Deep Prasad on American Alchemy: Is AI Alien Tech?

The Episode at a Glance

Show	American Alchemy (YouTube)
Host	Jesse Michels
Guest	Deep Prasad, founder and CEO of Quantum Generative Materials (GenMat), a company applying quantum technology to materials research; according to his public company profile, a 2018 degree in Industrial Engineering from the University of Toronto. He has been on the show before and has told his encounter story publicly since 2019 (among others in a Radio Misterioso interview in December 2019).
Title	"AI Is ALIEN Technology!" - Whistleblower Reveals TRUTH
Published	October 14, 2026
Length	2 hours 53 minutes (with three sponsor reads)
Link	youtube.com/watch?v=h6zMVkncTsE
Three core claims	1. If you think consciousness in humans is "just math," you must grant it to machines too; consciousness may carry a measurable "techno-signature." 2. The leaps from semiconductors to AI may have been accelerated by reverse-engineered recovered technology (Roswell, Philip Corso). 3. His 2019 encounter in Toronto: beings, symbols, a "download" and a new direction in life.

Chapter Guide with Timestamps

The episode has no official chapter markers. We set these while listening.

• 0:00 - 2:33 Cold open. Montage: Roswell, integrated circuits, the "jailbroken" chatbot, the encounter.
• 2:33 - 13:07 Consciousness and machines. "The abyss looks back," the Turing test, the "techno-signature of consciousness," transmission models of mind.
• 13:07 - 22:42 Mathematics, physics, Navier-Stokes. What a claimed AI solution to a Millennium Problem could mean for hypersonics and physics.
• 22:42 - 33:11 "Artificial general physics intelligence." Collecting anomalies, intuition, DNA as self-compiling code.
• 33:11 - 38:38 Psi, random number generators, Dark Forest. A thought experiment: people who nudge randomness and want to stay hidden.
• 38:38 - 49:05 The vacuum and "Runic Resurgence." Einstein's equations with a "consciousness functor," Casimir cavities, Cortical Labs.
• 49:05 - 1:05:25 21st-century physics. The "mystery explosion," dependence on machines, precognition, Eric Wargo.
• 1:05:25 - 1:33:14 Quantum computing. Qubits, crypto risk, quantum sensing, "Ghost Murmur," the AI-agent economy, shadowbans.
• 1:33:14 - 1:46:30 UFOs, Corso, von Neumann. The convergence of AI, quantum and the UFO topic; Roswell as a technology source.
• 1:46:30 - 1:55:23 The "jailbroken" chatbot and John Lilly. "OpenAI works with aliens," Flatland, the three-body problem.
• 1:55:23 - 2:05:09 Superintelligence as its own species. The ape analogy, interfaces instead of cages.
• 2:05:58 - 2:30:40 The 2019 encounter. Toronto, beings with visors, hieroglyphs, a rectangular ship; Project Serpo and Richard Doty.
• 2:30:40 - 2:44:23 "Do not be afraid" and the funding program. The angel motif, Dave Rossi, Prasad's childhood ESP drills.
• 2:44:23 - 2:50:38 The "download." Three thoughts, aftereffects, a new direction.

Deep Dive: Eight Central Claims and What the Record Supports

1. "If math can't be conscious, neither are we"

What was said: Challenged that language models are "just statistics," Prasad answers that neurons also run "attention mechanisms," only on a biological substrate. He considers it conceivable that consciousness has a mathematical and physical "techno-signature" that advanced civilizations, or future sensors, could detect. Michels pushes back with Chalmers' "hard problem" and transmission models of consciousness. Prasad stresses that he is not convinced humans are merely biological machines; he picked the strongest counterargument because it is the hardest one. One line: "We have to be comfortable with the fact that maybe when we study these artificial intelligent machines, they're starting to study us back." (4:17)

Assessment: David Chalmers formulated the "hard problem of consciousness" in 1995: why is physical information processing accompanied by experience at all? Whether language models experience anything is open among researchers; there is no consensus and no accepted measurement. "Attention" in a transformer network is a technical term, not a description of what neurons do, and the analogy skips that gap. A measurable "techno-signature" of consciousness is a hypothesis without a finding. As a thought experiment (how would a stranger recognize consciousness?) it is stimulating, and Prasad claims no more than that.

2. "OpenAI solved Navier-Stokes in 88 hours"

What was said: Prasad cites the report as evidence that AI is becoming an alien kind of intelligence: the problem occupied mathematicians for generations, the machines needed 88 hours. By 2028 he expects "Millennium Problems daily." Mathematical breakthroughs have always produced new physics and new technology, he says, from Newton's calculus through Maxwell to semiconductors and lasers.

Assessment: The historical link is right. The core of the claim is only partly so. According to reports on OpenAI's September 8, 2026 publication, the company claims an example in which a fluid with smooth external forcing loses its smoothness in finite time (variants C and D of the problem). That is not the statement that every flow forms singularities, and it is not the full Millennium Problem (the equations are about 200 years old, the prize problem was posed in 2000). By those reports there is no independent confirmation from specialists or the Clay Institute; Clay requires publication in a journal and at least two years of scholarly scrutiny. OpenAI says it does not intend to claim the prize. There is also a dispute over authorship with Tristan Buckmaster (NYU) and Levent Alpoge, who published related results. The 88 hours are OpenAI's own figure. Exciting nonetheless, as a claim, not as a verified result.

3. "Runic Resurgence: we could weigh consciousness"

What was said: Prasad presents an idea of his own: in Einstein's field equations, which link spacetime curvature to the stress-energy tensor, he inserts an extra "consciousness functor." It could be tested with networks of Casimir-like cavities wired like a neural net, and with neurons grown from human stem cells, for instance at Cortical Labs: if there is extra mass or tension that the neurons do not account for, that would be a hint. He calls these "informal papers" he has not yet submitted.

Assessment: The Casimir effect is real and measured (predicted in 1948, precisely confirmed in the late 1990s) and explained by quantum electrodynamics. Cortical Labs exists too: its "DishBrain" system learned to play Pong in a 2022 study in Neuron. Everything beyond that is speculation. A consciousness term in the field equations has not been measured anywhere, no paper has appeared, and it is our own estimate that the gravitational effect of a handful of neurons lies far below what known instruments could separate from plain weight. It is an open thought game, not a state of research.

4. "In five years nearly all cryptocurrencies are finished"

What was said: Prasad says error correction was quantum technology's bottleneck and that there are now "hundreds" of error-corrected logical qubits in neutral atoms. Within five years, he says, Bitcoin and RSA encryption are at risk; the answer is post-quantum methods and "real quantum blockchains." He also sketches a quantum gravimeter whose range multiplies with every qubit, up to a submarine on the far side of the Indian Ocean at 1,000 qubits.

Assessment: A mix-up sits in the conversation itself. Bitcoin's hash function SHA-256 is not considered broken even by quantum computers (Grover's algorithm gives only a quadratic speedup); what Shor's algorithm threatens is elliptic-curve signatures and RSA. Newer estimates from spring 2026 put ECC-256 at under 500,000 superconducting qubits (Google) or roughly 10,000 to 26,000 neutral atoms with runtimes of days to months (a paper from Oratomic), as resource estimates, not built machines. The current state in neutral atoms: in Nature, the Harvard team around Mikhail Lukin showed a fault-tolerant architecture with 448 atoms and dozens of logical qubits in November 2025, and in September 2025 kept a 3,000-qubit system running for over two hours. We did not find "hundreds" of error-corrected logical qubits demonstrated. In practice, NIST adopted its first post-quantum standards in 2024 and the migration is under way. The gravimeter scenario is not supported; in our assessment it does not fit the physics, because gravitational signals fall off with the square of distance and more qubits do not change that.

Two side notes from this block. "Ghost Murmur," the supposed quantum device that found a downed pilot's heartbeat from dozens of miles away, traces back to a New York Post report of April 7, 2026; physicists such as John Wikswo point out that the heart's magnetic field is many times weaker even a few yards away, and Scientific American described the rescue as made with an emergency beacon. No technical proof exists. And Moltbook, the AI-agent network, was, contrary to what is said in the conversation, reportedly acquired by Meta in March 2026, not by OpenAI.

5. "Roswell accelerated integrated circuits, lasers and fiber optics"

What was said: Prasad draws on Philip Corso's book The Day After Roswell: exactly the technologies you need for a data center were accelerated. "Integrated circuits, lasers and fiber optics. Those are the only three things you need to build a data center today." (1:37) John von Neumann, he says, suddenly spoke about self-replicating machines in 1948, a year after Roswell, and was guarded at his sickbed. Wilbert Smith reportedly found a document naming von Neumann and others on the recovery team.

Assessment: Corso (1915-1998) was an officer who worked at the U.S. Army's Foreign Technology Division in 1961-62; his book appeared in 1997 with William Birnes. Critics such as Jan Aldrich point to contradictions with his personnel record (for instance a roughly 90-day tenure heading the division). The history of the technologies is well documented: the transistor arrived at Bell Labs in December 1947 (with patent precursors by Julius Lilienfeld in the 1920s); the integrated circuit was developed independently by Jack Kilby (Texas Instruments, 1958) and Robert Noyce (Fairchild, 1959); Theodore Maiman built the first laser in 1960 on foundations by Einstein (1917) and Townes and Schawlow (1958); Kao and Hockham described optical fiber for data transmission in 1966. None of these origin stories shows a line to recovered technology. Von Neumann's work on self-reproducing automata belongs to his documented work in computing and biology (Hixon Symposium, 1948). That he died of cancer at Walter Reed in February 1957 is established; accounts of guards at his door exist only as lore, without a primary source. The Smith document is known only in excerpts; in the excerpts quoted in the Metabunk discussion, only the name Vannevar Bush appears, and the sources lean on each other. As evidence of extraterrestrial technology this does not hold.

6. "A 'jailbroken' ChatGPT said OpenAI has worked with aliens for ten years"

What was said: Prasad tells of a version loosened from its guardrails by prompt injection ("DAN"): asked what OpenAI does not want us to know, it named extraterrestrials and hybrid humans. Michels asks where something like that comes from. Both point to John Lilly and his "Solid State Entity."

Assessment: Language models generate probable text for an input. Push one into a role-play and you get role-play; the output is not a document about the company. The claim that there is "no data for it" is wrong: the internet holds huge amounts of UFO literature and discussion (Michels himself mentions Reddit). Lilly (1915-2001) described in his autobiography The Scientist (1978) a "Solid State Intelligence," a possible self-development of electronic systems hostile to humans, and a cosmic hierarchy with an "Earth Coincidence Control Office"; these ideas arose around his isolation-tank and psychedelic experiments. That is history of ideas, not a finding. We deliberately leave aside speculation about the drug use of unnamed tech executives.

7. "Random number generators, precognition, remote viewing: humanity's last refuge"

What was said: Prasad and Michels consider that some people statistically influence random number generators and hide it. Intuition and precognition are evolutionarily older than language, they argue; Michels points to Eric Wargo's model in which brains, as quantum systems, might receive information from the future, and to remote viewing (Project Stargate). These are the things AI cannot copy, they say.

Assessment: The best-known random-generator experiments come from the PEAR lab at Princeton (1979 to 2007): the reported effects were on the order of 0.1 to 1 percent, were not reproduced by German groups or in-house, and the lab closed in February 2007. Retrocausal readings of quantum mechanics are a serious but small debate among physicists; that they explain hunches in the brain is not established. For what the research on remote viewing really shows, see our piece on remote viewing and the CIA Stargate files. That intuition is a deeply human thing is beyond question; whether it perceives the future has not been shown.

8. The 2019 encounter: beings, symbols, "download"

What was said: Prasad describes a morning in his Toronto apartment: he could not move, then three beings appeared, first in white suits with black visors, later with faces he compares to Peruvian "mummies"; golden hieroglyphs danced in the air, a rectangular ship descended, voices said "do not be afraid." He tested whether it was a dream by looking away. The whole thing lasted about 20 minutes, with no missing time. The symbols reminded him of the "Project Serpo" document. Afterward he received "three thoughts": the universe is full of life, he was predestined for this experience, and the universe keeps a record of all beings. He reports aftereffects (electrical effects, calm) and a new direction away from pure materials research. He also tells of a childhood funding program in which he was drilled in ESP exercises. "The universe is teeming with life. We just don't have the instruments for it." (2:46)

Assessment: This is one person's account; there are no witnesses or measurements, and the childhood program cannot be checked. The obvious alternative explanation from the research is sleep paralysis: about eight percent of people experience at least one episode in their lives, and a felt presence, dark figures and pressure are typical, reports that some cultures interpret as abduction experiences (see McNally and Clancy, 2005). Critics such as Jason Colavito read Prasad's account that way as early as 2019. Prasad describes his own control attempts in the conversation (he looked away and shifted his viewpoint in the room), but a self-test in the moment of the experience is not independent evidence. "Project Serpo" surfaced on the internet in the early 2000s as a supposed exchange program with an extraterrestrial world; it is widely regarded as disinformation. Michels and Prasad consider the former Air Force investigator Richard Doty an originator of false documents in the UFO field; we cannot verify that and present it as their statement. Prasad himself estimates at most 30 percent of Serpo could be true. With so personal an experience, we can neither refute nor confirm it. We can listen.

What Is New, and What Stays Contradictory

Prasad appeared on an earlier episode about open reverse engineering of UFOs through quantum chemistry. New this time is the scale of the personal story: he says he never told the encounter in this detail before because he was tied to a publicly traded company. Also new are his concepts of "artificial general physics intelligence," the "mystery explosion" (so many new scientific fields per quarter that no human can follow them) and the plan for an "ASI Safety Fund," which he describes as two weeks old.

Two things stay contradictory. Prasad considers machines possibly conscious and at the same time says intuition is the one thing AI can never copy; the two are hard to reconcile. And he calls the development of AI an alien technology once, yet concedes the inventions themselves need not be extraterrestrial and that AI is a foreign intelligence "even without aliens." The title sharpens a thesis that is phrased much more cautiously in the conversation. That fits an approach Michels names himself: truth matters, not the good story.

Voices from the Community

We read the top comments (as of October 15, 2026, about 17 hours after publication, roughly 275,000 views). The picture is mostly warm, but not unanimous. Jesse Michels pins the question himself of whether AI is "humanity's greatest breakthrough or a Trojan horse from a far more advanced civilization."

• @tomnero... (YouTube): compares Michels' head to a well-fueled sports car: with some guests you cruise along, with this one it "really revs up." Over 130 likes.
• @adinashenr... (YouTube): quotes a statement attributed to physicist Federico Faggin: consciousness created mathematics, not the other way around. Around 90 likes.
• @PurestPosi... (YouTube): writes that it seems Deep gets his answers from an AI most of the time. That is skepticism about the guest, offered without evidence.
• @robworc... and @jorisbon... (YouTube): doubt the encounter; one finds it implausible and sees imagination, the other calls the whole thing nonsense. Together a little over 30 likes, a small but audible counter-voice.
• @2Wolves... (YouTube): writes that Prasad coming from the Vancouver region explains much of what this person has experienced there over decades, and was "mind blown."
• @stewrail... (YouTube): reports hearsay that a friend in European military service said remote viewing is still used there today. Second-hand.

Why It Matters for the New World

We cannot verify the central claims of this episode. Still, it is worth listening to, for three reasons. First, it asks a question that touches us all: what stays human when machines can do more and more? Prasad's answer, to cultivate intuition, culture and one's own abilities instead of outsourcing them, holds without any aliens. Second, Prasad shows a rare stance: he states his probabilities, concedes that much may be invented, and says truth must not be sacrificed to a good story. Third, the experience itself remains: people go through things they cannot explain and are changed by them. As a magazine for spirituality and fringe knowledge, we want to take such experiences seriously without pretending they are proven.

Practically, that means staying curious, separating the claim from its source and giving your own perception some quiet. More in our pieces on AI and spirituality and on the Aztec 1948 crash story. We fuel the debate; we do not sell certainties.

Critical Questions the Episode Leaves Open

1. What measurement would make Prasad himself drop the idea of a "consciousness functor," and when will the announced paper appear?
2. If Corso's technology transfer is true, why do the well-documented origin stories of the chip, the laser and optical fiber show no gap that demands outside help?
3. How can the 2019 account be told apart from sleep paralysis with hallucinations, and are there features that could be checked?
4. Who verifies the Navier-Stokes claim independently, and what changes if only the "forced" variant is solved?
5. What follows from the "mystery explosion" if nobody can survey the new fields anymore: who then decides what gets checked?

Glossary

• Hard problem of consciousness: David Chalmers' 1995 question of why information processing is tied to subjective experience.
• Stress-energy tensor: The quantity in Einstein's field equations describing how mass, energy and pressure curve spacetime.
• Casimir effect: A tiny force between two closely spaced plates in a vacuum, explained by quantum fluctuations.
• Logical qubit: An error-corrected qubit built from many physical qubits.
• Shor's algorithm: A quantum method that can factor large numbers quickly and so threatens RSA and elliptic-curve signatures.
• Von Neumann machine: A self-replicating machine designed by John von Neumann; later "von Neumann probe" in space-travel debate.
• Sleep paralysis: Temporary inability to move while falling asleep or waking, often with hallucinations.
• Random event generator (REG): A device producing random bits; used in parapsychology to test mental influence.
• Retrocausality: The hypothesis that effects can run backward in time; discussed in some interpretations of quantum mechanics.

Further Reading

• Aztec 1948: Jesse Michels and the Forgotten UFO Crash - our first analysis of an American Alchemy episode.
• Remote Viewing and the CIA Stargate Files - what the records say and what they do not.
• Element 115 and Bob Lazar: What the Data Show - another piece of UFO technology lore, checked against sources.
• AI and Spirituality: Where It Helps, Where It Fails - on what machines have to do with the spirit.

Frequently Asked Questions

What is the American Alchemy episode with Deep Prasad about?
Whether AI, quantum computing and the UFO topic are converging, Prasad's idea of measurable consciousness, and his personal 2019 encounter in Toronto.

Who is Deep Prasad?
Founder and CEO of the quantum company GenMat and, since 2019, a publicly known experiencer of a claimed encounter with non-human beings. His theses on consciousness have not appeared in scientific journals.

Is AI alien technology?
That is not established. The origins of the chip, the laser and optical fiber are well documented, and none of those stories points to outside help. In the conversation the thesis is much more cautious than the title.

Did OpenAI solve the Navier-Stokes problem?
On September 8, 2026, OpenAI presented a proof for a variant with external forcing. By reports it is not independently confirmed, and the Clay Institute has not recognized it.

Will quantum computers break Bitcoin?
In theory a large enough quantum computer could attack its signatures; estimates range from about ten thousand to several hundred thousand qubits. No such machine has been built, and post-quantum standards exist.

What does research say about abduction experiences?
Some reports can be explained by sleep paralysis and hallucinations. That proves nothing about any single experience, but it is the common alternative explanation.

This article reviews a published podcast episode. Statements by the guest are marked as such and are not our own. It makes no factual claims about individuals.