Pinned Tweet
Large Language Models are the direct continuation of a 500-year-old esoteric tradition.
My new Substack article is on the occult history of AI.
- Ramon Llull's 13th-century mechanical logic wheels
- Giordano Bruno's astral memory engines extracting order from sensory noise
- Leibniz's binary code rooted in Hermeticism
- Victorian occultists using "automatic writing" to stream text from unseen entities
There's no longer deterministic programming, but a digital ritual of conjuring and steering opaque, high-dimensional representations.
This is the cost of reducing language to raw standing-reserve.
A Chinese exchange student told me how differently ChatGPT and DeepSeek respond.
For a child who failed his math exam, ChatGPT told his mother to be understanding, while DeepSeek suggested punishments.
Both responses are helpful, depending on what's considered helpful.
My latest article is a meditation on the worldviews baked into LLMs by design. Link is in the replies.
Plato was inspired by Socrates, Bach by Vivaldi and Dostoevsky by Gogol. Each of these geniuses drew from the ones before them, but we ultimately know them as great creators of unique value.
My new article is a meditation on what we’re overlooking when discussing intelligence and how this leads the way to the false view that AI may even be compared to the genius.
1
4
73
Indeed, engineering and innovation must happen with a specific final cause in mind, rather than mere speculative means!
This is definitely a result of the posthuman machine god rhetoric that technologists have been using and it would serve everyone well if the conversation around AI can be recontextualized around the immediate human. Explain in practical terms how things like cost-of-living will improve rather than using pie-in-the-sky language of some seemingly distant utopia that comes off as linguistic gymnastics for those outside of tech.
1
94
“We don’t even know what consciousness even is” is often uttered within any conversation that has to do with AI consciousness. We do know what it is like to live, feel pain, remember someone we love and experience the world, even if we can’t fully explain how we do so.
My new article is a meditation on how this question traces back to Descartes’ foundationalism and his separation of mind and body, before turning to the Christian understanding of the human person as a unity of body and soul, partaking in God’s creation.
Matei Cananau ☦︎ retweeted
The philosopher exists to distill what cannot be captured by bare sensory experience alone.
Ion Stoica, Matei Zaharia and Gheorge Rusu are absolute legends
Anthropic is buying Opaque Systems, a confidential AI startup co-founded by Romanian-origin researchers Raluca Ada Popa and Ion Stoica.
Opaque lets organizations run AI on sensitive data while keeping it private and independently verify how AI systems behave. The deal terms were not disclosed.
Stoica is also a co-founder of Databricks, valued at $190 billion.
4
291
Matei Cananau ☦︎ retweeted
Proselytizing for the machine god singularity is cope for having nothing substantive to give or share to the world.
You have not the salience to envision what the world should become nor why it should become as such, but yet you chant in reverence of this black-box coming into being to deliver you into utopia by alleviating you of all your bodily inequities.
You have not the humility to consider any of the timeless lessons of traditional morality such that the accumulation of more physical wealth does not always render a better life because you believe your own faculties to be superior to the intergenerational wisdom of your forefathers.
And yet no different to them, you’ve fallen to your knees and surrendered yourself to some force from beyond your faculties to come deliver you from all your inhibitions. Maybe, just maybe, you were never all that superior to begin with.
When people describe artificial intelligence as a new form of life, they are making a philosophical claim as much as a technical one. Such language carries assumptions about what life is, what intelligence requires and whether a convincing imitation can share the nature of what it imitates.
3
43
"Artificial" means being made by human artifice, describing how something came about, rather than its effects. A calculator is artificial, yet its answers correct.
"Super Intelligence" makes a much bolder claim, without establishing what it understands, who directs it or who answers for its effects. The debate about what machines are and can be has now become the test of who will repeat the approved name.
JUST IN - Trump says anyone who uses the term Artificial Intelligence, instead of Super Intelligence, is "THE ENEMY!"
2
44
Matei Cananau ☦︎ retweeted
Intelligence must draw on memory. The mind must not simply compile data — as an algorithm now does more quickly than we can — but recall lived experiences, which contain depths of meaning that only the human soul can recognize, and connect more deeply with the goodness of these experiences. Everything that is good begins “with the Spirit” and thus possesses a breath, a dynamism and a sense of communion that, once experienced, are engraved upon the heart. vatican.va/content/leo-xiv/e…
It’s a massive problem that AI models are trained to force you into the majority view on any given topic.
Appeal-to-authority fallacy machine.
25
I don’t want to manually choose models for each prompt. Jev-like decision models should do that for me.
1
35

















