[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74289":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":15,"stars7d":16,"stars30d":17,"stars90d":14,"forks30d":14,"starsTrendScore":18,"compositeScore":19,"rankGlobal":9,"rankLanguage":9,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":21,"topics":24,"createdAt":9,"pushedAt":9,"updatedAt":33,"readmeContent":34,"aiSummary":35,"trendingCount":14,"starSnapshotCount":14,"syncStatus":16,"lastSyncTime":36,"discoverSource":37},74289,"thereisnospoon","dreddnafious\u002Fthereisnospoon","dreddnafious","A machine learning primer built from first principles. For engineers who want to reason about ML systems the way they reason about software systems.",null,"Python",1146,94,5,0,1,2,19,3,58.83,"MIT License",false,"main",true,[25,26,27,28,29,30,31,32],"deep-learning","engineering","fundamentals","machine-learning","neural-networks","primer","transformers","tutorial","2026-06-12 04:01:14","# There Is No Spoon\n\n![There Is No Spoon — A Machine Learning Primer](figures\u002Flogo.svg)\n\nA machine learning primer built from first principles.\nWritten for engineers who want to reason about ML systems the way they reason about software systems.\n\n\u003Cbr>\n\n## 🎯 Who This Is For\n\nYou're a strong engineer. You can draw a software system on a whiteboard from your own hard-earned mental model. You understand tradeoffs — maintenance vs elegance, performance vs complexity.\n\nYou have a gut for software design. **You don't have that gut for machine learning yet.**\n\nYou know the tools exist but you can't feel when to reach for which. This primer builds that intuition.\n\n\u003Cbr>\n\n## 💡 What Makes This Different\n\nThis isn't a textbook or a tutorial. It's a **mental model** — the abstractions you need to reason about ML systems the way you already reason about software systems.\n\nEvery concept is anchored in **physical and engineering analogies**:\n- Neurons as polarizing filters\n- Depth as paper folding\n- Gradient flow as pipeline valves\n- The chain rule as a gear train\n- Projections as shadows\n\nThese analogies aren't decorative — they're the primary explanation, with math as the supporting detail.\n\nThe focus is **when to reach for which tool and why** — not just what each tool does, but the design decision it represents and the tradeoffs it implies.\n\n\u003Cbr>\n\n## 📐 What It Covers\n\nThe primer is organized in three parts:\n\n\u003Cbr>\n\n🧱 **Part 1 — Fundamentals**\n\nThe neuron, composition (depth and width as paper folding), learning as optimization (derivatives, chain rule, backprop), generalization, and representation (features as directions, superposition).\n\n🏗️ **Part 2 — Architectures**\n\nThe combination rule family (dense, convolution, recurrence, attention, graph ops, SSMs), the transformer in depth (self-attention, FFN as volumetric lookup, residual connections), encoding, learning rules beyond backprop, training frameworks (supervised, self-supervised, RL, GANs, diffusion), and matching topology to problem.\n\n🚦 **Part 3 — Gates as Control Systems**\n\nGate primitives (scalar, vector, matrix), soft logic composition, branching and routing, recursion within a forward pass, and the geometric math toolbox (projection, masking, rotation, interpolation).\n\n\u003Cbr>\n\n## 📖 Read It\n\nThe primer is a single markdown file with inline visualizations:\n\n### **[ml-primer.md](ml-primer.md)**\n\n\u003Cbr>\n\nJump to a specific topic:\n\n| | Topic | What it covers |\n|---|-------|---------------|\n| ⚡ | [The Neuron](ml-primer.md#-the-neuron) | Start here — dot product, bias, nonlinearity |\n| 📄 | [Composition](ml-primer.md#-composition-depth-width-and-paper-folding) | What depth buys you — the paper folding model |\n| 📉 | [Learning](ml-primer.md#-learning-as-optimization) | Derivatives, chain rule, backprop, loss landscape |\n| 🎯 | [Generalization](ml-primer.md#-generalization) | Why overparameterized networks work at all |\n| 🧠 | [Representation](ml-primer.md#-representation-what-networks-actually-store) | Features as directions, superposition |\n| 🔀 | [Combination Rules](ml-primer.md#-the-combination-rule-family) | Convolution vs attention vs recurrence vs graph vs SSM |\n| 🤖 | [The Transformer](ml-primer.md#-the-transformer) | Self-attention, FFN, residual connections |\n| 🏋️ | [Frameworks](ml-primer.md#%EF%B8%8F-frameworks) | Supervised, self-supervised, RL, GANs, diffusion |\n| 🗺️ | [Topology](ml-primer.md#%EF%B8%8F-topology-for-the-problem) | Matching architecture to problem — worked examples |\n| 🧩 | [Design Patterns](ml-primer.md#-design-patterns) | Common problems → which tool to reach for |\n| 🚦 | [Gates](ml-primer.md#-gates-as-control-systems) | The practitioner's gating and control toolkit |\n| 🔧 | [Diagnostics](ml-primer.md#-appendix-diagnosing-and-fixing-training-problems) | Loss curve symptoms, sanity checks, LR tuning |\n\n\u003Cbr>\n\nThe syllabus shows the full topic map: **[SYLLABUS.md](SYLLABUS.md)**\n\n\u003Cbr>\n\n## 🧭 How to Use This\n\nThis primer was built through conversation — one concept at a time, each stress-tested with questions until the mental model held up. You can use it two ways:\n\n\u003Cbr>\n\n**📚 Solo reading**\n\nRead it front to back, section by section. When something doesn't click, stop and re-read the section it depends on.\n\nThe primer is designed so that each section builds load-bearing intuition for the next. Don't skip ahead — the later sections assume you've internalized the earlier ones, not just read them.\n\n\u003Cbr>\n\n**💬 Interactive exploration with an AI agent**\n\nThis is the more powerful approach, and closer to how the primer was actually built. Feed the primer (or a section of it) to your preferred AI coding assistant and explore it conversationally:\n\n```\nRead ml-primer.md. I'm an engineer learning ML fundamentals.\nWalk me through the section on [topic]. I want to understand\nit well enough to reason about design decisions, not just\nrecite definitions. Push back if I get something wrong.\n```\n\nAsk \"why\" questions. Propose wrong answers and see if the agent catches them. Ask for concrete examples. Ask what would happen if you changed one thing. Ask how two concepts relate.\n\nThe primer gives both you and the agent a shared vocabulary and a correct conceptual framework — the conversation fills in everything a static document can't.\n\n> The primer is the map. The conversation is the territory.\n\n\u003Cbr>\n\n## 🖼️ Visualizations\n\n12 figures covering neurons, activation functions, paper folding, derivatives, chain rule, attention, FFN volumetric lookup, residual connections, dot products, loss landscapes, combination rules, and gating operations.\n\nAll generated from Python scripts in [`scripts\u002F`](scripts\u002F). To regenerate:\n\n```bash\npython3 scripts\u002F01_neuron_hyperplane.py\npython3 scripts\u002F02_activation_functions.py\n# ... etc\n```\n\nRequires `matplotlib` and `numpy`.\n\n\u003Cbr>\n\n## 📝 Origin\n\nBuilt through an extended conversational exploration between a software engineer and Claude, where every concept was stress-tested through questions, analogies were iterated until they landed, and misconceptions were corrected in real time.\n\nThe result is closer to a distilled mentorship than a reference document.\n\n\u003Cbr>\n\n## 🤝 Contributing\n\nPRs welcome. The goal is high signal — if you can explain a concept more clearly, fix an error, or add a section that fills a gap, open a PR.\n\nKeep the tone:\n- Direct, concrete\n- Analogies over notation\n- When-to-use over how-it-works\n\n\u003Cbr>\n\n## 📄 License\n\nMIT\n","这是一个面向工程师的机器学习基础教程，旨在帮助他们以软件系统思维理解机器学习系统。项目通过物理和工程类比（如将神经元比作偏振滤光片、深度比作纸张折叠等）来构建对机器学习核心概念的直观理解，并强调何时及为何选择特定工具。适合已经具备扎实软件工程背景但希望增强在机器学习领域直觉与设计感的技术人员使用。采用Python语言编写，涵盖了从基本原理到架构设计乃至控制系统的全面内容，包括但不限于神经网络、优化方法、表示学习以及最新的Transformer模型等。","2026-06-11 03:49:49","high_star"]