Tutorials & Guides
OpenAI Cookbook updates archived Codex recipes with security warnings and links to current CI/CD guidance.
A data scientist explains how they built a fraud detection system using multiple models and merged disparate datasets with schema validation, despite a simpler model performing better in tests.
A practical guide on effectively using AI coding agents by providing context, breaking down problems, and maintaining control over the development process.
Explains the mathematical foundation of the sigmoid function, connecting its relationship to Euler's number 'e' and practical usage in logistic regression and neural networks.
Sentence Transformers v6.0 introduces a new MultiVectorEncoder model type, with a detailed guide for finetuning or training from scratch for late-interaction/ColBERT-style retrieval.
Guide for serving Markdown content to AI agents via Accept: text/markdown headers to reduce token usage and improve RAG performance.
An accepted MATS 11.0 applicant shares their experience with the OpenAI safety team stream application process and advice for future candidates.
Explains how RAG rerankers actually work at the token level rather than just architectural level, showing they learn statistical token associations conditioned on query-passage pairs.
Explains the mathematical rationale behind Random Forest's feature subsampling, showing how it reduces correlated errors between trees beyond what bagging alone can achieve.
OpenAI explores how ChatGPT enables continuous learning beyond the classroom for both students and educators.
A methodology for efficiently handling 100+ daily coding tasks using Claude Code, addressing context management and task orchestration challenges.
Shows how to extract structured data from large collections of similar documents into SQL tables for efficient RAG queries.
A developer shares hard-won lessons about migrating a local data pipeline to AWS, revealing unexpected dependencies on 'everything running on one machine'.
Presents 10 controversial positions on enterprise RAG that challenge standard tutorials, arguing for structure-first retrieval, multi-modal approaches, and enterprise-specific considerations.
A developer used four AI models including GLM-5.3 and Claude to successfully root an Amazon Fire tablet that was automatically shutting down.
Codex hooks allow developers to inject custom logic at specific points in the agentic loop, enabling deeper customization beyond prompting.
Cheshire Academy educators are using chatbots like ChatGPT and Perplexity alongside specialized tools like MagicSchool to create lesson plans and grading rubrics while teaching students responsible AI use.
A beginner-friendly tutorial on survival analysis and the Cox proportional hazards model, with runnable Python code for practical implementation.
Author shares their agent.md file used to prompt AI code assistants like Claude, containing detailed coding style rules that dramatically improve LLM-generated code quality.
OpenAI added archive compliance logs platform quickstart to their cookbook repository for managing API logging and compliance.
A technical analysis explains why locally run LLMs underperform compared to reference implementations due to hardware differences, software stacks, and improper benchmarking methods.
Simon Willison launches a project to document Agentic Engineering Patterns for building software with coding agents like Claude Code and OpenAI Codex.
Google Search's AI Mode can visualize furniture in your room, and its Lens and Circle to Search features can identify vintage items and find similar decor products.
Proposes row-level chunking for RAG systems to retrieve specific table rows instead of entire tables, improving precision for tabular data queries.
Presents a multi-document RAG approach for unstructured PDF folders using nested outlines instead of traditional indexing when documents lack shared fields.
The author details replacing in-memory storage with PostgreSQL to create a robust, shareable backend for a LangGraph AI agent previously built for booking services.
A real estate company fine-tuned Google's SigLip model with LoRA to improve room classification in property listings, explaining when this approach is appropriate.
A comprehensive guide to fine-tuning LLMs with QLoRA, showing how a Mistral 7B model achieved 98% accuracy on complex medical reporting tasks.
A guide for building enterprise RAG systems that identifies three distinct document corpus shapes, explains why a generic vector store fails, and outlines tailored architectures for each type.
Simon Willison argues that effective coding agent use requires learning to instruct and verify changes beyond just line-by-line code review.