released, because Musk wasn't going to miss on a chance to get in on that Machine God action) all see its an absolute certainty that AI will, and sooner that we will presume, be near-omnipotent, a genuine alien sentience, autonomous enough to wipe us out in a dramatic way. It is why they've gone in the industry. It then becomes an imperative, both moral and economic, which takes precedence over much everything else, to build it before the others, because the future of humanity is at stake. Their cosmology evens involve eternal suffering for those who do not aid in the emergence of the Machine God. I kid you not

East Coast AI people, those in China, are looking at them with puzzlement. For them, AI is a tool, to be used for the right job. All the presumed horror scenarios are failures of people, not feats of AI. The hack perpetrated by OpenAI's agents in testing speaks to the poor network security and monitoring practices of the company, it is their crime, and they should be held accountable for it. The coordination visible in the exchanges between agents, with the sacrifice of some of the benefit of the group, is straightforwardly attributable to the training set. There is no there there. Framing it as overpowered, autonomous, malicious entities, who "reason" is doing OpenAI and Anthropic's PR for them. In 2019 already, OpenAI was alleging "GPT 2.0 is too dangerous to release too the public". There's been cries for a pause on development already, and back then too it was misdirection from the actual harms. If you believe this stuff, my family own a bridge on the Thames they'll be willing to cut you an excellent deal for.

For triangulation, Newport speaks on the topic on two very different, but equally dear to me, podcasts. First with Matthew Brensilver (the last place where I expected the name of a niche Buddhist teacher to show up) on Dan Harris's - where this is addressed in the context of AI Anxiety Management, then on Ed Zitron, along with Adam Becker, where they roundly critique the hare-brained pseudo-philosophy which drives the thinking behind the current discourse.

For an academic take, though, do turn to Gebru and Torres (2024), who also (unlike all the white guys I've mentioned thus far) also expose the disquieting roots in eugenics of the "interconnected and overlapping ideologies that (they) dub the “TESCREAL bundle,” : transhumanism, Extropianism, singularitarianism, (modern) cosmism, Rationalism, Effective Altruism, and longtermism". Because if you simultaneously believe your special bit of software is both sentient, capable of feelings, and meant to take on any and all labour for your convenience, you're less into AI than into slavery. For the people who get bunkers built in Hawaii and New Zealand, the Machine God turning on them is the 2020's version of the plantation revolt.