
Eron
@0xEronn • 11,257 subscribers
AI researcher & builder | Content creator |
Videos

A month ago, Springer Nature published a 400-formula book that may have quietly solved quantum gravity - no strings, no extra dimensions, no new particles. Space-time itself undergoes a phase transition at extreme energies: from continuous fabric to scale-fluctuating noise. The same phase transition math that defines when every LLM training run on GPU clusters crosses from chaotic gradients to convergent loss landscapes. The singularity disappears - replaced by a "dark particle," a clot of pure fluctuating space-time ~100x larger than the Planck scale. Zero spin means they bypass the Pauli principle and merge freely - a dark matter candidate that doesn't evaporate. The entire theory rests on one constraint: total energy of the universe equals zero. This kills the 10¹²⁰ vacuum catastrophe, explains dark energy, and eliminates the need for the anthropic principle. Occam's razor destroys string theory in comparison. 50 minutes. Bookmark & watch today. The zero-new-particles theory that solves singularities, dark matter, and dark energy at once.
Eron207,893 Aufrufe • vor 17 Tagen

You can turn a sphere inside out without tearing, poking holes, or creating a single crease - and this is not a trick, it is a proven mathematical theorem. You cannot do the same thing to a circle. The reason is a topological invariant called the turning number, and it is the same invariant governing the geometry every LLM embedding space is built on. Stephen Smale proved eversion was possible in 1957. It took seven more years before anyone found a way to actually do it. The mechanism is corrugation. Divide the sphere into guide strips connected by wavy sections that let the surface pass through itself without pinching. Push the caps past each other before a crease forms. Twist them in opposite directions - the same way a belt with a loop untwists when you rotate each end half a turn. Then push the equator through and uncorrugate. Bowls, domes and saddles cancel in a pattern that stays invariant throughout - the same structure that makes high-dimensional transformer geometry behave nothing like 3D intuition. 21 minutes. Bookmark & watch today. The mathematical proof that breaks every intuition you have about surfaces.
Eron287,629 Aufrufe • vor 1 Monat

Google's first rule for machine learning: don't use machine learning. A 3-line PID equation beats a neural network at balancing a robot. Proportional error, integral of accumulated error, derivative of error change rate - three terms, hand-tuned gains, no training data. The same control theory running inside every GPU cooling system and every server rack thermal regulator at every LLM data center on the planet. The difference shows up when physics change. Add a load, tilt the surface, face an unknown environment - the PID's hand-tuned K values break because they're locked to one set of physics. A neural network with ReLU activation functions captures nonlinear dynamics: a 10-degree lean with backward acceleration doesn't need the same response as a 10-degree lean standing still. Without activation functions a neural network is just y = mx + b. One bend in the function is what separates a linear controller from a system that adapts to anything. 8 minutes. Bookmark & watch today. The 3-line equation that beats AI until the world stops being predictable.
Eron53,323 Aufrufe • vor 18 Tagen

In 1957, Stephen Smale proved you can turn a sphere inside out without cutting, tearing, or creasing it. His own advisor thought he'd made a mistake. The surface passes through itself but keeps a smooth tangent plane at every point - the same topological invariance that governs how every LLM embedding manifold preserves continuity across high-dimensional transformations on GPU clusters. A circle can't do it - flipping reverses the turning number from +1 to -1, and no smooth deformation changes that. A sphere has an extra dimension. Thurston splits it into 8 corrugated strips, pushes caps through each other, twists each strip through a figure eight. The ripples let bands slide past each other instead of creasing. The same dimensional escape that lets attention heads resolve contradictions in latent space that look impossible in lower projections. 3 minutes. Bookmark & watch today. The proof that spheres break a rule circles never can.
Eron19,989 Aufrufe • vor 9 Tagen

Heinz Hopf discovered this in 1931. Roger Penrose called it "an element of the architecture of our world." It shows up in at least 8 areas of physics, yet most mathematicians spend entire careers never touching it. The structure that defines every qubit in quantum computing, every rotation a robotics controller computes, and the geometry inside every LLM embedding space operating on GPU clusters today - and nobody talks about it. The Hopf fibration maps a 4D hypersphere onto a regular sphere using circles that never intersect but each links through every other exactly once. The Bloch sphere every quantum computer operates on, the fiber bundles underlying gauge theory, and the quaternion math inside every 3D graphics engine and every embedding rotation on the planet. Bookmark & watch today. The most important shape nobody talks about.
Eron30,686 Aufrufe • vor 15 Tagen

In 1843, Hamilton realized complex numbers needed not one more dimension but two. One real part, three imaginary - i, j, k. He carved the equation into a Dublin bridge on October 16th. That number system now runs inside every 3D graphics engine, every robotics controller, and every LLM embedding space that maps language into high-dimensional geometry on GPU clusters today. Quaternion multiplication spins two perpendicular circles simultaneously - something that can only happen in four dimensions. i times j is k, but j times i is -k. Rotations don't commute, like a Rubik's cube. That's exactly why game engines and spacecraft use quaternions instead of matrices - no gimbal lock, no numerical drift. The same rotation math that every transformer uses when computing rotary position embeddings to encode token order in the complex plane. 2 minutes. Bookmark & watch today. The number system that powers every 3D engine and nobody taught you about.
Eron26,181 Aufrufe • vor 13 Tagen

Hamilton spent 10 years trying to build a 3D number system. On October 16, 1843, walking across a bridge in Dublin, he realized he needed 4 dimensions instead - and carved the equation into the stone on the spot. Quaternions are what every robotics system, 3D game engine, and LLM positional encoding falls back on when rotation math gets hard. This video builds the full 4D geometric intuition using stereographic projection - the same trick that maps a sphere onto a flat screen. 30 minutes. Bookmark & watch today. The Alice in Wonderland connection alone is worth it.
Eron159,228 Aufrufe • vor 3 Monaten

Jeff Hinton dismissed gradient descent in the 1970s because he thought models would get stuck. He won the Nobel Prize in 2024 partly for being wrong about that. Every time ChatGPT gets better after a training run, it's because 1.2 billion parameters fell through a wormhole in a loss landscape - not because they crawled downhill one step at a time. This video visualizes exactly what that looks like inside Llama 3.2. 20 minutes. Bookmark & watch today. The misconception that almost stopped AI is more interesting than the solution.
Eron130,619 Aufrufe • vor 2 Monaten

Heinz Hopf discovered this in 1931. Roger Penrose called it "an element of the architecture of our world." Eric Weinstein brought it up on Joe Rogan - and the silence in the room said everything. The Hopf fibration maps a 4D hypersphere onto a regular sphere using circles that never intersect but each links through every other exactly once. It shows up in at least 8 areas of physics - including the Bloch sphere geometry that every qubit in a quantum computer lives on. 6 minutes. Bookmark & watch today. The most important shape nobody talks about.
Eron122,559 Aufrufe • vor 2 Monaten

In 1859, Riemann wrote one sentence about where a function equals zero. 10 trillion cases checked, not a single counterexample - and it's still unproven. Worth $1 million. The zeros of the zeta function encode the exact position of every prime number as a sum of waves - the same Fourier decomposition that every transformer uses when its positional encoding layers convert token positions into sine and cosine frequencies. Each zero at 1/2 + iγ produces a wave in log x with height growing like √x and frequency γ. The first zero at 14.13i repeats every time x gets multiplied by 1.56. Add 100 pairs of waves and the prime count at 100 converges from 98.16 to within 0.4 of truth. If a single zero sits off the critical line, its wave would grow faster than √x and eventually swamp every other correction. In 2025, an AI agent called Gauss completed a machine-checked proof of the prime number theorem in Lean - verified line by line. But the full hypothesis needs an idea no human or AI has found yet. 10 minutes. Bookmark & watch today. The $1 million sentence that decides where every prime number lives.
Eron31,159 Aufrufe • vor 19 Tagen

In 1957, Stephen Smale proved you can turn a sphere inside out without cutting, tearing, or creasing it - even though you can't do it with a circle. The proof was so abstract it took 7 years before anyone found a practical method. The key is a topological invariant called the turning number - the same class of invariants that every LLM uses when its embedding layers preserve structure across high-dimensional transformations on GPU clusters. The trick is corrugation. Divide the sphere into guide strips connected by wavy segments - the waviness gives enough flexibility for strips to twist past each other without creasing. Push the polar caps through, twist them in opposite directions to convert loops into twisting, then push the equator back through the center. 16 rotated copies of one building block construct the entire eversion. A circle can't do this because its turning number changes from +1 to -1 when reversed. Spheres have an extra dimension where saddles, bowls, and domes interact - their invariant stays the same regardless of which face is out. The same dimensional escape that lets attention heads resolve contradictions in higher-dimensional latent space that look impossible in any lower projection. 19 minutes. Bookmark & watch today. The proof that spheres break a rule circles never can.
Eron28,367 Aufrufe • vor 21 Tagen

MIT just published a paper proving that scaling LLMs has a hard mathematical ceiling - and doubling the width to get there costs 6x more each time. A frontier model costs $300 million to train today. Double it once: $1.8 billion. Four more times: $390 billion for a single training run. Even if energy became free, the error rate would still drop at exactly the same speed. The exponent is locked by the structure of human language itself. The mechanism is superposition. GPT-3 has 12,000 dimensions but over 50,000 concepts to store - so concepts like "elephant" and "balloon" end up pointing in almost the same direction in embedding space. When the model activates one, the other bleeds through. That bleed is the interference causing hallucinations. Doubling the width cuts interference in half - but once every concept gets its own direction, scaling buys exactly nothing. 7 minutes. Bookmark & watch today. The hard limit on LLM scaling that money cannot fix.
Eron51,552 Aufrufe • vor 1 Monat

Every GPU running Claude and ChatGPT is built from billions of these. A MOSFET is just a voltage-controlled switch - raise the gate voltage, current flows. Drop it, current stops. That's it. That's the entire foundation of modern computing. No moving parts, no chemistry, just a thin layer of silicon responding to an electric field. Trillion-parameter AI models are nothing more than this switch flipping on and off at 3 gigahertz. 10 minutes. Bookmark & watch today. One transistor. Infinite scale.
Eron102,043 Aufrufe • vor 3 Monaten

Someone built a full quantum atom simulator in C++ with GPU ray tracing - and it proves the model you learned in school is completely wrong. Electrons don't orbit. They exist as standing waves across 100,000 possible positions at once, rendered in real time using the Schrödinger equation. The probability clouds look nothing like the textbook diagram. 14 minutes. Bookmark & watch today. The actual physics is more beautiful than the lie they taught you.
Eron89,243 Aufrufe • vor 2 Monaten

In 2015, every LLM researcher hit the same wall: deeper networks performed worse. Training a 56-layer model gave worse results than a 20-layer one — and no one knew why. The fix turned out to be a one-line hack that changed all of AI. Skip connections. Instead of forcing each layer to learn the full transformation, ResNet lets gradients flow directly through addition — bypassing the "shattered gradients" problem that made deep GPU training chaotic. The residual stream became working memory: the same architecture now lives inside every transformer, diffusion model, and LLM you use today. The most cited paper of the 21st century is 4 pages long. 20 minutes. Bookmark & watch today.
Eron56,900 Aufrufe • vor 1 Monat

William Tutte cracked Nazi's hardest cipher at 24. Decades later he published a paper called "How to Draw a Graph" — where the answer was just springs. Pin the outer boundary, convert every edge into a spring, let go. If the graph is polyhedral, the network untangles itself with zero crossing edges, every time. The reason it works: three-connectivity. Every node secured by at least three anchoring forces means no part can dangle, collapse, or fold. The same principle drives GPU mesh unwrapping in every AAA game — Tutte's spring equilibrium is literally how 3D artists flatten character skins into flat textures. Solving one linear system replaces 8,000 bouncing springs instantly. 22 minutes. Bookmark & watch today.
Eron52,227 Aufrufe • vor 2 Monaten

In 1852, Ludwig Schläfli proved that above four dimensions only three regular shapes can exist. The paper sat unpublished for 34 years - too far ahead of its time. Two dimensions have infinitely many regular shapes. Three have five. Four have six. Then the universe hard-codes exactly three forever - and every LLM embedding space and every quantum chip operates in the dimensions where this permanent pattern lives. The reason five platonic solids exist and not six is pure angle arithmetic. Stack regular polygons around a shared corner - sum below 360 degrees and you get a valid 3D corner, hit exactly 360 and you get a flat tiling, exceed it and the surface warps like a potato chip. Triangles give three valid solids, squares one, pentagons one. In four dimensions the same logic produces six polytopes - including the 600-cell with 600 tetrahedral faces and 1152 symmetries. Above four dimensions only three families survive: the simplex, the hypercube, and the cross-polytope. The hypercube tiles n-dimensional space the way cubes tile 3D - the same structure indexing positions across every transformer attention grid today. 25 minutes. Bookmark & watch today. The sequence 5, 6, 3, 3, 3, 3 and the geometry behind every dimension your model operates in.
Eron24,592 Aufrufe • vor 28 Tagen

Iron Man's arc reactor is fiction. The physics inside it is not. Copper coils generate magnetic fields that trap plasma at 100 million degrees. The Lorentz force spirals every charged particle along field lines - it can never escape. The same electromagnetic math governing electron flow through every transistor on every GPU running LLM inference today. Deuterium and tritium fuse into helium plus a free neutron - 17.6 MeV per reaction. A tiny amount of missing mass becomes energy via E=mc². Double the current in the coils and stored energy grows 4x. At 35% thermal efficiency, 10 MW of fusion heat gives 3.5 MW of electricity - and the machine keeps a chunk for itself. Every real tokamak on Earth - ITER, JET, KSTAR - runs on this exact physics. The same energy conversion pipeline that every AI data center uses when turning raw power into the computation behind every transformer. 3 minutes. Bookmark & watch today. The real physics behind the most famous reactor in fiction.
Eron10,257 Aufrufe • vor 10 Tagen

In 1971 the US government funded a speech AI that used a hand-crafted knowledge graph of 14,000 nodes. It worked. Then hidden Markov models replaced it entirely by learning from data - and outperformed it by orders of magnitude. Richard Sutton called this the bitter lesson. In 2025 Sutton said LLMs are repeating the same mistake. Training on human-generated text is just a more sophisticated version of building human knowledge into the system. AlphaGo Zero - trained with zero human games, only reinforcement learning - discovered strategies experts called "from an alternate dimension." 22 minutes. Bookmark & watch today. The most important question in AI nobody is asking loudly enough.
Eron57,301 Aufrufe • vor 2 Monaten

As someone who ships LLM systems in production, this Laplace transform video is the closest thing to a "why Mamba beats transformers" explainer I've ever seen released for free. Everyone thinks attention is what makes LLMs powerful. Laplace proved in 1785 that any signal breaks into exponential pieces - locate the poles, expose the frequencies. Mamba and Claude run on this exact trick. 34 minutes. Bookmark & watch today. The math is older than the United States.
Eron67,007 Aufrufe • vor 3 Monaten