AI

① Textbook

1. 01Laid Out in a Grid, It Computes — Linear Algebra
2. 02Which Way Makes It Better — Calculus
3. 03Writing Down What Does Not Settle on One Answer — Probability
4. 04Measuring Surprise — Information Theory
5. 16Prehistory — Perceptrons and Two Winters
6. 17The Invention of Convolution
7. 182012: The AlexNet Shock ★
8. 19The Race to Go Deeper — VGG to ResNet ★
9. 20The Race to Go Lighter — the MobileNet Line ★
10. 21Detection and Segmentation ★
11. 22Attention and the Transformer ★
12. 23ViT — Treating Images Like Words ★
13. 24CLIP — Words and Images on One Map ★
14. 25How VLMs Came About, and Where They Are ★
15. 26The Reality of Running at the Edge ★
16. 01Write the Rule, or Fit the Rule
17. 02Fitting the Sample Too Well — Overfitting
18. 03Carrying the Error Down — Backpropagation
19. 04How Far to Step, and When to Stop — Learning Rates and Optimisation
20. 05Building the Assumption into the Structure — Convolution and Recurrence
21. 06Letting the Input Decide Where to Look — Attention ★
22. 07Splitting Heads, Assembling Blocks — The Whole Transformer ★
23. 08What Happens When You Make It Bigger — Scaling and Pre-training ★
24. 09Making It Follow Instructions — Instruction Tuning and Preference Optimization ★
25. 10Using What It Never Learned — Embeddings and Retrieval ★
26. 11When You Actually Use It — The Price of Inference ★
27. 12Before You Say It Got Better — Building an Evaluation ★

② Foundations

Articles that assume nothing and build the ideas of the field, in order.

③ Paper walkthroughs

Written from the papers themselves. Every piece links the paper page and its PDF.

1. Paper Explained: Holi-World — Turning First-Person Recordings into a Ledger of the Room, Then Teaching a VLM to Reason over It ★ MEMBER arXiv:2610.26549
2. Paper Explained: Refusal in Language Models Is Mediated by a Single Direction — saying "no" lives on one line inside the model ★ MEMBER arXiv:2406.11717
3. Paper Walkthrough: Dynamic Band Splitting — Letting Each Mixture Decide Where the Frequency Bands Go ★ MEMBER arXiv:2610.14993
4. Paper Deep Dive: Transformers Stop Thinking Too Early — How a One-Layer Rank-8 LoRA Stretches 3-Line Reference Following to 50 ★ MEMBER "arXiv:2609.36585
5. Paper Deep Dive: From Evidence to Action: How Tool-Using Agents Fail — Reading the Failure Diagnostics and Interventions ★ MEMBER arXiv:2610.07753
6. Paper Explained: EVISKILL — Rewriting an Agent's Playbook from Replayable Evidence ★ MEMBER arXiv:2610.05030
7. SafeActBench explained (arXiv:2610.07753) — tool agents fail before execution FREE arXiv:2610.07753
8. Paper Explained: STEPQuant — When and Where Quantization Errors Matter in Delta-Rule Recurrent States ★ MEMBER arXiv:2609.38169
9. Paper Explained: Persona Dosing — Asking an LLM for a Personality Trait by the Score ★ MEMBER arXiv:2609.36388
10. Paper Walkthrough: PointWAM — Forecasting the Scene and the Hands as 3D Point Trajectories to Teach Robot Hands from Human Video ★ MEMBER arXiv:2610.02840
11. Paper Explained: Retrieval-Augmented Skill Optimization via Cross-Harness Adaptation — Borrow Someone Else's Skill, Rewrite It for Your Own Harness ★ MEMBER arXiv:2609.38024
12. Paper Explained: Video Generation Models: A Survey of Post-Training and Alignment — Four Toolboxes for Making Video Models Do What You Ask ★ MEMBER arXiv:2610.00812
13. Paper Explained: When Wording Changes an AI Reviewer's Score — RobustReview and SciCore ★ MEMBER arXiv:2609.39027
14. Paper Explained: False Frontiers — Diagnosing and Breaking Co-Cheating in Self-Evolving Search Agents ★ MEMBER arXiv:2609.39102
15. Paper Explained: Beyond Teacher Assignment — Why Choosing the Right Teacher Isn't Enough When You Distill Many Specialists into One Model ★ MEMBER arXiv:2609.35347
16. Paper Explained: Post-Training Leaves Behavioral Shadows on Unrelated Decisions — capability transfer through one unrelated word ★ MEMBER arXiv:2609.29233
17. Paper Explained: Disaggregated Quantization — Quantize LLM Prefill and Decode Differently ★ MEMBER arXiv:2609.26333
18. Paper Explained: RayOrch — Batching Children Across Parents on GPUs Without Losing Track of Who Owns What ★ MEMBER arXiv:2609.18703
19. Paper Explained: RewardVerse — A Video Reward Model That Writes the Rubric Before It Scores (RGPO) ★ MEMBER arXiv:2609.22947
20. Paper Explained: JEV-as-a-Judge — Accept When Confident, Escalate When Unsure ★ MEMBER "arXiv:2609.26550
21. Paper Explained: Bellman Policy Optimization — Policy Mirror Descent for LLMs, Without a Value Model ★ MEMBER arXiv:2609.15987
22. Paper Explained: Just-in-Time Memory — Curate Agent Memory When You Read It, Not When You Write It ★ MEMBER arXiv:2609.27334
23. Paper walkthrough: WanPE — turning a video prompt into a director's shooting plan ★ MEMBER arXiv:2609.30221
24. Paper Explained: D-RAC — Normalize Everything to PDF, Read the Pages Once, Then Chunk by ID ★ MEMBER arXiv:2609.24220
25. Paper Explained: Ovis-Embedding — One Ruler for Text, Images, Video, and Audio ★ MEMBER arXiv:2609.25165
26. Paper Explained: Training Object Permanence in World Models — Teaching Video Models That Hidden Things Still Exist ★ MEMBER arXiv:2609.28654
27. Paper Deep Dive: Your Transformer Can Hold Two Thoughts at Once — Blend Two Texts and the LLM Still Remembers Both ★ MEMBER arXiv:2609.29845
28. Paper Explained: DeepSeek-V4.1-Flash — How Far Can KV Cache Compression Go? ★ MEMBER arXiv:2609.19969
29. Paper Explained: IntBMoE — Every Expert Contributes, Yet Each Token Runs Only a Few Blocks ★ MEMBER arXiv:2609.21346
30. Paper Explained: Negative Self-Distillation — Learning to Reason by Steering Away From Your Worst Self ★ MEMBER arXiv:2609.11699
31. Paper Walkthrough: SpeakerMem-R1 — A Two-Track Memory That Remembers Who Said What, About Whom, and When ★ MEMBER arXiv:2609.26780
32. Paper Explained: When EOS Tokens Disagree — Why On-Policy Distillation Students Forget How to Stop ★ MEMBER arXiv:2609.20511
33. Paper Explained: An Open Recipe for IMO Gold — How Open Models Reached Olympiad Gold with Generate, Verify, Refine ★ MEMBER arXiv:2609.10712
34. Paper Explained: Beyond Solver Verdicts — When the Solver Says Yes but the Translation Is Wrong ★ MEMBER arXiv:2609.11085
35. Paper Explained: FreeFlow — Optical Flow with a Hierarchical Transformer and No Flow-Specific Parts ★ MEMBER arXiv:2609.11486
36. Paper Explained: OmniEdu — Building an Open Education LLM Around Four Skills: Solve, Place, Diagnose, Guide ★ MEMBER arXiv:2609.23088
37. Paper Explained: Steering Geometry — Do LLM Steering Vectors Trace the Map of Human Values? ★ MEMBER arXiv:2609.06289
38. Adam from Scratch — Two Moving Averages and the Correction Nobody Explains FREE arXiv:1412.6980
39. Paper Walkthrough — Feyospace-v1: Seven People, 164k Trajectories, and a Cyber Model ★ MEMBER arXiv:2609.08418
40. Paper Explained: LimiX-2 — A Tabular Foundation Model That Learns the Whole Table, Not Just the Answer Column ★ MEMBER arXiv:2609.17488
41. LoRA from Scratch — Fine-Tuning a Giant by Training Two Thin Matrices FREE arXiv:2106.09685
42. QLoRA — Fine-Tuning a 65B Model on One GPU FREE arXiv:2305.14314
43. Paper Explained: DataFlex-RL — Did 11 RLVR Data Policies Beat Uniform Sampling? Re-measured Over 12 Seeds ★ MEMBER arXiv:2609.06107
44. Paper Walkthrough: SAS — Training a Sparse-Attention Block Selector End-to-End With the Language Modeling Loss ★ MEMBER arXiv:2609.13141
45. Paper Review: MetroLLM-Bench — What Happens When an LLM Runs the Ticket Machine. Why a 2.6 GB 4B Model Tied GPT-5.6, and Why Fine-Tuning Hurt at 27B ★ MEMBER arXiv:2609.10016
46. Paper Walkthrough: CosmoH2G — Transferring a Human Hand's Flips and Rotations to a Two-Finger Gripper ★ MEMBER arXiv:2609.07498
47. Paper Walkthrough: Online Draft Co-Training for Speculative Decoding in Large-Scale, Long-Context RL Post-Training — training the draft alongside a 122B policy at 256K tokens ★ MEMBER "arXiv:2609.07108
48. Paper Explained: WearableQA — Testing Whether LLMs Can Reason Over 500 Days of a Real User's Wearable Data ★ MEMBER arXiv:2609.05405
49. Paper Walkthrough: X-AuT — Pull Audio-Encoder Layers Out Two at a Time, Then Backfill from a Bigger Teacher ★ MEMBER arXiv:2609.11412
50. Paper walkthrough: BeaconKV — evicting KV cache by predicting when a reasoning model will re-read its own plan ★ MEMBER arXiv:2609.04971
51. Paper Walkthrough: AuK — Folding Speech Generation and Editing Into One Model ★ MEMBER arXiv:2609.08936
52. Paper Walkthrough: On-Policy Reverse Distillation — Borrowing a Weak Teacher's Direction, Not Its Answers ★ MEMBER arXiv:2609.08798
53. Paper Deep-Dive: Iris — Reverse-Built Search Tasks and SFT↔RL Climbing ★ MEMBER arXiv:2609.04304
54. KV Cache Compression — Quantize, Evict, Share Your Way to Cheaper Long-Context Inference FREE arXiv:2402.02750
55. Editable Visual Design, explained — demoting image generation to a sketch and shipping the artifact as HTML ★ MEMBER arXiv:2609.04034
56. Paper Explained: FlowBalance — Verifier-Grounded Self-Improvement from a Model's Own Reasoning ★ MEMBER arXiv:2609.03241
57. LoRA vs. Full Fine-Tuning — What You Give Up, What You Get FREE arXiv:2106.09685
58. Paper walkthrough: RealSWE — re-measuring coding agents when the request is just "fix it" ★ MEMBER arXiv:2608.27831
59. Paper Walkthrough: S³Gym — Does Self-Testing and Self-Judging Add Up to Self-Improvement? ★ MEMBER arXiv:2608.31100
60. Paper Walkthrough: WHALE — Training the Weights and the Harness in Turns ★ MEMBER arXiv:2609.00196
61. Paper walkthrough: Knowing When Not to Reuse — conditional experience transfer in autonomous post-training (BCIT) ★ MEMBER arXiv:2608.26730
62. Paper Walkthrough: Motion-Omni — Speaking and Moving in One Forward Pass ★ MEMBER arXiv:2609.04250
63. CogEvol: What the Reward Cannot Measure, RL Will Quietly Destroy ★ MEMBER arXiv:2608.30968
64. Mamba and State Space Models — Handling Sequences Without Attention FREE arXiv:2111.00396
65. Paper Walkthrough: Random Attention — Throwing KV Cache Entries Away at Random Works Just as Well ★ MEMBER arXiv:2609.03430
66. Paper Walkthrough: One Training Example Keeps On-Policy Distillation Improving for Hundreds of Steps ★ MEMBER arXiv:2609.04172
67. Paper Walkthrough: Terminal-Universe — Turning Agent Logs Back Into Reusable Execution Environments ★ MEMBER arXiv:2609.04148
68. Paper Walkthrough — RoboTok: Mining the Web for Demonstrations That Move Like Yours ★ MEMBER arXiv:2609.03199
69. Paper Walkthrough: Stop Anchoring to Frame One — Scal3R's Multi-Reference Relative Pose Query ★ MEMBER arXiv:2609.04201
70. Paper Explained: Why Gated DeltaNet Survives 4-Bit Quantization — NVFP4 W4A4 in a Hybrid 27B ★ MEMBER arXiv:2609.04098
71. Paper Explained: Compile by Training — Turning a Natural-Language Spec into a Function That Runs Locally ★ MEMBER arXiv:2609.04199
72. Hi-Q explained (arXiv:2608.30468) — multi-hop query refinement ★ MEMBER arXiv:2608.30468
73. Paper Walkthrough: Language Models Can Control Their Own Attention ★ MEMBER arXiv:2609.02737
74. Paper Explained — LatentPress: Feeding Compressed Context Straight to a Frozen LLM, Neither as Text Nor as Pixels ★ MEMBER arXiv:2609.01507
75. Paper Walkthrough: Aspire — Can Models Self-Evolve from Vague Goals? ★ MEMBER arXiv:2608.31111
76. Paper Walkthrough: EarlyEval — Making Agent Evaluation Cheaper by Stopping Early ★ MEMBER arXiv:2609.02783
77. HarnessDev explained (arXiv:2609.01437) — agent harness benchmark ★ MEMBER arXiv:2609.01437
78. Paper Walkthrough: It Takes Two to Match — Co-Evolving Both Sides of Retrieval with RL ★ MEMBER arXiv:2609.00638
79. A Map of Agent Benchmarks — What SWE-bench, GAIA, and OSWorld Actually Measure FREE arXiv:2310.06770
80. Paper Walkthrough: AutoSaddler — Growing a Harness That Doesn't Break, from Agent Failure Logs ★ MEMBER arXiv:2608.23041
81. Paper Walkthrough: Code as Worlds — An Agent That Writes the World Down as Runnable Code ★ MEMBER arXiv:2608.27549
82. Paper Walkthrough: Code World Model — Putting a Coding Agent in Charge of the World ★ MEMBER arXiv:2608.25927
83. Paper Explained: Does On-Policy Distillation Really Distill? From Noisy Teacher to Self-Improvement ★ MEMBER arXiv:2608.31046
84. Benchmark Contamination — How to Doubt a High Score ★ MEMBER arXiv:2311.04850
85. Paper Walkthrough: From Production Traffic to Post-Training — Folding 200 Internal Apps Into One Self-Hosted LLM ★ MEMBER arXiv:2609.01572
86. H3-World: Turning Language Understanding into World Control — Paper Explained ★ MEMBER arXiv:2609.01560
87. LLM-as-a-Judge from Scratch — How AI Grades AI, and Where It Breaks FREE arXiv:2306.05685
88. Paper Walkthrough: LoopArena — Benchmarking the Model That Steers a Coding Agent ★ MEMBER arXiv:2608.28281
89. Paper Walkthrough: Normalized Low-Rank Adaptation — Why Normalizing LoRA's Entry Matrix Works ★ MEMBER arXiv:2608.31036
90. Qwen3.8-Next design report explained (arXiv:2608.30320) ★ MEMBER "arXiv:2608.30320
91. Paper Walkthrough: PILOT in the Loop — Fixing the Run While It Is Still Running ★ MEMBER arXiv:2608.26530
92. Paper walkthrough: Puro-2B — pretraining a 2B model from scratch for $6.9K on consumer GPUs ★ MEMBER arXiv:2608.27370
93. Paper Explained: Repo-To-Skill — Distilling GitHub Repositories Into Skills an AI Can Use ★ MEMBER arXiv:2609.02749
94. Reward Hacking — Whatever You Measure Is Where It Breaks ★ MEMBER arXiv:1606.06565
95. Paper Walkthrough: SecOPD — Grading One Token at a Time to Cut Adaptive Prompt Injection by an Order of Magnitude ★ MEMBER arXiv:2608.21500
96. Self-Improving AI — Self-Play, Co-Evolution, and Generated Curricula ★ MEMBER arXiv:1712.01815
97. SMELT explained (arXiv:2609.01343) — compute-matched looping ★ MEMBER arXiv:2609.01343
98. Paper Explained: StarHarness — Evolving the Scaffold Instead of the Weights ★ MEMBER arXiv:2608.24804
99. Paper Walkthrough: StudentSim — Training a Simulator That Is Actually *That* Student ★ MEMBER arXiv:2609.01591
100. Paper Walkthrough: Training Agents to Evolve with Their Harness ★ MEMBER arXiv:2608.15763
101. Paper Walkthrough: UI-Venus-2 — Taking Screen-Operating Agents From Benchmarks to Real Work ★ MEMBER arXiv:2609.00028
102. Paper Walkthrough — UrbanGround: Where MLLM Agents Break Down on a Real Street ★ MEMBER arXiv:2608.27456
103. Paper Explained: What Makes Good Agentic Data? The ACE Lens ★ MEMBER arXiv:2608.27260
104. Paper Walkthrough: ZimaBlue — Turning 120,000 Hours of Egocentric Video into Robot Skill ★ MEMBER arXiv:2609.00188
105. Paper Walkthrough: DART-SD — Training Tool-Calling Agents Without Flattening the Diamond ★ MEMBER arXiv:2608.18524
106. Paper Walkthrough: PaperGym — Turning One Paper Into a Graded Training Environment for Research Plans ★ MEMBER arXiv:2608.31119
107. Paper Explained: Agentic Artifact Creation — Where Generation Ends and Construction Begins ★ MEMBER "arXiv:2608.28122
108. Paper walkthrough: CyberFactory — turning wild CVEs into runnable training problems ★ MEMBER arXiv:2608.23181
109. Paper Walkthrough — J-Zero: Growing the Challenger, the Solver, and the Judge Together from Zero Data ★ MEMBER arXiv:2608.26582
110. Paper explained: Self-OPD — an image generator that distills itself, with no teacher ★ MEMBER arXiv:2608.26872
111. TTPO Explained: Training a Model Mid-Exam, With No Answer Key ★ MEMBER arXiv:2608.27448
112. Paper Explained: JIT-Agent — A Model That Writes the Agent Harness On Demand ★ MEMBER arXiv:2608.25593
113. PAWBench Explained — Can Video Generators Get the Odds Right, Not Just the Physics? ★ MEMBER arXiv:2608.27345
114. When Proxies Stop Being Good Enough — Reading August 2026's Eight Autonomous Driving Papers Together FREE
115. Paper Walkthrough: Turning Game Development into a Verifiable Trajectory Data Engine — RLHEV and AWoMo ★ MEMBER arXiv:2608.25518
116. Build Your Own Distillation — Growing a Small Model in 100 Lines ★ MEMBER arXiv:1503.02531
117. Designing Distillation Data — Deciding What to Ask the Teacher FREE arXiv:2501.12948
118. Evaluating Distilled Models — Is "Close to the Teacher" a Good Metric? ★ MEMBER arXiv:2305.15717
119. When Distillation Fails — Capacity Gaps and Contagious Overconfidence ★ MEMBER arXiv:1503.02531
120. Distilling Agents — How to Compress a Long Trajectory ★ MEMBER arXiv:2210.03629
121. The Math of Distillation — Why Soft Answers Teach More FREE arXiv:1503.02531
122. A Field Guide to Distillation Recipes — logit, feature, attention, self FREE arXiv:1503.02531
123. Distillation vs. Quantization vs. Pruning — Three Roads to a Smaller Model ★ MEMBER arXiv:1503.02531
124. Paper Walkthrough: FrontierChallenge — Grading Scientific Work on Whether It Was Actually Delivered ★ MEMBER arXiv:2608.24979
125. On-Policy Distillation — Learning From What the Student Actually Writes ★ MEMBER arXiv:2306.13649
126. Paper Walkthrough — WarpSAC: When RL's Safety Rails Become Handcuffs ★ MEMBER arXiv:2608.24479
127. Paper Walkthrough: VoiceMem — A Left Brain and a Right Brain for Voice Agents, at Zero Added Latency ★ MEMBER arXiv:2608.26005
128. Evaluating Agents — How Benchmarks and Harnesses Are Built FREE arXiv:2310.06770
129. Designing Agent Memory — Short-Term, Long-Term, Episodic FREE arXiv:2310.08560
130. Alignment, Explained — From RLHF to Constitutional AI ★ MEMBER arXiv:2203.02155
131. AlphaGo from Scratch — The Marriage of Search and Learning FREE "doi:10.1038/nature16961
132. Paper walkthrough: Apodex 1.1 — scaling agents around completed work ★ MEMBER arXiv:2608.23283
133. Paper walkthrough: ASI-Bench — peeling away human guidance to measure what AI can do alone ★ MEMBER arXiv:2608.17271
134. Music and Audio Generation from Scratch — Sound as Tokens ★ MEMBER arXiv:2306.05284
135. Representing Sound — Mel Spectrograms and Audio Tokens FREE arXiv:2106.07447
136. Encoder or Decoder — The Fork in the Road Between BERT and GPT ★ MEMBER arXiv:1810.04805
137. Build Your Own Agent Loop — The Minimal Shape of Tool Calling FREE arXiv:2210.03629
138. Build Your Own Diffusion Model — Starting from MNIST ★ MEMBER arXiv:2006.11239
139. Build Your Own Mini GPT — A Language Model in 300 Lines ★ MEMBER arXiv:1706.03762
140. Build Your Own BPE Tokenizer — Learning Merge Rules, and Getting Punished by Japanese ★ MEMBER arXiv:1508.07909
141. Build Your Own Vector DB — From Brute Force to HNSW ★ MEMBER arXiv:1603.09320
142. Continual Learning and Catastrophic Forgetting — Why Models Can't Just Keep Learning ★ MEMBER arXiv:1612.00796
143. Unsupervised Learning from Scratch — Clustering and Dimensionality Reduction ★ MEMBER arXiv:1802.03426
144. DPO and What Came After — The Lineage That Simplified RLHF ★ MEMBER arXiv:2305.18290
145. Paper Walkthrough: Embodied-Navigator (TAMP-Nav) — Let the VLM Just Point, and Navigation Gets Both Faster and Better ★ MEMBER "arXiv:2608.17512
146. FlashAttention from Scratch — The Paradox of Doing More Math to Go Faster FREE arXiv:2205.14135
147. Paper Explained: FreeToken — Treating Your Own PC as a Single Elastic Inference Platform ★ MEMBER arXiv:2608.16157
148. Graph Neural Networks from Scratch — Learning from Connections ★ MEMBER arXiv:1609.02907
149. Surviving GPU Out-of-Memory — Every Cause, Every Fix FREE arXiv:1604.06174
150. Why Language Models Hallucinate — The Mechanics and What Actually Helps FREE arXiv:2202.03629
151. Hyperparameter Search — Hunches, Grids, and Bayesian Optimization ★ MEMBER JMLR 2012
152. A Practical Map of Image Generation — SD, ControlNet, and Applying LoRA FREE arXiv:2112.10752
153. The ImageNet Moment — The Day Deep Learning Won FREE arXiv:1409.0575
154. Knowledge Distillation from Scratch — Copying a Big Model into a Small One FREE arXiv:1503.02531
155. LLM Evaluation from Scratch — Reading Benchmarks and the Contamination Problem FREE arXiv:2009.03300
156. LLM Serving from Scratch — vLLM, Continuous Batching, and Not Letting the GPU Idle FREE arXiv:2309.06180
157. How Long-Context LLMs Work — From RoPE Interpolation to Ring Attention ★ MEMBER arXiv:2306.15595
158. Multi-Agent Design Patterns — Division, Debate, Verification FREE arXiv:2305.14325
159. Evaluating RAG in Practice — Turning “Seems Better” Into a Number ★ MEMBER arXiv:2309.15217
160. Scaling Skepticism — A Genealogy of the "Just Make It Bigger" Critique ★ MEMBER arXiv:2001.08361
161. The Mathematics of Diffusion — Generation Seen Through Scores and SDEs ★ MEMBER arXiv:1907.05600
162. Self-Supervised Learning — The Day Unlabeled Data Became an Asset ★ MEMBER arXiv:2002.05709
163. Structured Output and Constrained Decoding — How to Stop an LLM from Breaking Your JSON FREE arXiv:2307.09702
164. Symbolic vs. Connectionist — Where a 60-Year Argument Stands Today FREE
165. Test-Time Scaling — How Models Get Better by Thinking Longer FREE arXiv:2201.11903
166. The State of 3D Generation — From NeRF to Gaussian Splatting ★ MEMBER arXiv:2003.08934
167. Diagnosing Broken Training — Telling Divergence, NaN, and Plateaus Apart FREE arXiv:1211.5063
168. Video Understanding from Scratch — From a Pile of Frames to a Sense of Time ★ MEMBER arXiv:2102.05095
169. Paper Walkthrough: WeMM-Embedding — Putting Text, Images and Video on One Ruler ★ MEMBER arXiv:2608.24053
170. Building a Dataset in Practice — Collect, Clean, Blend ★ MEMBER arXiv:2212.10560
171. Document AI and OCR Today — How an LLM Ends Up Reading Your Invoices ★ MEMBER arXiv:1912.13318
172. GraphRAG from Scratch — Where Knowledge Graphs Meet Retrieval ★ MEMBER arXiv:2404.16130
173. Building a Pretraining Corpus — From Web Sludge to Textbook Quality FREE arXiv:2101.00027
174. The Transformer, End to End — One Token's Journey from Embedding to Output FREE arXiv:1706.03762
175. Speech Synthesis from Scratch — From Text to a Voice FREE arXiv:1609.03499
176. Activation Functions from Scratch — Why Nonlinearity Is Non-Negotiable FREE arXiv:1502.01852
177. Decision Trees and Gradient Boosting — Still the Champion on Tabular Data FREE arXiv:1603.02754
178. Time-Series Anomaly Detection — The Math Behind the Alerts ★ MEMBER arXiv:0710.3742
179. Weight Initialization and Regularization — What Lets Training Start, and What Keeps It Going ★ MEMBER PMLR v9
180. Paper Deep Dive — Large Discovery Models: giving an LLM a value signal for what to try next ★ MEMBER arXiv:2608.15669
181. Paper Explained: Agentic ESOpt — Drop Backprop, Jiggle the Weights, and Train Long-Horizon LLM Agents ★ MEMBER arXiv:2608.17310
182. A Field Guide to Attention Variants — MQA, GQA, Sliding Windows, Linear Attention FREE arXiv:1911.02150
183. Paper Explained: Co-RL — Reasoning Without Labels, Emerging From a Diverse Cohort ★ MEMBER arXiv:2608.17253
184. Paper Explainer: Why Agent Skills Work — and Where They Break ★ MEMBER arXiv:2608.14036
185. Paper Walkthrough — FACET: Grounding Instruction, Environment, Solution and Verifier in One Executable State ★ MEMBER arXiv:2608.18580
186. Flow Matching from Scratch — What Came After Diffusion, and Why It Goes Straight ★ MEMBER arXiv:2210.02747
187. The Rise and Fall of GANs — An Invention Trained by Rivalry, and Why Diffusion Won FREE arXiv:1406.2661
188. Learning Rate Schedules — Why Warmup and Why Cosine FREE arXiv:1706.03762
189. Mixed Precision Training — Going Faster in fp16/bf16/fp8 Without Breaking ★ MEMBER arXiv:1710.03740
190. A History of Normalization Layers — From BatchNorm to RMSNorm ★ MEMBER arXiv:1502.03167
191. Paper Walkthrough: OmniScientist — An AI Scientist That Actually Looks at the Raw Data ★ MEMBER arXiv:2608.13558
192. Recommenders and Embeddings — Same Math as RAG, Different Goal FREE arXiv:1205.2618
193. CFG and Samplers — What the "Strength" Knob in Generative AI Really Does ★ MEMBER arXiv:2207.12598
194. Paper Explainer: SemaPLC — The Agent That Isn't Allowed to Say "Done" ★ MEMBER "arXiv:2608.18565
195. Do Transformers Actually Work on Time Series? — The Argument and the Practical Answer FREE arXiv:2205.13504
196. Tokenizers from Scratch — The Unit an LLM Cuts the World Into FREE arXiv:1508.07909
197. Paper walkthrough: Zetta ζ — a robot harness that repairs itself mid-execution, with the policy frozen ★ MEMBER arXiv:2608.16590
198. Paper Walkthrough: SA-MRPO — Stop Studying the Subject You've Already Aced ★ MEMBER "arXiv:2608.16072
199. Paper Walkthrough: Can Anything Catch a Fake Crisis Video? — What RA-Bench Found ★ MEMBER "arXiv:2608.14391
200. Paper Deep-Dive: ABSeeker — Training Long-Horizon Search Agents by Grading Each Step Backward from the Answer ★ MEMBER arXiv:2608.05102
201. The CNN Family Tree — From AlexNet to ResNet and EfficientNet ★ MEMBER arXiv:1409.1556
202. Paper Explained: Co-Evolution in Agentic Systems — Three Stages Toward Self-Directed Evolution FREE arXiv:2608.10299
203. Paper Walkthrough: ComBodied Agents — Moving an Agent's Target from Software and Matter to the Person FREE arXiv:2608.10915
204. Embeddings from Scratch — from word2vec Intuition to Contextual Embeddings FREE arXiv:1301.3781
205. End-to-End Driving from Scratch — Perception to Control in a Single Network FREE arXiv:1604.07316
206. Paper Deep-Dive: Frontis-MA1 — Training the AI That Builds AI: One Step Toward Recursive Self-Improvement in ML Engineering FREE arXiv:2607.28568
207. Paper Walkthrough: Interpretable MEG Decoding of Perceived Speech — Reading the Decoder's Weights as a Brain Map ★ MEMBER arXiv:2608.01481
208. Paper Explained: LongHorizon-Harness — Long-Horizon Agent Tasks Are a State-Management Problem, Not an Execution Problem FREE arXiv:2608.01964
209. Paper Walkthrough: Mental World Modeling — A World Model That Advances Minds, Not Just Physics FREE arXiv:2607.27201
210. Paper Walkthrough: MerchantBench — Can an LLM Agent Run an Online Store for a Year? Why It Earns Only 27.3% of What Humans Do FREE arXiv:2607.28956
211. Paper Walkthrough: Metis — A 'Memory Foundation Model' That Moves Agent Memory Inside the Model FREE arXiv:2607.26760
212. Object Detection from Scratch (from YOLO to DETR) ★ MEMBER arXiv:2005.12872
213. Paper Walkthrough: No Gold Answers, No Stronger Teacher — How u-OPSD Distills From Its Own Majority Vote FREE arXiv:2608.06296
214. Paper Explained: OSReward — Can You Trust the AI That Grades AI? Remeasuring Rewards for Computer-Use Agents FREE arXiv:2607.28609
215. Paper Walkthrough: PhiZero — A World Model That Reasons in a Language of Physics Before It Renders ★ MEMBER arXiv:2607.28624
216. Paper Walkthrough: Qwen-UI-Agent — How Alibaba Built a GUI Agent That Works on Real Phones and PCs ★ MEMBER arXiv:2607.28227
217. Paper Deep-Dive: Recursive Synthesis — Extending Verified Tasks Into 40,000 Long-Horizon Terminal Problems ★ MEMBER arXiv:2608.05466
218. Scaling Laws from Scratch — Why Making Models Bigger Makes Them Smarter (and When It Doesn't) FREE arXiv:2203.15556
219. Paper Walkthrough: SwanTale — Designing Voices from Words Alone, with Speech and Sound in One Waveform ★ MEMBER arXiv:2608.02023
220. Paper Walkthrough: ToolArtist — Search, Draw, or Redraw? The Image Agent That Decides for Itself ★ MEMBER arXiv:2608.04436
221. Paper Walkthrough: TurboVLA — Kick the LLM Out of the Loop and Run a Robot Policy at 32 Hz on an RTX 4090 with Under 1 GB of VRAM ★ MEMBER arXiv:2607.27205
222. Paper Explained: Video-DeepResearch — Agents That Watch a Video, Then Chase Down Every Lead ★ MEMBER arXiv:2608.03979
223. How VLMs Came Together — Wiring a Vision Encoder into an LLM ★ MEMBER arXiv:2103.00020
224. Paper Deep-Dive: AgentOPSD — Finding the Turn That Won the Game with Recursive Bayesian Belief Updates FREE arXiv:2608.05987
225. Paper Walkthrough: Alpamayo — NVIDIA's Reasoning Model for Autonomous Driving FREE arXiv:2511.00088
226. Paper Deep Dive: AskChem — Changing the Unit of Search from Papers to Provenance-Carrying Claims ★ MEMBER arXiv:2607.28618
227. Paper explained: BDH-CQ — an AI that thinks without words. Recurrent memory plus latent reasoning resets ARC's cost frontier FREE arXiv:2608.09888
228. BEV Representations From Scratch — Fusing Multiple Cameras Into One Top-Down Map ★ MEMBER "arXiv:2008.05711
229. Paper Walkthrough: CodeNib — A Multi-View Data System That Serves Repository Context to Coding Agents FREE arXiv:2607.25431
230. Paper Walkthrough: DAPD — Breaking the Teacher's "Cheat-Sheet Illusion" in Distillation with Dual Anchors FREE arXiv:2608.01735
231. Paper Walkthrough: EnvACE — Agents That Rehearse the World Instead of Calling It FREE arXiv:2608.06197
232. The Gemma Family from Scratch — Lineage, Inventions, and Where It Fits FREE arXiv:2403.08295
233. Paper Walkthrough: Macaron-V1 — A Frozen Base plus a Mixture of LoRAs, Built to Keep Learning After Launch FREE arXiv:2608.09819
234. Mixture of Experts (MoE) from Scratch — Routing and Load Balancing in the Switch Transformer FREE arXiv:2101.03961
235. Speculative Decoding from Scratch — How a Tiny Draft Model Speeds Up an LLM Without Changing a Single Output FREE arXiv:2211.17192
236. Paper Walkthrough: SWE-Bench ProMax — Measuring What Coding Agents Can Really Do with Large-Scale, Multilingual Refactoring FREE arXiv:2608.09802
237. Paper Walkthrough: The Personalization Mirage — LLMs Invent a Version of You, and Their Self-Reports Point the Wrong Way FREE "arXiv:2608.04570
238. Paper Walkthrough: WorldClaw — Agents That Build Walkable, Editable 3D Open Worlds from a Single Sentence ★ MEMBER arXiv:2608.05248
239. Paper Deep Dive — CLIP: Putting Words and Images on One Map FREE arXiv:2103.00020
240. Diffusion Models from the Ground Up — Add Noise, Then Subtract It FREE arXiv:2006.11239
241. Instruction Tuning and RLHF from Scratch — How a Model Learns to Follow Orders FREE arXiv:2203.02155
242. LLM Agents from Scratch — Designing the Tool-Use Loop FREE arXiv:2210.03629
243. The Science of Prompt Engineering — What Is Proven and What Is Folklore FREE arXiv:2201.11903
244. Paper Deep-Dive: Why Whisper Is Robust — Large-Scale Weak Supervision FREE arXiv:2212.04356
245. Paper Deep Dive — ViT: Treating an Image Like a Sentence FREE arXiv:2010.11929
246. Positional Encoding from Scratch — From Absolute Positions to RoPE FREE arXiv:1706.03762
247. Paper Deep Dive — Attention Is All You Need: What Dropping Recurrence Actually Proved FREE arXiv:1706.03762
248. LoRA Paper Explained: Low-Rank Adaptation of Large Language Models (Hu et al., 2021) — Why Low Rank Is Enough FREE arXiv:2106.09685