We’ve Released Our Draft Blueprint
Announcement of GIFT framework v3.4, in plain language. About a ten-minute read.
There’s a magnet on your fridge. Somewhere in your kitchen, a leek sits in the vegetable drawer. The sun warms your window. If there’s a phone on nearby, its GPS knows where you are to within a few meters.
None of these things look like they have anything to do with each other. In fact, they all work because of about twenty very precise numbers. Shift any one of them by a few percent and everything I just described stops existing. The magnet falls off. The sun stops warming you the same way. The very chemistry that lets you read this sentence is gone.
We’ve measured all these numbers, to staggering precision. We have never known where they come from.
That’s what this page is about.
Roughly Twenty Empty Boxes
When a physicist writes the operating manual for reality, the final equation contains something like nineteen empty boxes. Genuinely empty. No reasoning gives their values. To fill them, you go outside, you measure, you come back, you write the result in the matching box. We’ve been doing this for decades, with vertiginous success.
A few examples to ground things.
The first box, called alpha, equals 1/137.036. This number sets how strongly light interacts with matter. It’s the reason a magnet sticks to your fridge, the reason colors exist, the reason your skin warms in the sun. Shift it up by 4% and stars stop manufacturing carbon: no more carbon, no more leeks, no more you. Shift it down by 4% and atomic nuclei collapse. No one has ever explained why this value and not another.
Another box is called Ngen and equals exactly 3. It’s the number of “families” of matter in the universe. Every elementary particle of matter comes in three copies: a light version, a heavier one, a heavier one still. Not two, not five. Three. Always. No one knows why.
Other boxes hold mass ratios between particles, mixing angles between families, the amount of dark matter compared to ordinary matter, and so on. About twenty in total. All measured, all unexplained.
The Blueprint Hypothesis
About a year ago I had a naive, maybe absurd question: what if these numbers aren’t arbitrary?
What if, the way a musical form determines which notes an instrument can play, a precise geometric shape, at a deeper level than the one we usually look at, fixed the allowed values in those nineteen boxes?
The idea isn’t that the numbers would be chosen. The idea is that they would be constrained. As in an instrument, where the length of a string and the shape of the body aren’t free decisions: they’re what decide the notes.
This hypothesis has a name in the mathematical tradition: a structure. A blueprint. An assembly logic hidden behind the measurements.
GIFT (Geometric Information Field Theory) is my attempt at reconstructing that blueprint.
What We Did
We picked a very specific geometric shape, one that lives in seven internal dimensions. Don’t panic: there’s nothing to picture. These dimensions aren’t directions you could walk in: they’re abstract degrees of freedom, the way “color” and “size” are two independent descriptive dimensions for a leek.
This shape has a rare property: it leaves very little room for choice. Once you decide it exists, almost all of its features are determined. It even has a name in the mathematical baptismal records (G₂), but the name isn’t important for this page.
Then we methodically followed the logical consequences of this shape. What mass ratios should we observe if nature is built this way? What angles between families? What value of alpha?
Ninety-Five Predictions, Zero Dials
The result fits in a sentence: starting from this shape, with no adjustable parameters, we can produce 95 numerical predictions about the world. Not nineteen. Ninety-five.
Not all of them land perfectly. Some are barely within one percent, some are off by a wider margin, and several pieces of the mathematical reasoning are still under construction. This isn’t a “final solution” we’re announcing, it’s a working structure, part of which is starting to match the world.
Three concrete examples to give you a feel.
The Koide ratio. A mathematical ratio between the masses of the three charged leptons (electron, muon, tau). We predict it equals exactly 2/3. The measured value is 0.666661. Gap: 0.001%.
The tau-to-electron ratio. The tau is the heaviest version of the electron’s family. We predict its mass is exactly 3477 times that of the electron. Measurement gives 3477.15. Gap: 0.004%.
The dark-matter to ordinary-matter ratio. How much dark matter is there in the universe, compared to the matter we know? We predict 43/8, which is 5.375. Cosmological observation gives 5.375. Gap: invisible.
Across all these predictions, 11 are accurate to better than one part in a hundred thousand. 53 are accurate to better than 1%. Average precision is around the percent level.
Important: none of these results were tuned. The shape was chosen, the logic was unrolled, and the numbers came out. That’s what makes the thing interesting. A bad theory can always be recalibrated until the numbers fall into place. A theory with no dials either lands right or lands wrong.
Checked by a Tireless Bookkeeper
While all of this was happening, we also asked a computer program called Lean to read the reasoning back to itself. Lean does the work of a bookkeeper: it verifies, line by line, that nothing has been swept under the rug. It trusts nothing, skips no step, asks stupid questions at every comma. It is insufferable, and that’s exactly what we want from it.
Lean re-read 213 key statements in the reasoning and found no internal inconsistency. That’s about as close to a certificate of no-mistakes as mathematics knows how to produce today.
The Block We’re Pulling from the Jenga Tower
A good theory has to be able to be wrong, so we’ve announced in advance which block we’re pulling from the tower.
There’s a number, inside the neutral matter we call neutrinos, that measures the asymmetry between matter and antimatter in their behavior. It’s called delta-CP. We predict it equals exactly 197°. This is a very precise number, already partially measured by current experiments, and which will be measured much more precisely by an experiment called DUNE between 2028 and 2040.
Here are the rules of the game, stated publicly: if DUNE measures delta-CP between 187° and 207°, GIFT may have told the story correctly. If DUNE measures delta-CP outside that interval, we’ll have been wrong. Not vaguely wrong: documentably, datedly, archived-in-advance wrong.
That’s exactly the posture of a Jenga player announcing to the table which block they’re about to pull, before pulling it. Either the tower stands or it falls. And the win isn’t being right at all costs: it’s having known, in either case.
What’s Still Open
Part of the mathematical reasoning isn’t yet complete: the precise geometric construction of the seven-dimensional shape is still an open problem that geometers are working on right now. What we have is a blueprint that works on the outside (the numbers land, the computer checks pass), with a mathematical promise still to be fully fastened to the rest. The specialists know this.
Why I’m Making This Public
Everything I’ve just described is now freely available, with no paywall and no sign-up, on several academic platforms like Zenodo or ResearchGate. The source code for the computer verifications is on GitHub. Version 3.4 of the framework, released today, is the reference state from which we move forward.
I’m publishing this now for two reasons.
First, because I could be wrong. Any theory can be wrong. The only way to find out is for other people, more competent than me on this or that point, to come look under the hood. The sooner it’s read, the sooner we know.
Second, because scientific theories are never the work of a single person. They’re collective blueprints. We correct them together. We inherit them, pass them on, tinker with them. This page is one step among many in a very old conversation.
If you want to go further:
the complete blueprint, in mathematical detail, is on Zenodo (link on the blog’s home page),
the computer verifications are at github.com/gift-framework/core,
earlier episodes of this blog tell the pieces out of order, in small touches.
And as at the end of the Lego episode: maybe we’ve all been doing the same thing for centuries. Sorting pieces, comparing patterns, testing assemblies, correcting our mistakes, starting again.