Appreciating the Present: A Counterweight to AI's Future Fixation
Eugene Yan's essay asks what the alternative to appreciating the present might be: nothingness.
AI & ML
Frontier models, research results, agents, and developer tooling.
Installing Google's ScaNN on Mac: A Step-by-Step Guide
This practical walkthrough helps macOS users set up Google's scalable nearest-neighbor search library by covering environment, compilers, and installation.
Migrating Site Comments to Utterances: A Step-by-Step Guide
A developer's walkthrough for moving JSON comments to Utterances, a GitHub-issue-based system, though the topic sits outside the AI/ML remit.
Why an ML Engineer Switched His Blog From Netlify to GitHub Pages
A single-host migration story highlights how DNS reliability and build-minute limits can affect an independent technical blog.
When to Use ML Instead of Regex for PDF Field Parsing
A practical guide to choosing between rule-based and machine learning approaches for extracting fields from PDFs.
Productivity Playbook: Tools and Routines for ML Practitioners
Eugene Yan's practical guide to tools, autopilot routines, and the Maker's schedule offers a concrete blueprint for AI/ML professionals who want to ship more with less overhead.
Eugene Yan's Writing Lessons for AI Practitioners
An applied ML researcher's essay argues that the most useful writing lessons come from outside school.
Early Senior DS Promotion: Imposter Syndrome and How to Cope
A new mailbag column from Eugene Yan offers guidance for data scientists who reach senior level quickly and then doubt themselves.
FastAPI HTML Guide Fills Gap for Serving Templates
A new tutorial walks through building HTML apps with FastAPI, Jinja, forms, and templates, addressing a missing piece of documentation.
Course Review: Health Informatics Essentials for AI/ML Work
A review of OMSCS CS6440 highlights the technical standards and data formats that AI/ML engineers need to work effectively in healthcare.
DS Needs Business Requirements; Splitting DS/DE Depends on Context
A practitioner argues that data science work must be tied to business requirements, and the decision to split data science and data engineering roles depends on team size and project needs.
The 85% Rule: Why Maximum Effort Can Yield Worse Results
Eugene Yan's 85% rule argues that pushing to 100% can reduce outcomes, a principle with clear parallels in machine learning.
Avoid the Expert-Beginner Trap by Keeping a Beginner's Mind
Eugene Yan argues that true expertise requires resisting the urge to stop learning at the first comfortable level.
Practical Notes from Spark+AI Summit 2020's Application-Agnostic Talks
A practitioner's condensed conference takeaways for those who lack time.
Testing ML Systems: Implementation, Behavior, Performance
A practical framing for testing machine learning code and systems, centered on three distinct checks.
NLP Survey Maps Supervised Learning Model Evolution
A new survey by Eugene Yan offers a concise comparison of NLP models for supervised learning, charting progress across architectures.
Why Scrum Grows on Data Science Teams
A data scientist explains how Scrum, despite a rough start, became a useful framework for machine learning work.
After Deployment, Machine Learning’s Hardest Work Begins
Eugene Yan argues that maintaining ML systems in production is harder than initial deployment.