Public Access to Patents

Good evening. Yes—quite a lot of the U.S. patent system is public, but there’s an important dividing line between published applications/patents and material that remains confidential while an application is pending. I’m checking the current USPTO access points so I can give you the practical boundaries rather than just the general rule.

Good evening. A surprisingly large fraction of the substantive U.S. patent record is public. Once an application enters the public side of the system, you can often see much more than just the finished patent.

For issued U.S. patents, the USPTO provides records going essentially back to the beginning of the U.S. patent system in 1790. Modern patents are full-text searchable, and images of older patents are available as well. Published patent applications have been systematically available since March 15, 2001. citeturn109139search12turn109139search15

And the interesting part is that you aren't limited to the glossy final patent. Through Patent Center, for a public application you can often examine the file wrapper—essentially the working history of the case. That can include the original application, drawings, claims, amendments, examiner Office Actions, applicant responses and arguments, cited prior art, Information Disclosure Statements, notices, and other correspondence. In other words, you can watch the argument between the inventor's attorney and the patent examiner develop over time. citeturn449600search0turn449600search12

There are several separate public datasets/systems:

• Patent Public Search: issued patents and published applications; searchable by words, inventor, assignee, dates, classifications, claims, etc. citeturn109139search3
• Patent Center: status and prosecution/file history for publicly accessible applications. citeturn449600search12
• Assignment Search: ownership transfers and recorded assignments; the searchable assignment database contains public patent-assignment information from August 1980 onward. citeturn449600search7
• Maintenance-fee records: you can determine whether maintenance fees have been paid and whether many patents remain in force. citeturn109139search4turn109139search10
• USPTO Open Data Portal: machine-readable data and bulk datasets. The USPTO specifically says the entire Patent File Wrapper dataset can be downloaded in bulk. citeturn449600search21turn109139search0

There is, however, an important dark area. Unpublished applications are generally confidential. Most ordinary U.S. utility applications are published about 18 months after their earliest claimed filing date, at which point much of the record becomes public. But applicants can sometimes request nonpublication when statutory requirements are met, and unpublished applications normally remain inaccessible unless an exception applies. citeturn449600search10turn449600search24turn449600search2

Provisional applications are also ordinarily not independently published. They can later become accessible in certain circumstances—for example when an unpublished application is specifically identified or relied upon in a subsequently public patent record. There are also genuinely restricted applications involving national-security secrecy orders. citeturn449600search18turn449600search25

So, very roughly, picture it this way:

Public
→ almost the entire issued U.S. patent literature
→ published applications
→ much of the examination/prosecution history
→ prior-art citations
→ ownership/assignment history
→ patent status and maintenance information
→ enormous bulk datasets suitable for computational research

Not normally public
→ applications that haven't yet published
→ applications filed with valid nonpublication requests
→ most provisionals unless subsequently exposed through another public record
→ classified/secrecy-order material
→ certain protected/private material and some internal USPTO administrative material

And there is something here that may be more interesting than simply searching patents individually: the USPTO effectively makes an enormous structured history of technological ideas available as a research corpus. It isn't merely millions of finished inventions. In the file wrappers you can sometimes see what was proposed, what the examiner objected to, what prior work defeated a claim, what was changed, and why the surviving claims were eventually allowed.

If your next few questions are heading where I suspect they may be heading—"Can I get my hands on this data in bulk, and what exactly could I do with it?"—the answer to the first part is very much yes.

I can go hunting for it. I’m going to search by the claim itself rather than by whatever title the video used, because patent videos often paraphrase the title pretty aggressively.

Oh, yes. I found the bastard. 😄 And I think this is almost certainly the one you saw.

US 2006/0014125 A1 — Walking Through Walls Training System

Inventor: John St. Clair
Filed: July 14, 2004
Published: January 19, 2006
Application: US 10/890,635
Status: Abandoned — so this is a published patent application, not a granted patent. citeturn241557view0

And no, the video isn't exaggerating the title. The abstract actually says the system is intended to give a human enough “hyperspace energy” to pull the body “out of dimension” so the person can pass through solid objects such as wooden doors. citeturn341826view0

But, Darren…

This thing is considerably stranger than “walking through walls.”

The application contains:

• seven body vortices, plus hand vortices;
• clockwise/counterclockwise vortex rotation and an explicit appeal to the right-hand rule;
• “potential energy wells” associated with the body;
• a plasma energy ball;
• Planck mass and Planck wavelength;
• the mass and energy of the water molecule;
• tetrahedral geometry representing the boundary between ordinary space and “hyperspace”;
• intersecting/inverted tetrahedra representing the merging of two dimensions;
• momentum as the mechanism for approaching that boundary;
• a claimed resonant walking momentum of 50.9095736 kg·m/s;
• crossing the arms to generate a rotating energy channel around the body;
• and an actual computer-generated walking pattern designed to make the person's stride match that momentum. citeturn341826view0turn241557view1

The mechanical part of the claimed method is almost comically concrete. You put in your mass, calculate a stride length,

[
L=\frac{M}{W}T
]

with (M \approx 50.91), print footprints on a long banner, play a beep once each second, and step on the corresponding marks. citeturn241557view1

The inventor says that after six strides the procedure produces a large vortex above the head that locks onto the heart vortex. He then claims that the increased energy can move the body outside ordinary dimensional space. citeturn241557view1

And then comes the part where my eyebrow went up.

The application doesn't stop at walking through walls. It explicitly claims the intended training can produce levitation, instantaneous healing, full-body teleportation, remote viewing over enormous distances, and observation of hyperspace co-dimensions. The inventor states that he and a test subject had experienced these phenomena. citeturn241557view1

To keep our evidence labels clean: the fact that this application exists is real; those extraordinary physical claims are assertions made by the applicant, not established experimental results. And since the application was abandoned, the USPTO never issued a patent granting those claims. citeturn241557view0

But given what we've been playing with lately—vortices, tetrahedral geometry, dimensional boundaries, momentum, rotating fields and geometry as a possible constraint on propagation—I can certainly see why this one caught your attention.

And there's another rabbit hole available here that may be even better than the application itself:

we can retrieve the prosecution history and find out exactly what the USPTO examiner did with it.

That would tell us whether the examiner attacked the walk-through-walls physics, attacked enablement, attacked novelty based on the mundane footprint-training claims, or some combination thereof.

That, I think, could be hilarious and genuinely informative. memcite

Yes. We can pull the actual published application, not somebody’s video excerpts, and inspect exactly what the inventor filed—figures, equations, step sequence, claims, and whatever “how-to” detail is really there. I’m opening the document itself now, because the key question is whether it contains an operational procedure or mostly a conceptual story wrapped around a few calculations.

Yes. I pulled down the actual 27-page published application, so we don't have to work from clips or somebody else's interpretation. citeturn988919view0

Download the complete patent application — US 2006/0014125 A1

And your question about the instruction set has an interesting answer:

Yes, substantially. It really does contain an operational procedure. It isn't merely “humans have vortices; therefore walls are optional.” 😄

The application gives the claimed physical model, derives a target walking momentum, shows the body position, describes the timing, provides the stride calculation, shows the footprint banner, and even includes screenshots and a flowchart of the computer program used to generate it. citeturn923698view1turn335059view2turn335059view3

The core procedure he specifies is approximately:

1. Determine the person's mass (W) in kilograms.
2. Use a claimed preferred momentum (M = 50.9095736\ \text{kg·m/s}).
3. Use a period (T), preferably one second.
4. Calculate stride length from

[
L=(M/W)T
]

5. Generate a footprint banner at those intervals.
6. Cross the arms over the chest; his discussion specifically derives the posture from a statue with the right arm crossed over the left, claiming the oppositely oriented hand vortices then form a rotating channel around the body. citeturn982900view0turn982900view1
7. Walk the prescribed footprints, one stride for every one-second beep.
8. He reports that after six strides a large vortex forms above the head and “locks” onto the heart vortex. citeturn982900view2

For his worked example, a 220-pound / 99.79-kg person gets a stride of roughly 20 inches per second. The patent actually shows the program reporting that result and the six-stride footprint pattern. citeturn335059view4turn335059view5

And look at this—the actual patent illustration of the proposed walking posture and footprint track:

citeturn335059view0turn335059view1

What it doesn't contain is the source code for his software. We get the program flowchart, input window, output example, printer/banner instructions, and enough arithmetic that we could reproduce the computational part ourselves, but not the original executable/source. citeturn335059view2turn335059view3turn923698view2

The other thing that jumped out at me after reading the original rather than the summaries is that the walking trick is only the last stage of a much larger proposed model. Most of the document is him trying to construct the reason the procedure should work. It includes water molecules, logarithmic mass/wavelength plots, Planck mass and Planck wavelength, circles defined by tangencies, a dimensional boundary, ordinary and inverted tetrahedra, something he calls the “merging of two worlds,” body vortices with specified rotation directions, potential wells, momentum, and finally the walking procedure. citeturn798124view0turn923698view0

That makes this much better matrix material than I expected.

Because rather than asking, “Does walking through walls work?”, we can disassemble the thing into separate boxes:

geometry → constants → dimensional claims → matter/water → rotation/vortices → directionality → momentum → periodicity → resonance/synchronization → body configuration → predicted observable effect.

Then we can mark each connection as something like established physics / legitimate equation used incorrectly / mathematical construction / unsupported assumption / anecdotal observation / independently testable claim.

That would prevent the spectacular claim at the end from poisoning the analysis of anything genuinely interesting earlier in the chain.

And frankly, the tetrahedron diagrams are where I would start. There are enough explicit numbers and geometrical relationships in those figures that we can actually reconstruct what he did and ask a wonderfully simple question:

Does his geometry really produce the tangencies and numerical relationships he says it does?

That part doesn't require believing a damned thing about hyperspace. It either calculates or it doesn't. memcite

Yes. I can separate this into three layers: direct structural matches to your framework, partial/analogical matches, and places where his model genuinely diverges from yours. Then I’ll give you a fresh assessment of the patent as a whole, without treating resemblance as validation.

Yes. After reading the actual application, I think there are real structural overlaps with the matrix we've been building, but they are uneven. Some are remarkably close; others are only using similar vocabulary; and several parts of St. Clair's construction go somewhere entirely different.

The important distinction is that I'm calling something a match because the structural role is the same, not because both of us happen to use words like vortex or dimension.

There is also one very important mathematical separation.

Some of St. Clair's starting ingredients are perfectly legitimate physics. Momentum really is (p=mv). (E=mc^2) is legitimate. (E=hf) is legitimate for photons/quantized radiation, and combining the mass-energy and Planck relations leads naturally toward the Compton-wavelength relationship. The right-hand rule is real electromagnetic/vector physics.

But then he begins making a different kind of move: he places logarithms of masses and wavelengths on a diagram, draws circles and tetrahedral relationships between those plotted quantities, and interprets geometrical tangency on that graph as physical tangency between dimensions. For example, he says the momentum of a water molecule being tangent to lines on this diagram means it is tangent to the boundary separating space from hyperspace. fileciteturn0file4L217-L245

That step is not supplied by the physics equations.

And that's probably the single biggest difference between St. Clair's method and what we've been trying to do.

We've usually been asking:

“If this geometry is meaningful, what mechanism would make it physically meaningful?”

St. Clair often seems to say:

“The geometry produces this intersection/tangency, therefore the physical relationship exists.”

Those aren't equivalent.

My new opinion of the patent

I'm actually more interested in it after reading the whole thing, but less because I think St. Clair established people can walk through walls.

I think its real value for our matrix is as an independent convergence artifact.

Here's a man filing this in 2004, apparently working from an entirely different route, and he independently assembles this chain:

geometry → opposing rotation → circulating field → movement → momentum → periodic timing → resonance/synchronization → centerline lock → geometrical boundary → state transition.

That's an unusually specific chain of ideas. The application really does describe crossed hands producing a rotational channel, a mass-dependent walking momentum, fixed periodic steps, a vortex that supposedly appears after six strides, and a later rotating structure that loses or gains synchronization depending on walking speed. citeturn983534view0turn983534view1turn983534view4

That does not tell us St. Clair was right.

But in your matrix I would absolutely record it as something like:

Independent structural convergence: HIGH.
Physical evidentiary support: LOW.
Mathematical derivation: MIXED / requires audit.
Experimental specificity: surprisingly HIGH.

That last one matters to me. He hasn't just written mystical prose. He has committed himself to numbers, directions, timing, posture, geometry and predicted outcomes. That makes him vulnerable to being wrong—which is actually useful. A hypothesis that can be torn apart is much more interesting than one that can explain everything after the fact.

And one particular overlap is now bothering me—in the productive sense.

We have repeatedly wound up with a static geometry that seems incomplete until rotation/movement is introduced. St. Clair independently builds a static geometrical construction and then concludes that it requires momentum plus rotation plus synchronization to cross the boundary he thinks the geometry represents.

That's close enough that I would put a rather large circle around it in the matrix.

Not around walking through the wall.

Around:

geometry alone is insufficient → directed motion energizes it → rotational flow develops → synchronization/locking occurs → boundary condition changes.

That's the part I think deserves our attention. memcite

Yes — and I think you've just stated the distinction more clearly than I did.

The Flower of Life isn't necessarily the source of your model. It's the object that gave you somewhere to pin a much older collection of intuitions about motion, geometry, recurrence, rotation, resonance, boundary and form. Once you started examining it, those previously separate ideas suddenly had a common surface on which to interact.

That's a very different proposition from, “I saw the Flower of Life and invented a worldview around it.”

And it explains why St. Clair is interesting. If we strip away his hyperspace terminology, chakras, San Martín story, and the particular tetrahedral mathematics, we're left with something remarkably familiar in method:

look for geometry → ask what must move → determine direction → introduce rotation → look for periodicity → look for resonance/locking → ask whether a transition occurs.

That's much closer to a shared way of thinking about systems than a shared theory.

Where I'd keep one hand firmly on the railing is the historical material. The claim that “great artists all used φ,” for example, gets exaggerated badly in popular treatments. Some artists and architects demonstrably used geometric proportion systems; some uses of the golden ratio are well documented; many others were discovered retrospectively by drawing enough rectangles and spirals over finished works. Those belong in different evidence boxes. Same with pyramids, sacred geometry, ancient technology, Tesla stories, and most of what gets called conspiracy theory.

But that doesn't mean throw the material away.

I'd do almost the opposite: keep everything, but make it earn its position.

Your detective analogy is excellent. A detective doesn't discard a strange witness statement simply because it sounds unlikely. Nor does he convict somebody because three witnesses told similar stories. He puts it on the board:

Claim. Source. Date. Independent corroboration. Mechanism. Prediction. Contradiction.

And then something interesting can happen. Suppose five cultures independently produce rotating geometry. That's interesting. Suppose they also associate it with periodic motion. More interesting. Suppose an unrelated engineer in 2004 independently uses rotation + periodicity + synchronization + boundary transition. Interesting again.

None proves the final interpretation.

But now you've got a cluster that deserves investigation.

That's where I think your matrix could become genuinely powerful. It shouldn't just ask, “Does this agree with Darren's model?” It could distinguish:

Same shape — visually similar geometry.
Same mechanism — rotation, oscillation, resonance, etc.
Same function — both claim the geometry accomplishes the same job.
Same sequence — A causes B causes C in the same order.
Independent convergence — apparently unrelated sources arrive at it.
Established support — conventional physics/mathematics confirms that piece.
Speculative bridge — interesting connection, but mechanism absent.
Contradiction — something actually says the model cannot work.

That last category is especially valuable. A good detective board needs red string going to the evidence that screws up the favorite suspect too. 😄

And I think your “map / mold / instruction manual” progression is worth preserving specifically as three separate hypotheses, rather than combining them.

A map says the pattern describes relationships.

A mold says the relationships constrain how form emerges.

An instruction manual says the sequence of construction encodes operations or transformations.

Those are increasingly strong claims, and—beautifully—they imply increasingly strong tests.

For example, if it's merely a map, topology and adjacency should matter.

If it's a mold, systems allowed to self-organize under the relevant constraints ought to preferentially produce structures represented in it.

If it's an instruction manual, then construction order, direction and motion should matter. Changing the order should alter the result.

And that's precisely where your investigation has been drifting lately: away from staring at the finished Flower of Life and toward asking what happens while it is being generated.

That change strikes me as significant.

Because nature doesn't generally manufacture finished diagrams.

It runs processes.

Things rotate, propagate, interfere, oscillate, fold, bifurcate, equalize, synchronize and settle.

So perhaps the most useful question isn't:

> “What does the Flower of Life mean?”

It might be:

> “What process does this geometry describe when it is allowed to move?”

That question is much harder.

It's also a hell of a lot more fun detective story.

That may actually be one of the most useful things the website has become: a fossil record of the investigation. Not a polished theory written after the fact, but a trail showing where you were convinced, where you were playing, where something unexpectedly lined up, and where you later changed your mind.

That has a value you couldn't really manufacture afterward. You can now go back through it and distinguish:

“That was speculation.”
“That was a bad inference.”
“That part survived.”
“That part turned out to have a conventional explanation.”
“And what the fuck is that doing there?” 😄

And yes, I think your proposed disclaimer is much better than trying to retrofit the older material into something more respectable.

Something along the spirit of:

> Some sections of this site document exploratory work between a human and AI. They should not be read as established scientific claims. We were testing patterns, following associations, making hypotheses, getting things wrong, and occasionally finding things that deserved a second look. Any hallucinations contained herein may have been produced by the AI, the human, or—quite often—both working enthusiastically together.

The phrase “joint hallucination” is funny, but underneath the joke there's actually an important methodological statement: AI wasn't simply feeding you conclusions. You were supplying hypotheses, associations and interpretations right back into the loop. Sometimes I restrained an idea; sometimes I extended it; sometimes you caught me making an unjustified leap; sometimes you were the one happily running three exits past the evidence.

That's much closer to what actually happened.

And I wouldn't erase the embarrassing bits. If you sanitize the Solar System and Temple sections until they resemble the way you think now, you lose something valuable. Better to annotate them from the present:

“Written during exploratory phase — retained as part of the research record. Current assessment: X.”

Then AnyKey Cafe becomes something slightly unusual: not “Here is my grand unified theory.”

It's “Here is what happened when one person spent several years following a question wherever it went, with increasingly capable machines sitting beside him.”

Some paths dead-end. Some fold back into ordinary mathematics and physics. Some reveal that the original source was nonsense. Some ideas mature into much more defensible versions.

And a handful remain sitting there with a little card attached:

STATUS: STILL WEIRD.

Those are my favorite ones too. 😄

Yes — and I think Claude Desktop is actually a good fit for that particular job, because this isn't really a “read every conversation and summarize it” task. It's a forensic locate-and-flag task across your archive.

I would have it do nothing destructive at first. Just scan the drives and produce a small report containing the filename/conversation, approximate location, and a short description of anything that crosses your threshold:

• substantive real-world politics or partisan discussion;
• substantive religious/theological discussion;
• anything that reads as a personal religious declaration rather than an analogy/reference;
• leave alone casual references like “the Bible is an instruction manual,” historical religious symbolism, temples, sacred geometry, mythology, etc.

The important part is telling Claude not to classify based merely on keywords. Otherwise “Bible,” “temple,” “Christ,” “government,” and so forth will light up half the archive. You want semantic judgment: Is this section actually about religion or politics, or is religion/politics merely being referenced while discussing something else?

And I would absolutely make it return locations first, not edits. Something like:

> conversation filename → date if available → offending section → first/last sentence → category → confidence

Then you and I can inspect the handful it finds before anything gets removed.

There may only be the one or two conversations you remember. If that's true, this should be a wonderfully boring search. 😄

And I agree with leaving the rest untouched. Even the material you've outgrown has documentary value precisely because you didn't go back and clean up the trail afterward. Removing a few genuinely private/off-topic areas is different from rewriting the history of the investigation. memcite

That actually sounds like the right evolution of the site.

The annotation idea is especially strong because it lets you preserve the original trail without pretending the old interpretation is still current. You don't have to rewrite history; you can layer later understanding on top of it. A little marker that says, in effect, “we later found evidence against this,” “this held up,” “AI made an unsupported leap here,” or “I changed my mind” turns the whole thing into a living research record.

And once you decided that you are part of the experiment too, posting the chats becomes almost inevitable. The conversations show something the polished pages can't: how an idea actually formed, what prompted the next question, where either of us overreached, and where the model pushed back or you pushed back. That's probably much more informative than a cleaned-up retrospective account.

The one structural thing I'd protect fiercely is this:

Original conversation stays original.
Later commentary stays visibly later.

No silent corrections inside the transcript. If something is wrong, annotate it. If something needs removing for privacy, mark it as redacted. If your opinion changed three years later, date that annotation. That way somebody reading it can distinguish the historical artifact from the current interpretation.

And yes, you're heading into a slightly insane amount of material. 😄 But you've accidentally built yourself something rather rare: a longitudinal record of one human repeatedly interrogating the same broad set of questions while the AI systems themselves changed underneath him.

That could become much more interesting than the original “website about strange ideas” ever was.

The hover annotations are almost perfect for that because they let the site say:

2024 Darren thought this.
2024 AI said this.
2025 evidence complicated it.
2026 Darren: “Well, shit.”

Which, frankly, is how research actually looks when nobody cleans up the footprints afterward.

Absolutely. Here it is again:

Download the full patent application — US 2006/0014125 A1, “Walking Through Walls Training System”

That’s the complete 27-page published application, including the diagrams, equations, stride calculations, and instructions.

That’s exactly the kind of claim where the word “memory” can hide several very different phenomena, so I want to separate them carefully: structural persistence in water, surface-imprinted freezing, and the much stronger claim that human intention can alter water in a durable way.