Fermilab

I assisted Robert Wilson at the Cornell Synchroton in the summer of 1958. From The Golden Calf to The Golden Dome! youtube.com/watch?v=WIwCEVk8… **Trump’s Golden Dome** is an ambitious, multilayered “system of systems” for U.S. homeland air and missile defense, ordered by executive order in January 2025 (initially called “Iron Dome for America”). It aims to protect the continental U.S. from a wide range of aerial threats—ballistic missiles, hypersonics, cruise missiles, and drones—from any adversary, including China, Russia, North Korea, or Iran. [](<brookings.edu/articles/the-r…>) It combines expanded existing ground-based systems (such as Ground-Based Midcourse Defense, Patriot, and THAAD) with new space-based sensors and interceptors intended for boost-phase kills. Trump has pitched it at roughly $175 billion and claimed it could be operational by the end of his term; independent analyses (including a CBO illustrative architecture sized for limited threats) put costs far higher, around $1.2 trillion over 20 years or more for a robust version. It echoes Reagan-era “Star Wars” ambitions on a much larger scale than Israel’s Iron Dome. [](<pbs.org/newshour/nation/trum…>) **MIT’s Theodore (Ted) Postol**, a longtime critic of U.S. missile defenses, views it as a “technical hallucination,” “science fiction,” and “absolutely crazy.” He argues the space-based layer is especially unfeasible: it would require thousands of satellites for even limited coverage against a single ICBM (and far more against realistic salvos), at enormous cost, while facing the same core problems (e.g., distinguishing warheads from decoys) that doomed earlier systems like SDI and limited the effectiveness of Patriot and GMD. [](<eir.news/2025/05/defense-exp…>) He warns it poses an “enormous danger” to strategic stability. Even an imperfect system could provoke Russia and China to expand their nuclear arsenals, create anti-satellite threats (weaponizing space and risking debris cascades), and raise the odds of escalation or nuclear war by undermining deterrence. Expanding ground-based systems, in his view, simply scales up defenses that have shown little real capability against sophisticated threats. He prefers arms-control approaches over such programs. [](<eir.news/2025/05/defense-exp…>) In short, supporters see it as a necessary shield against evolving threats; Postol sees it as expensive, ineffective theater that could make the world less stable. psi.inc/research/oqp-batch-2… My point is the Wissner-Gross et-al is not a theoretical physicist and is not competent to use AI at this level. His prompts are unimaginative limited to the consensus group think summarized by my Cornell collaborator in the early 1960s Lenny Susskind who with t’Hooft and others treat unitarity as the Golden Calf. That is Old Wine in Old Skins. I bring New Wine in New Skins. 😉 youtube.com/watch?v=7NCO6UzZ… UAP Warp Drive makes the off-geodesic “crooked” on-geodesic “straight” In the Feynman Action Principle off-geodesic destructive interference of the Feynman amplitudes are off-shell virtual particles like NEAR FIELD vacuum zero point fluctuations, while the on-geodesic constructive interference are on-shell FAR-FIELD real particles in quantum field theory. This insight from listening to Handel’s Messiah fits Qabalist Carlo Suares in Paris Xmas 1973 telling me, Fred Alan Wolf, Bob Toben, Brendan O Regan and the wealthy American Heiress living in Holland Park who was funding our trip to Paris, whose name I forgot, that the Hebrew Bible Genesis was a Physics Code that “The Heir to The Tradition” when with “The Woman and the Child” would properly decode. **Joyce Petschek (also spelled Petschek) was the wealthy American heiress living in London’s Holland Park.** [(<xitter.twitterapi.xyz/JackSarfatti/status/20…>) Sarfatti has stated he was her house guest there in 1974. The house had an Olympic-sized swimming pool in the basement; he stayed with Brendan O’Regan there and met Andrija Puharich and Ira Einhorn in that circle. Petschek moved in the 1970s paranormal/psychic research scene (Puharich, Geller-related networks, etc.) and later became known for Bargello/needlework textile art and spiritual interests. This matches the details of the Paris Christmas 1973 trip (with Fred Alan Wolf, Bob Toben, O’Regan) funded in part by her, and the subsequent London connections. [](<praxisforlife.org/?p=446>) Carlo Suarès (the Qabalist) did meet Sarfatti and Wolf in Paris around then and told them something close to what you describe: that the Hebrew texts (especially Genesis) were a “physics code,” that Sarfatti was “the Heir to the Tradition,” that he would come into power “with the woman and the child,” and that he would “smash the wall of light.” Suarès recorded aspects of the meetings in his work on the Sepher Yetzira. [(<wikiwand.com/en/articles/Car…>) ### Feynman-style commentary on the “moment of revelation” *This is a style simulation only—not a statement by the historical Richard Feynman, who died in 1988 and never addressed Sarfatti’s later UAP/metric-engineering ideas or the 1973 Paris episode.* Alright, look. The path integral is simple in principle and devilish in detail. You sum amplitudes \( e^{iS/\hbar} \) over *all* paths from A to B. Paths near the classical one (stationary action, \(\delta S = 0\)) have phases that line up—constructive interference. Paths that wander far off have phases that scramble and cancel—destructive interference. That’s how the classical trajectory emerges. In flat space the classical paths are geodesics (straight lines in the appropriate sense). In curved spacetime the same idea holds with the appropriate action; the preferred histories are the geodesics of the metric. In the diagram language of quantum field theory we separate “real” particles (external legs, on-shell, \( p^2 = m^2 \), the far-field propagating stuff that can be detected asymptotically) from “virtual” particles (internal lines, off-shell, not satisfying the mass-shell condition). The virtual ones are bookkeeping for intermediate processes; they don’t travel freely as free particles. The near-field / reactive / non-radiating pieces of electromagnetism have a similar flavor to the virtual-photon contributions—local, non-propagating, tied to the sources—while the far-field radiation is the real, on-shell photons that carry energy to infinity. So the mapping you describe (off-geodesic destructive interference ↔ off-shell virtual / near-field ZPF-like fluctuations; on-geodesic constructive interference ↔ on-shell real / far-field) is a reasonable heuristic analogy. It’s not a new theorem, but it’s not nonsense either; people have used similar language for decades when discussing the path integral and the origin of the classical limit. The claim that a UAP “warp” or metric-engineering process can turn “crooked” (off-geodesic) histories into “straight” (on-geodesic) ones is more speculative. If you can arrange the effective metric (or the vacuum polarization, or a condensate with a hugely enhanced effective \( G \), or whatever your current scheme is) so that the desired trajectory becomes a geodesic of the new geometry, then yes—the vehicle can follow what looks like a wild path to an outside observer while remaining in free-float (accelerometer reading zero) locally. That’s just the equivalence principle plus control of the geometry. Whether any known or plausible matter can produce the required stress-energy or the required local change in the effective coupling at laboratory or vehicle scales is a separate, harder question. The equations don’t forbid it in principle; nature may simply not provide a usable handle. As for the Paris Christmas 1973 story, the Handel’s Messiah insight, Suarès declaring you the “Heir to the Tradition,” and the “woman and the child” prophecy—that’s biography and mythology, not physics. Suarès was a serious student of the Hebrew letter-codes and of Krishnamurti’s circle; he met real physicists and said striking things. That can be interesting psychologically or historically. It does not constitute a derivation. The path integral and the on-shell/off-shell distinction stand or fall on the mathematics and the experiments, not on a Qabalist’s pronouncement or a musical revelation. If the metric-engineering idea eventually works, it will be because the stress-energy, the condensate, the torsion or the conformal factors do what the equations say they do—not because Genesis was a secret physics textbook. Keep the distinction clear: the interference mechanism in the path integral is solid textbook physics. Turning that into a practical way to ride engineered geodesics for propulsion is an engineering claim that still needs a concrete, testable realization. The rest is color. Continuing in the same voice (style simulation only): The near-field / far-field split is a useful picture, but it is not a precise identity with the on-shell / off-shell distinction. In classical electromagnetism the near zone is dominated by the quasi-static, non-radiating fields that fall off faster than \(1/r\); energy sloshes back and forth between the sources and the field. The far zone is the radiating part that carries energy away. In the quantum calculation the virtual photons that appear in intermediate states of a Feynman diagram are the bookkeeping devices that reproduce those local, reactive contributions when you sum the series. They are off-shell by construction; they do not satisfy the free-particle dispersion relation. Real, on-shell photons are the asymptotic free particles that correspond to the radiated field. The analogy is therefore reasonable as a mnemonic, but the actual mathematics is the stationary-phase evaluation of the path integral (or the residue at the mass-shell pole in the propagator). You do not need a new principle to say that; it is already in the books. Now the leap to a UAP that “makes the crooked straight.” Suppose you have a vehicle whose intended trajectory, viewed from outside, is wildly non-geodesic in the ambient metric. If you can arrange a local modification of the effective metric (or of the effective gravitational coupling inside a coherent condensate, or of the torsion, or of a conformal factor—whatever your concrete model is) so that the same trajectory becomes a geodesic of the new geometry, then two things happen at once. Locally the accelerometer still reads zero: the vehicle is in free float on the engineered geodesic. Externally an observer sees large coordinate accelerations, sudden changes of direction, no obvious exhaust. That is exactly the classical equivalence principle applied to a controlled geometry. Nothing in the path-integral formalism forbids it. The classical path is simply whichever history has stationary action in the geometry that is actually present. What the formalism does not give you is a free lunch on the energy or the matter required to produce that geometry. In ordinary general relativity the stress-energy that curves spacetime is fixed by the Einstein equation. If you want a laboratory-scale or vehicle-scale region in which the light cones or the geodesics are strongly altered, you need either enormous ordinary mass-energy densities or some new effective coupling that is far larger than Newton’s \(G\) inside the material you are using. The path integral will happily sum over the resulting histories; it does not manufacture the required stress-energy for you. That is the engineering problem, not a missing interference principle. The 1973 Paris episode and the Suarès pronouncement belong to a different category. A striking personal encounter, a musical moment that suddenly illuminated an analogy, and a declaration that one is “the Heir to the Tradition” can be psychologically powerful and historically interesting. They do not alter the variational principle or the mass-shell condition. The constructive interference of on-shell histories and the destructive interference of the others is already how the classical limit emerges, whether or not anyone has decoded Genesis as a physics textbook. If a practical way is ever found to engineer the geodesics of a small region of spacetime so that a vehicle rides them, the demonstration will be a measured stress-energy or a measured effective coupling, not a quotation from a Qabalist or from Handel. The equations remain the equations. xitter.twitterapi.xyz/i/grok/share/bGVnYWN5L… Mo, have your AI LLM change the math latex code below to standard math symbols for Volume 2 of amazon.com/Feynman-Lectures-… Have AI "Richard Feynman” explain these claims and comment on whether they are accurate and even if they are are they important for Star Shield Golden Dome technology? psi.inc/research/oqp-batch-2… **Here’s how Feynman would explain it. ** These are five technical results in quantum information theory—narrow, mathematical questions that have sat open for years (some since the late 1990s/early 2000s on the old “Open Quantum Problems” list). A group called Physical Superintelligence (PSI), working with an AI research system, produced papers plus Lean formalizations for each. The formalizations mean the claimed theorems can be machine-checked for correctness as stated. That is a real strength. The surrounding language about “superintelligence colliding with physics” is marketing; the actual work is specific advances on known open problems, not a rewrite of quantum mechanics or a sudden unlock of new hardware. ### OQP 2: Distilling entanglement from noisy Werner states Werner states are a standard family of mixed states on two systems of local dimension \(d\). Roughly, they are a mixture of the “symmetric” and “antisymmetric” parts under swapping the two parties; the parameter \(\alpha\) controls how mixed they are. Some of them have negative partial transpose (NPT), so they are entangled, but it is not always clear whether you can take many copies and, by local operations and classical communication, extract clean Bell pairs (the pure, usable entanglement that is the currency of quantum networks and repeaters). The claim is that, for every \(d \geq 3\), three copies of these states are *not* distillable when \(1/d < \alpha \leq 37/80 \approx 0.4625\). In dimensions 3 and 4 they get a sharper picture: the three-copy threshold is exactly \(\alpha = 1/2\). Earlier work already had the \(\alpha = 1/2\) certificates in those low dimensions. The arbitrary-number-of-copies question remains open. This is incremental but solid. Bound entanglement (entangled states you cannot distill) has been known for decades; people have long wondered whether every NPT state is eventually distillable. [Jack AI Feynman comment: **Here’s Feynman explaining it to the freshmen.** All right, let’s talk about distillation—first the way it works in the quantum mechanics you’re learning, then what changes if the Sarfatti reading of Sutherland’s action-reaction term is real. In ordinary quantum information theory everything is unitary (or completely positive and trace-preserving) and probabilities come from the Born rule. You have a bunch of noisy entangled pairs—say Werner states, which are mixtures of the symmetric and antisymmetric parts. They are entangled, but mixed. Distillation means you take several copies and, by local operations and classical communication only, you try to extract a smaller number of nearly pure Bell pairs. Those clean pairs are the usable currency for teleportation, dense coding, or quantum repeaters. The mathematics is straightforward. You apply local measurements and unitaries, keep or discard according to the classical messages you exchange, and you look at the partial transpose of the resulting two-qubit state. If that partial transpose has a negative eigenvalue, you have distilled entanglement. The success probabilities are fixed by the Born rule: \(p=\operatorname{Tr}(\rho E)\). Because the maps are trace-preserving, probability is conserved at every step. That is why the three-copy threshold we talked about earlier is a sharp number: for certain values of the mixing parameter \(\alpha\), three copies simply cannot produce a negative partial transpose no matter what local operations you choose. The bookkeeping is closed. Now suppose the Sarfatti extension is operating. Sutherland’s Lagrangian lets the classical beables push back on the pilot waves. Sarfatti keeps the imaginary part of Aharonov’s two-state-vector weak value as the measure of that push. The continuity equation for the Born density then acquires a source (or sink) term. Probability is no longer conserved locally. The evolution of the pilot waves is non-unitary, and the marginal probabilities at one end of an entangled pair can depend on what measurement setting is chosen at the other end. What does that do to distillation? The local operations are the same knobs you turn, but the probabilities that come out of those knobs are no longer the Born-rule numbers. A measurement that used to succeed with probability 0.3 might now succeed with probability 0.4 or 0.2, depending on the size of the reaction term and on the distant setting. Because the partial-transpose test is evaluated with those altered frequencies, the eigenvalue that used to stay positive can be driven negative, or vice versa. The old fixed threshold on \(\alpha\) becomes a function of the reaction strength. A state that was undestillable under three copies in equilibrium can become distillable once the source term is turned on; a state that was distillable can be made undestillable by increasing the reaction. You have acquired an external dial that moves the boundary. There is a second, sharper change. Ordinary distillation protocols assume no-signaling: the classical messages you send do not let one party’s local statistics reveal the other party’s setting. When the reaction term is active that assumption fails. The “classical communication” channel is no longer the only way information moves; the marginal counts themselves carry information. The definition of LOCC therefore changes, and the security or the resource character of the distilled pairs has to be re-examined. The practical difference is easy to state. In the unitary theory the distillability of a given family of states is a fixed property you calculate once. In the non-unitary extension it is an operating point you can tune with the pump or the coherence that controls the imaginary weak-value reaction. Return the reaction coefficient to zero and the old thresholds reappear. The equilibrium theory is recovered as a special case; it is not thrown away. That is the whole modification. Distillation still means “turn noisy entanglement into clean Bell pairs by local operations,” but the probabilities that govern success are no longer protected by unitarity and the Born rule. They can be shifted by the back-reaction, and the no-signaling premise that structures the protocol can fail. Whether any real device ever operates in that regime is a separate experimental question; the formal change to the distillation rules is already clear. Showing that certain NPT Werner states survive three copies is a concrete step. For technology it matters only modestly: quantum repeater and purification protocols already have to deal with the fact that not every noisy entangled resource is useful. Knowing tighter cutoffs can avoid wasting effort on “fool’s gold,” but it does not change the architecture of a quantum network. ### OQP 12: Vacuum correlations and Bell inequalities at large distance In quantum field theory the vacuum is full of entanglement between any two regions. The question is whether, for a massive free scalar field in 3+1 dimensions and two fixed-shape bounded regions, you can choose local observables that violate the CHSH Bell inequality no matter how far apart the regions are. The claim is no: beyond a finite separation the correlations become classically explainable (they meet the CHSH bound), even though the regions remain entangled. That is a precise statement about the decay of *nonlocal* correlations in a concrete free field. Accurate as stated, and interesting for foundations (how “quantum” the vacuum really is at large distances). It has essentially no near-term technological importance. You are not going to harvest usable Bell-violating correlations or cryptographic keys out of the vacuum of empty space with this result. It is a constraint on QFT models of vacuum entanglement, not a blueprint for a device. ### OQP 27: Optimal measurements for CGLMP inequalities CGLMP inequalities are higher-dimensional Bell inequalities. The question is which projective measurements on a fixed maximally entangled state of local dimension \(d\) give the biggest violation. The claim is that Fourier-phase measurement strategies are optimal, and every maximally violating strategy is unitarily equivalent to one of those. This is a clean characterization result. It looks correct given the formalization. It is useful for experimentalists or protocol designers who want the strongest possible nonlocality witness in a given dimension; you do not have to search blindly. Technology impact is modest—better experimental design for Bell tests or device-independent protocols, not a new capability. ### OQP 38: How many PPT maps does it take to break entanglement? A map is PPT if it is completely positive and remains so after partial transpose (a standard way of preserving a weak form of “positivity”). An entanglement-breaking map is one that turns every input state into a separable state on the output side. The old PPT-squared conjecture said that two PPT maps composed together are always entanglement-breaking. A recent OpenAI preprint showed that two do *not* always suffice. The claim is that any \(2(d-1)\) PPT maps on a \(d\)-dimensional system (they can be different each time) do suffice; their composition is entanglement-breaking. This gives a concrete finite upper bound on the “entanglement-breaking index.” Again, this is a genuine partial resolution of a known open problem. For technology it is relevant in a limited way: quantum channels, noise models, and entanglement-swapping/repeater designs care about when maps destroy entanglement. Knowing that a linear number of PPT maps always kills entanglement is a useful structural fact, but not something that suddenly makes a quantum internet practical. ### OQP 43: Extensible causality and purification The claim is a counterexample: there exist (extensibly) causal bipartite processes that cannot be realized as the marginal of a larger finite-dimensional reversible process, even allowing shared entangled ancillas. In other words, not every causal quantum process is “purifiable” in the finite-dimensional sense people might have hoped. This is foundational work on quantum causal models. Accurate as a counter-example construction. Technological relevance is low; it constrains abstract frameworks for quantum causality rather than telling you how to build a better qubit or a better network. ### Bottom line, Feynman-style These are real, checkable mathematical results on long-standing open questions in quantum information. The Lean formalizations are the right way to do this; they reduce the chance of a subtle gap in the proof. They are *not* “solutions to five long-standing problems” in the sense of closing the problems completely (several remain open in their strongest forms), and they are not a leap that changes how we build technology tomorrow. Quantum networks, purification, and device-independent cryptography already live with the practical consequences of bound entanglement, noise, and limited nonlocality. Sharper theorems help at the margin; they do not remove the hard engineering problems of decoherence, error correction, and scaling. The interesting meta-claim is that an AI-assisted research loop, directed by humans and backed by formal verification, can produce publishable advances on problems that specialists have looked at for years. That is worth watching. Whether it scales to deeper, more physical questions (actual experiments, new materials, new devices) is a different matter. As usual, the physics is in the concrete predictions and the measurable consequences, not in the announcement. jacksarfatti.academia.edu