InLevel Up CodingbySlava Efimov·1d agoGraph Neural Networks: GCN, MPNN, and GAT, Explained SimplyA visual guide to how graph neural networks work under the hoodA response icon1
Dr. Theophano Mitsa·5d agoFrom TF-IDF to Transformers: How Machines Learned to Search by MeaningRule-based retrieval, classical ML, embeddings, and fine-tuned transformers in PythonA response icon4
Chaim Rand·Sep 8Overcoming the PyTorch Recompilation Dilemma in Real-Time Inference WorkloadsHow to Reduce the Runtime Interference of Recompilation SpikesA response icon2
InTowards Deep LearningbySumit Pandey·Sep 2A 30MB Rust Browser Fixed the Worst Part of My AI News PipelineA small AI news digest that runs itself, the one job it kept failing at, and the tiny open-source browser that finally fixed it.A response icon8
The NewGenCoder ✨·Sep 2California wants my 3D printer to scan every design for a gun.A bill that just cleared the state Assembly would put a government-approved snitch in the firmware. The mechanism isn’t new. We just never…A response icon41
InPredictbyNov Tech·Sep 8He Invented a Core Technology Behind ChatGPT. Now He’s Betting $1 Billion It’s a Dead End.Here’s why the rest of the industry should be listening.A response icon68
InTowards Deep LearningbySumit Pandey·Sep 9I Asked GPT-6 and Fable 5.1 to Build an End to End Software, Outcome Shocked MeOne of them won the build and lost the analysis, and daily token limits shaped the outcome more than any benchmarks could.A response icon19
InData Science CollectivebyAnubhav·Sep 8One Command Tells You What Local LLM Your Computer Can Run in 2026It’s right about what fits in memory. The tokens per second beside it claims a 30% error band, and the only test guarding that number…A response icon10
InTowards Deep LearningbySumit Pandey·May 8A Single CLAUDE.md File Went Viral. The Reason Is Embarrassingly Simple.91,000 stars on GitHub. No code. Four rules from Andrej Karpathy that every coding agent should have been following from day one.A response icon115
InTowards Deep LearningbySumit Pandey·Jul 27Andrej Karpathy Asked for a Tool. 48 Hours Later, Graphify Went Viral.It’s called Graphify, and it turns any folder of code, docs, and images into a queryable knowledge graph: no vector database, no config…A response icon37
YaLinChen (Amber)·Just nowWhat does it mean when the model loss goes down?If you are training machine learning models, a loss curve is something you always track. The curve plots the model loss as a function of…
Charryl the Empath anempathsheart.com·Just nowTHE HAMSTER WHEEL OF OUTRAGE:The internet loves a hanging. BY Charryl the Empath 💜
Ritvik Rastogi·Just nowPapers Explained 620: Normalized Low-Rank Adaptation (NoRA)Normalized Low-Rank Adaptation (NoRA) is an improvement over standard LoRA that normalizes the down-projection matrix A along the rank…
sangavi·Just nowDay 9: Gradients — How a Model Actually “Learns” From Being WrongPart 9 of a 50-day journey from zero to building AI agents.
