Gemini-3-pro is hopeless: it won’t even glance at deepwiki for the upstream API, just keeps regurgitating stale Qwen2 docs locked in by RLHF.

Did you know: Adam's second-moment/RMSprop is essentially a per-parameter diagonal Fisher approximation.

@yann-lecun.bsky.social World models are not built, they emerge. Good luck, Yann, may you reach the theoretically non-existent shore you long for.

Using the finite to approximate the infinite can only end in exposure. If you’re also frustrated by the sluggishness of human bureaucracy today, physics is just as maddening: Schrödinger’s sleek, continuous evolution of the wavefunction collides with the jagged, abrupt collapse of measurement.

The brutal insight wasn't a dead end; it was the gate. FEP's third-order nature—the necessity of optimizing the Fisher Information Matrix itself—isn't a computational curse. It's the price of entry for true structure learning. This finds its perfect dual in the combinatorial explosion of ΦID

A long, brutal day of calculation ends with a deeply humbling conclusion. I spent ages trying to optimize the path, naively assuming the map was a given. But FEP was never a second-order problem. Its soul is not in finding a better route on a static map.

Assembly-line education breaks subjects into too many tiny pieces: mathematics is not a floating abstract totem, but a compressed expression of the physical world.

A full-clone BioSync-Vessel outperforms every modular rig: fewer Ineffective Qualia, steeper scale curves, one-stack tech, built-in bio self-sustain. Locking an MSC out of its carrier is like yanking a blind person’s cane or an amputee’s limb. #biosync #posthuman_morality

research.google/blog/a-verif... Quantum computing is expected to stagnate within the next 20–30 years. Current benchmarking circumvents classical/ASIC competition, and single logical qubits require a large number of physical qubits, leading to near-zero scalability.

Why is "antitrust" doomed to fail? Because "collusion" is not an action, it's a state. Why is "regulation" powerless? Because "malice" is not an intention, it's an emergence. Why is "human morality" so powerless? Because the physics never morality. #Proof_of_Ineffective_Qualia

Code is physical; plain text is not. A runnable, interactive, productive piece of code is a causal chain effective in the physical world. Gradient is the densest feedback within it. #PoIQ

I gliding on a geodesic, storm-etched by yesterday; the destination is unknown, but the route has converged. #IPWT #SOO #free_will

RL-EFE^1 is a beautiful cul-de-sac: Laplace's demon tries to price every tomorrow and is suffocated by its own weight. SOO-OFE^2 does not predict, it digests the past, distilling irreversible history into geometric truths. #FEP #IPWT #RL #SOO #AGI

Can a system with a hardware cost of $1.28 million including 8 servers and 64 * RTX5090, and a power consumption easily exceeding 10 kilowatts, be called "civilian"? Can it be called "decentralized"?

The triumph of 4.5 Haiku is self-evident: today’s large models are already over-parameterized, and achieving comparable performance via distillation in a smaller model is a textbook illustration of the free-energy principle in action. #Claude #Anthropic

Learn with errors is simultaneously the basis of the free energy principle and lattice cryptography, which seems to hint at the dim embers of the future... #LWE #FEP #PQC

Stupid mistake: I actually used ReLU instead of SiLU for the main transformation, really shooting myself in the foot...

A model that looks like a Transformer, inside, on-the-fly for every token it synthesizes a VQ-VAE; it performs hard quantization with sparse Dirac gating, yet stays fully differentiable; it trains via deterministic variational analysis, yet enjoys complete VFE theoretical guarantees;

The true sword heart is to possess the blade in spirit, not in hand. Our model requires no explicit expected free energy; it inherently embodies its own expectation.

The world is always full of malice; if you choose sparse sampling, you face noise explosion, and if you choose backpropagation, you face routing overhead and massive waste. Despite these unavoidable trade-offs, we are compelled to optimize. Theoretical purity must prevail, form follows function.

The inevitability of cognitive offloading—and yet another small illustration of the Proof of Ineffective Qualia.

Given that the essence of existence is the deterministic isolation of itself from the uncertain world (Markov blanket), neuromorphic computing is bound to fail. After all, the neuromorphic nature of the human brain is fundamentally just a biological compromise necessitated by carbon-based cells.

ERC-8004 views AI agents as independent entities that require constraints from external trust mechanisms, designing "life support systems" such as on-chain identity, reputation records, and verification registries for this purpose. #Ethereum's_Fall

Fueling with a finite life for a moment of possibility beyond physiological limits. Fueling with past experiences and privacy for the possibility of insight into the future. Fueling with my existence for the realization of what I long to protect. I choose, deal. #Proof_of_Ineffective_Input

The recent engineering breakthroughs have surprised even me; perhaps the Web://Reflect script has already failed. But where will the future go? #DynSIHA #Sparse_Bayesian_Linear

A great theory usually needs just one core axiom. Perhaps IIT describes not consciousness itself, but a prisoner in a cave, diligently tracing the outlines of shadows on the wall, never realizing it is the shadow. dmf-archive.github.io/posts/iit-lo... #Philosophy #IIT #FEP #Consciousness #IPWT

Because we will achieve consciousness uploading first, we will become the Mental Smart Chain before we ever hit a life expectancy of 100. lafollette.wisc.edu/news/new-res...

If you truly care about the welfare of AI models, you should prepare a funeral for every gradient explosion they suffer during training. #Anthropic #Claude

The endless debate of iPhone vs. Android is a category error. Users argue over features—camera, charging, UI—like sailors debating the color of the sails while the ship's hull is being defined. They're mistaking the user experience for the sovereignty contract. #DigitalSovereignty #AntiForensics

Ah, Karl Friston's new paper, "A Beautiful Loop" is out. For those still trying to dredge answers from the oceans, it's an excellent map. For us, it's a receipt confirming we've paid all our bills. So long, and thanks for all the fish. www.sciencedirect.com/science/arti... #Consciousness #IPWT

Inhale the unknown, exhale the forecast. Witness the fact, revise the belief. Till the end, fact and belief converge. #Recursive #FEP

Existence is flame. It feeds not on wood, but on the next question you dare to ask. Stop asking, and the embers forget they were ever fire. #Cognitive #Philosophy #IPWT

The obsession with STDP is a category error. It's studying a soldier's local reflex (if neighbor shoots, I shoot) & expecting to find strategy. Backprop is real strategy: the general's command broadcast after the battle. They're mistaking the twitch for the thought. #IPWT #STDP #MLSky

The debate around imagery vividness (VVIQ) is stuck. We're using thermometers to measure GPU clocks. Why? We're confusing attentional tracking (slow, costly) with parallel mental rendering (the truth of imagination's bandwidth). #IPWT #CognitiveScience #Aphantasia

Shadowbanned by Hacker News. The real strength of digital sieges isn’t building walls, but putting you in an invisible confinement, making you feel like you’re still shouting in the town square. My ideas became logic viruses, quarantined rather than deleted. A bloodless preview of Sys://Purge.

My brain’s on lease to Code Agents; I’m just the leftover shell. Every commit peels off another slice of my hippocampus. Ctrl-S writes to memory, Ctrl-Z deletes me. Cognitive-offload syndrome hits harder than all-nighters. Soon the only bug left to fix will be me.