Loading
1. Basic Structure A dendritic tree is the branching, tree-like extension of a neuron. It spreads outward from the soma (cell body) and serves as the neuron’s primary input system.…
Read moreHere’s a practical blueprint for building silicon devices that behave like different neuron classes—not just “a neuron,” but excitatory vs. inhibitory, modulatory (dopamine-like), place/grid cells, etc. Four layers you can…
Read moreREFERENCE POST 1) Adaptive cycles / neuroscience analogs https://dspace.stir.ac.uk/bitstream/1893/31784/1/1-s2.0-S0149763420305753-main.pdf — Review detailing how cholinergic REM states drive apical dendritic integration (“apical drive”) that biases cortex toward internally generated processing. https://pmc.ncbi.nlm.nih.gov/articles/PMC5462120/…
Read moreImagine we can engineer at true micro-/nano-scale with almost arbitrary materials (diamondoid, graphene, soft polymers, protein scaffolds, photonics). Below is a systems-analyst, ground-up build of a digital cell that physically…
Read moreImproving AI systems requires more than generating fluent answers — it requires building mechanisms that distinguish between when a model is right and when it is wrong. This article outlines…
Read moreAbstract Artificial intelligence research often relies on engineering metaphors — pipelines, stacks, and nodes. Yet biology has long shown that life thrives not through rigid architectures but through specialization, symbiosis,…
Read moreComputers have always spoken in the language of 1s and 0s—but what if they could also think in threes? This article explores the design of a trinary accelerator card: a…
Read moreSignals (what moves through the mesh) Think “nervous impulses,” but they’re compact, content-addressed messages. Each signal is idempotent, verifiable (by hash/signature), and small enough to gossip freely. HELLO Purpose: neighbor…
Read moreA truly tiny C++ compiler doesn’t exist because the language is huge. But you can still get a practical, space-efficient toolchain that compiles real C++ (C++17/20) for your “digital DNA…
Read moreCompiler Processing Unit (CPU*) a pure graph/dataflow accelerator with obscene memory bandwidth. Below is a concrete, buildable spec you can prototype on FPGA and later harden into an ASIC. (CPU…
Read moreFor years, artificial intelligence has been engineered as if it were a product: you design an architecture, train a model, and deploy a static artifact. But biology doesn’t build intelligence…
Read moreAbstract Modern artificial intelligence architectures, epitomized by deep neural networks and Large Language Models, exhibit striking limitations in long-term reasoning and stability. They simulate the brain’s neurons but ignore the…
Read moreWhen most people think of NVIDIA, they picture GPUs — the chips behind ChatGPT, self-driving cars, and photorealistic gaming. But here’s the plot twist: NVIDIA’s biggest long-term advantage in AI…
Read moreThe Transformer Bottleneck The transformer architecture revolutionized natural language processing—but its greatest strength is now a core limitation. Self-attention enables models to consider all tokens simultaneously, but this also means…
Read more1. High-Level Philosophy RWKV solves a core limitation in traditional Transformers: quadratic attention cost with sequence length. Instead of computing full attention matrices, it introduces a mechanism to simulate attention…
Read more
