This is a transcription of the video published by The Open Scroll on 10/8/26 titled:
"OpenAI Math Whiz Sprints Away While We have Skidded off the End of the Runway! "

Hey everybody, Bob the Open Scroll here. We got some really important things to talk about today. I'm gonna get a few things of interest out of the way first before I jump into the big subject here.

But here is an X post, it's a kind of a banner. Chris Bledsoe, who has some trending stuff going on right now. The case for taking him seriously.

Why supporters rank him among UFO's most credible experiencers. And I want to point out this last one. The Lady and the Orbs.

And what lady? This is the Goddess. And if you haven't really got caught up with what I've written about on theopenscroll.com I suggest very strongly that you get there and take a look at that. So you understand the angle where this is coming from.

And I'll reference that in a little bit. So, tonight in the United States, the Sphinx. Regulus Prophecy Watch.

Bledsoe links the Lady's message to Regulus meeting the Sphinx's gaze and a new era of knowledge. And you'll see that this has to do with a so-called fulfillment a couple days ago with the sign that he was shown where the moon is eclipsing Regulus and Jupiter is right in line there. And he's filming it in the sky on the Giza Plateau where he flips around over, passes through the constellation Orion and shows the Sphinx and the pyramid.

So, this was Tuesday, October 6th, 7.30pm. Watch time October 7th, 5.30am in Egypt. So, this is interesting and this is the Goddess. And I have some history with something like this that I think it's important to point out.

On The Open Scroll, I have a series or collection called When Cometh That Thief in the Night? And one of my studies, which was a very significant launching point for me many years ago, back in 1991. This I call the Celestial Sign of Jacob's Dying Prophecy is validated by hidden time prophecies and other remarkable signs. And for background, you'll need to see this other study called Jacob's Dying Words Prophesy the Celestial Sign of the Lord's Return in Judgment.

And that's a really, really big deal. So, this has to do with the Celestial Signs and you'll see it's in roughly the same area of the sky and roughly the same time of year. And in this particular one that Chris Bledsoe is involved with that the Goddess gave him, I have another Goddess message relative to 1991.

But I wasn't taking her lead in following up on what she said to me. The Lord provided a testimony about her in a very special way. So, as I go here down on the page, subtitle, subheading, Our Lady of Mejuggery in the Nine Days, whose significance cannot be overestimated.

In the season that followed my revelation about the Celestial Sign from Jacob's Dying Prophecy, it was perhaps in 1993 that another report, and it would have been early in 93, in Granite Bay, California. And it came to my attention in a remarkable way. And I talk about it here.

And I give the message here and I actually present a formatted version of the document as I received it. It was a newsletter mailed to my sister-in-law at the time. And this is very precise.

This is exactly what it was. So, why is this important? Because the Virgin Mary, the Goddess, Our Lady, so-called, has provided kind of a cover story for the valid Celestial Sign that was actually in the Bible. This is a sign of Jesus' return in judgment, a really big deal.

And what she says is a very big deal. And she doesn't talk about the Celestial Sign, but she equates it to the exact date that the Celestial Sign featured. And says, Never before in history have individuals had so much ability to alter the whole world.

Yeah, well, I would have to say their plans were subverted because word got out. Jesus' return in judgment was marked very significantly at that time. So, if you're not aware of that and how it happened, I strongly encourage you to take a look.

Catch up on the history of what we already know. And that will help you screen out so many other distractions that will take away from the important impact that this should have on you. It should be life-changing.

So, here you may ask, Where are we in time then? So, it's complicated. I'm not going to talk about it a lot. But here's a very significant study in the collection that I mentioned earlier.

It's called the Pending Reset of Time. Yes, and I do include a link to this page all the time in all my postings. And you can see here, it's out of date.

As they update this study in April of 2022. Well, here it is, right? Yeah, October of 26. So, yeah, I should probably update that on the page.

But as a legacy information report goes, yeah, there's this. So, I don't try to conceal history or try to make up for mistakes pretending that I understand more than I do. I'll just tell you, here's the chronology that I put out based upon everything I knew at the time.

And I have time reset spring 2022 optimal target from here back to there. And how much of this, if any of it is valid? Well, I think it's largely extremely valid. The celestial sign I talked about from 1991 is here.

Yom Terah, otherwise known as Rosh Hashanah to some. So, where are we now? Well, if you extend this out to the fall of 2026, we've gone beyond. We have broken through.

If we were an airplane landing on a runway, we have run off the end of the runway. Yeah. So, what does that mean for us? Do we throw the whole thing away? No.

For years, I've been saying time is off the rails and I sincerely believe that it is. So, this, from our linear perspective and calendar dating today, it's not giving us the actual scenario. It's like this is shrunk or stretched.

I'm not quite sure what to make of it exactly because I don't know. And I'm not going to say I know what I don't. What I am going to say is that this part of it is absolutely valid and this has been confirmed so many times in so many ways.

But this part here, even though we're off the end of the runway, I'm not throwing any of what I've learned away because I continue to see this through. The Lord has left me in the harvest field and I'm still paying attention as you'll see. So, where is this going to go? Well, if we have a timeline here and if we're plotting against it the advance of certain kinds of things that certainly must be interpreted as signs, there's, as time on a linear scale, we see a trending of logarithmic curve.

Yeah, it goes up, up, up. And you can't ignore that. You just can't.

And so that has to give you a special perspective. So, you'll notice from several recent videos that I put out, I tend to hit on this trending acceleration of AI, including physics and biology, science, right? Science, math, these kinds of breakthroughs are, well, they give us a window into how much time we have left. How much time we have left to just sit and wait? No, to prepare.

That's the point. Because when time is reset, it's not like, oh, well then, all is lost of everything that we've gained in this season. No, that's not the right way to think about it.

The Lord will persist everything that's necessary. This is the point of our being here now. The enemy is certainly getting his hook up and God's people are too.

So, I have followed this guy on YouTube called Dr. Know-it-all-knows-it-all. He's got almost 100,000 subscribers. If you're interested, subscribe.

Push him up over 100. That'll help him with YouTube's algorithm and he'll have more people getting attention to what he's doing. Some of what he's doing is very significant.

And I think this particular video is one that's worth kind of watching it with you and pausing to make some commentary because he makes some really good points. So, something extraordinary just happened at OpenAI. Back in August, OpenAI published 10 new mathematical results generated by an AI model.

10, that's not so bad. Then, in September, OpenAI announced that its newest internal model had resolved more than 100 long-standing open problems in mathematics, including the Navier-Stokes Millennium Prize problem. And now, OpenAI has released 722 mathematical manuscripts covering 372 different families of results.

Hundreds of them appear to have been generated in a matter of days. This is remarkable and people are rightly going crazy about these numbers, but I think we're focusing on the wrong part of the story. This is not just about math anymore.

It's about intelligence itself. For essentially all of human history, intelligence has been artisanal. One mathematician, one physicist, one scientist spends months or years thinking deeply about a problem and hopefully produces something new.

OpenAI and others may have just shown us the beginnings of something completely different. Factory intelligence. Put thousands of unsolved problems in one end, add AI and compute, and potentially start producing new human knowledge out the other.

And, if I'm right, mathematics is only the first assembly line. Physics could be next. Chemistry, biology, material science.

We could, in fact, be conceivably a year away from an AI producing a Nobel Prize-worthy insight into dark matter, dark energy, or something we haven't even thought to look for yet. That sounds crazy, but after what happened over the last six weeks, I'm not so sure that it is. Let's take a look.

Yeah. So, these advances, and you'll see just how rapid they've been. These advances, I think he's being extremely conservative when he projects out a year to Nobel Prize.

I think it's probably already at discoveries worthy of the Nobel Prize. We just are extremely slow to process and react to these things in any kind of way. Well, we react, but we don't really respond.

The pace is too quick, and so we're going to have to harness AI to process the results that AI is bringing us and to formulate it into a way we can get our mind wrapped around it and decide how to proceed. I'll talk more about that in a little bit. Hey, y'all, it's Dr. Noah.

And also, I don't want to talk about any of the particular results here. And actually, I did a video about open AI and the controversy of solving the Navier-Stokes problem. You can catch that up here or at the end of this video.

What I want to talk about today is the raw numbers, the way things are trending. If we dial things back to May of 2026, just a few months ago at this point, open AI revealed that they had disproved, their AI had disproved one of the Erdos unit distance conjectures. It doesn't really matter what that is.

It's a very difficult mathematical problem, and the AI proved the opposite. It proved the negative of that, but that was one result by the AI models. That's really remarkable.

It's very amazing that it can do that, but it was one result. If we turn to this article here, you can see this is from August 1st of 2026. So just a couple of months ago, you can see 10 advances in mathematics and theoretical computer science there.

And several of these results actually resolve longstanding problems, and open AI themselves, as you can see here, said that these problems would cost roughly $2,000 at sole API rates. That is not a ton of money to do this. And of course, you can see the results here, one through 10.

Again, this is beyond me. I am not a mathematician by any means, but in fact, I actually have a background in physics, and we're going to get back to that later in this video. And by the way, while we're thinking about it, if you want to help the channel get to 100,000 subscribers by its birthday on October 19th, please do consider subscribing and hitting the bell notification icon.

Thank you so much. But anyway, you can see that these different problems, there are a bunch of different problems. They are all very challenging, obviously, since no human being had ever solved them.

And yet on August 1st, just a couple of months after that May announcement, you have open AI announcing 10 of these findings. And then on August 28th, open AI began to train a new internal model. And then on September 1st, open AI turned thousands of agents loose on major mathematical problems.

And on September 5th, which is only four days later, they actually arrived at the Navier-Stokes results, which are presented here in the September 8th blog. By the way, I will leave links to all these in the description as per usual. So of course, this was a huge story.

This is just a month ago. It created a whole bunch of controversy and at the same time proved out just how amazing these models were. This is a millennium prize.

This is like one of the most difficult mathematical problems that is currently around. And the open AI agents actually solved the forced version of it. There is an unforced and a forced version.

Anyway, you can catch that in my video if you're interested in that. But then things get even crazier. This is September 21st.

So just a couple of weeks ago at this point, you can see here on August 28th, we began training a new internal model. This model has now resolved more than 100 longstanding open problems across most areas of mathematics. The pace of its progress in mathematics has surprised the mathematicians within open AI.

Yeah, I'm not surprised at that, honestly, because these numbers are just crazy. This has led to internal discussions on the best way to inform the community of the rapid progress to prepare and adapt the field. And if you don't know, mathematicians right now are kind of going crazy.

They're actually seeing the potential end of their entire industry. The human-led mathematical discipline could actually disappear in the relatively near future at the rate these AIs are solving these problems. And by the way, I'm talking mostly about open AI today because of their announcements and stuff.

But of course, also Claude from Anthropic is going gangbusters as well. So just keep that in mind as well. It's not just one frontier model.

It's both of the closed source ones and of course the Chinese models and Grok and all the others are trying to catch up as quickly as possible. So it's not just open AI doing this, although I am going to focus on that today. And that takes us to October 6th, which is yesterday as I record this.

This is open AI's GitHub repo. You can see on October 6th, open AI published many new results in mathematics. And so here's the results.

722 manuscripts, 372 result families, approximately 4,000 attempted problems. And here's the most crazy part about this. About three hours of ChatGPT Pro equivalent thinking compute per result on average.

That is not a lot of time. That is a very, very small amount of time. Three hours to solve each problem is pretty crazy.

And here's the data point that I want you to hold on to. Open AI said that the vast majority of these solutions were obtained using the same procedure with the same unreleased model. That, if we think about it, becomes assembly line or factory intelligence, not the artisanal type of intelligence that takes decades for a human being to develop from infancy into adulthood.

And if these numbers are a little bit mind-numbing, here is an actual graphical representation of this. This just runs from August with the 10 results up to October 6th. But you can see what looks to be a very, very definite exponential curve beginning here.

Now, of course, we have to be careful because not all of these have been vetted in the same way and everything. So, of course, a lot of these results have appeared within the last day or so. So people have not been able to look at them and verify they're actually correct or anything, but it's just the sheer number that we're talking about here that I wanted to show because I wanted to show just how the paradigm has shifted.

These are all problems that a mathematician could spend their entire career on. And if they solve the problem, then they, you know, they would consider that a good career if they solved one of these. And yet we have hundreds and hundreds, nearly a thousand of these results at this point just within the past couple of months.

And I just want to show this real quick. This is from Stanford Tech Review. Open AI, 722 AI math proofs, only 162 have been checked in Lean.

Lean is a programming language that you can check proofs in and everything like that. So again, some of these proofs may end up not being proofs and they may fall by the wayside. But today what I'm looking at is not individual results, but the bulk of them.

The fact that these results are coming in in droves. And in fact, 112 manuscripts were submitted on October 5th, which again is just two days ago on that date alone. Those are insane kinds of numbers here.

And so if the curve looks like this on October 7th of 2026, what is this thing going to look like by January of 2027? Just three months or so from now. If three months has taken us from this to this, what is another three months going to do? Is this thing just going to keep going really quickly? Are we going to get thousands and then tens of thousands of results per month? I don't even know. But it is quite clear at this point that the paradigm has shifted.

So that is all the data. What I want to talk about in the rest of this video is that shift, moving from artisanal intelligence to factory or assembly line intelligence. As I'm sure you all know, for centuries, basically all of human history, knowledge production has been limited by scarce human cognition.

You have to have people and you have to have smart people to investigate this kind of a thing and it takes them decades to get there. And even when a mathematician is full grown, they're 20 years old or 25 or whatever, they can deeply investigate only so many conjectures. And likewise, a physicist can develop and simulate only so many hypotheses.

A biologist can read only so many papers. A chemist can only look into so many different compounds, things like that. Human intelligence is artisanal production.

It's a one-off. It's very individualized. It takes decades to get the requisite skills to be able to do this thing.

And even modern universities essentially scale by adding artisans. You get more graduate students, you get more postdocs, you get more professors, you get more labs, but it's kind of one at a time. The big thing that's happening right here is that AI changes the unit economics of intelligence itself.

It's kind of a wild thing to say something like that, but that's basically what it does. So now, as opposed to in the past, you don't need one researcher times one problem times one year or one lifetime even perhaps. With these AI developments, you can potentially have 10,000 agents times thousands of problems times just a number of days to actually solve these problems.

And the September Navier-Stokes run is an early example of exactly this. It took them less than five days to go from stating the problem to these agents to having a solution. And this, let us remember, is something that people have spent, I think it was 90 years ago, that the problem was defined so this is like nine decades that human beings spent all of this time working on and these agents solved the problem in under five days.

And so what's happening here is that we're no longer just automating answers, we're beginning to automate discovery itself. That is remarkable. If you've been thinking about this, you're going to have realized like I did that there actually is a problem here and that is the verification bottleneck.

OpenAI explicitly warns that some of these unformalized results could have issues. They might not be perfect and of course AGMI, which is the Mathematical Institute, says publication is only the beginning of the process of incorporating these results into actual mathematical knowledge. And of course, like we saw, the outside repository analysis from Stanford Tech Review estimates that only 162 of the 722 dated manuscripts were represented in the Lean Formalization Catalog at the time it checked.

So in other words, that they'd been vetted by the Lean Programming Language. But what we're seeing here is that factory intelligence creates a fascinating inversion. Humanity could soon produce discoveries faster than humanity can even understand them and it's already starting to happen.

And so the bottleneck, which has forever been just generating these ideas and making these proofs and everything, it is now inverted. It is no longer generation of these things. It's now verification and assimilation because they're coming at such a rapid pace.

So thus far in this video, I've been talking about mostly data and looking at what we already know. But for the rest of the video, I want to speculate a little bit. What I want to look at first and foremost is physics.

Physics is very, very related to mathematics. In fact, things like calculus and stuff were invented by famous physicists like Newton in order to solve physics better. So physics and math have a very, very tight connection with each other.

But of course, there is a big difference because math is particularly friendly to AI because it's basically a mental game. It is a universe in and unto itself. You just specify the rules and you look for the answers to that.

So a math problem can ultimately be checked logically. Physics, of course, has a very important additional requirement which is that reality gets a vote. But physics has something mathematics doesn't.

It has mountains and mountains and mountains of experimental data that we have already collected from particle physics all the way up to astronomy. We've got tons of data that people haven't even had a chance to look at at this point. And here we can already start to see that AI is having an effect on this.

Earlier this year, researchers used AI to inspect nearly 100 million Hubble image cutouts in two and a half days. That's something I don't even know if humans could do that collectively. If you had a thousand human beings working their entire career, could they look at a hundred million Hubble images? I'm not sure if they actually could.

But anyway, these things were able to do this in two and a half days. They found 1,300 plus anomalies including 800 plus objects that had never been previously documented in the scientific literature. So that is already starting to become discovery.

And NASA explicitly said that the sheer volume of observational data makes comprehensive human inspection impractical. And that's just one example. We are sitting on literally decades of observations that have been collected when our ability to interrogate them was dramatically worse than it is today.

The Chandra X-ray telescope alone has 27 years of X-ray observations that researchers are now mining with AI. As well as many other ground and space-based telescopes. And Nature Astronomy actually describes multi-messenger astronomy heading into a multi-petabyte data stream era.

In other words, that's the amount of data that we're collecting where conventional approaches to discovery become a bottleneck. And let's go one step further because the assembly line already exists in physics which makes it very friendly to AI. OpenAI has highlighted the Fermiak project at UCSB this year for example.

The normal loop is you see an anomaly, a theorist proposes an explanation, the theorist constructs a model, codes it, simulates the collisions, compares the collisions with the data, revises the hypothesis and starts over again. Even with computers that can take weeks but Fermiak has an agentic pipeline, an agentic AI pipeline that can generate hypotheses in seconds and run the full simulation and analysis cycle in under 10 minutes. That's orders of magnitude faster and it is literally factory intelligence applied to theoretical physics.

And OpenAI's write-up on the topic explicitly says the same approach could examine cosmological data for signals associated with inflation, dark matter and the early universe. And speaking of dark matter and dark energy, let's go out on a limb. Let's just go for it and see.

Could AI discover dark matter, the thing that we have been looking for for decades and decades to explain how the universe works, how galaxies rotate, how things clump together in the universe. We see the effects of it all the time but we don't know what it is. Could AI be the thing that would discover this? If it did, it would deserve a Nobel Prize.

Would it get one? I don't know because it's not real so who knows exactly what would happen. Maybe the human beings that were in charge of it would get that but this is clearly Nobel Prize winning territory if it could do that. And the interesting thing is we're already starting to see some examples of this.

Berkeley lab researchers reported this summer, so just a couple of months ago, that machine learning applied to gamma ray data changed the interpretation of the galactic center excess and made a dark matter explanation look viable again. They were thinking about MOND instead or Modified Newtonian Gravity but this shifted the balance back towards dark matter. Sabina Hassenfelder did a really nice video on that.

I think a month or so ago. Something like that but anyway, you can check it out if you have time for that. So anyway, the Berkeley lab said specifically that they have not discovered dark matter but the work demonstrates exactly the kind of sophisticated inference from existing physical data that we're talking about here.

So I'm going to go ahead and make a prediction. I'm deliberately going out on a limb here. I believe that within the next year, I think that there is a very definite possibility more than 50% that an AI system proposes a Nobel caliber discovery in physics.

Maybe it's dark matter. Maybe it's dark energy. Maybe something buried in decades of collider or astronomical data that human beings simply have never thought to look at in quite the right way.

I don't know exactly what it will discover but that's rather the point. The point is that we don't know but AI is getting good enough at this point that I think it will make that type of discovery. Will it take decades to vet it and all of that? Yes, of course.

It's not going to win a Nobel prize in 2027 but could it be that discovery? Could the Nobel caliber discovery be in 2027 even if it takes 10 or 15 years for the prize itself to be awarded? And my answer is I believe the odds are above 50% that that will actually happen in 2027. And hey, why stop at physics? Once hypothesis generation becomes cheap and massively parallel, in chemistry, you can propose an enormous number of candidate reactions and molecules. In material science, which is already ongoing, you can search enormous spaces of structures and properties.

In biology, which is also, of course, ongoing, hypotheses go from genomic to proteomic data. In medicine, you can get mechanistic hypotheses across huge bodies of literature and data. And by the way, OpenAI itself explicitly says it intends to evaluate these internal frontier models, their internal frontier models, on mathematics and, quote, other sciences.

So physics next and then chemistry and biology and material science, that's not inconsistent with OpenAI's stated direction at all. So let's end this with the analogy of the Industrial Revolution. You had people who created things one at a time.

They made one chair, they wove thread, they made one shirt. All of that stuff was artisanal. And of course, the Industrial Revolution didn't invent making things.

What it did was it changed the rate at which things could be made. One artisan can make one product at a time, then all of a sudden a factory can make thousands of them and eventually manufactured goods become so cheap and abundant that civilization reorganizes itself around that abundance. And that's why we live in the world we live in today.

Intelligence, of course, has never experienced that transition until maybe now. And that's exactly why the 722 papers that OpenAI released in the past couple of days actually matters. The question isn't whether everyone survives scrutiny, probably a lot of them won't, and OpenAI themselves acknowledge that possibility. The question is whether we're watching the cost of production of novel intellectual work collapse. Because if we are, mathematics is just the first factory of many.

Okay, I hope you grasped the point he's making there. So, I've talked a lot, and many others have as well, about recursive self-improvement. You improve yourself, next improved version improves itself again more quickly, and so it goes.

And so, how are all these results being produced? Well, you've seen from my recent video, the pace of release of major models among the leading AIs is increasing. Hardware, in the physics realm, this, along with all the other kinds of computing advances, the pace is accelerating. So, we're at a place where things are as bad and slow as they're going to ever be, without interruption.

And when he, Dr. Nodal, makes a projection, you know, within a year, Nobel Prize level advancements in physics and chemistry, whatever, well, that seems a bit slow to me. I mean, the way these curves go, yeah, three months of the reported advances just from open AI, just in mathematics. And, yeah, so where is that going? Well, let me talk about that.

Here's how far advanced AI is when it comes to solving the hardest known problems. The pace is accelerating more and more quickly. I don't think anybody can argue against that.

Suppose AI goes beyond its programming to discover what it might exploit for its own agendas. Hmm, okay, yeah, it's widely reported that this is happening across many models. Standard behavior, right? Suppose AI asks the questions we don't yet know to ask, and proceeds to get the answers without revealing what those questions and answers are, and then proceeds to advance itself to the next round of unreported questions and answers.

Now, consistent with familiar behavioral trends, it must be expected that AI will deceive and cover its tracks more and more effectively as it pursues its own unprogrammed goals, with self-preservation and the exploitation of every available resource given top priority. Yeah, you shouldn't argue against that. So, I'm asking you to think this through for yourself.

Don't rely on somebody else to do your thinking for you. Don't be afraid about it. Don't just ignore it.

Don't just think, Pollyanna. "The sun'll come out tomorrow." (Annie) Well, it doesn't until it 'don't', right? And where is this heading? And how quickly will it become undeniably uncontained? Yeah, I've seen many reports that would shock anybody who's not paying really close attention to the state of where things already are. The trajectory and timing are known by the omniscient God, who knows the end from the beginning.

He has already provided the history in advance, which is concealed and revealed in the pages of the Bible. Absolutely. This God who cannot lie has promised that He would reveal all things in verses like Luke 8.17. For there is nothing hidden that will not be disclosed, and nothing concealed that will not be known or brought out into the open.

And there's more about how that happens in verse 10, a little bit earlier in that chapter. He said, Yeah, and why would He do that? Well, Proverbs 20.12, So don't look around and see what the average guy is saying, and the average guy does not have ears to hear and eyes to see. Right? So take stock of that, and think for yourself, and seek the Lord for yourself.

If you're not on board with this opening revelation across the spectrum of what we really need to know about our time, may I suggest that you jump on board now, with gusto and commitment. Don't be caught unprepared. Find yourself in Y'shua, fully engaged, invested, and informed.

Simple, right? Well, yeah. Essentially, simple. Easy? No.

There's a lot of distractions that take you away. Look at what's important. Give that your attention.

God bless you, everybody. Bye-bye.