Clemence Ayekple·1h agoYour Animation Is Making Someone DizzySome of your visitors have already asked their device for less motion. Here’s how to hear them with prefers-reduced-motion, and what WCAG…
Clemence Ayekple·2h agoHow to Build a Data Science Portfolio That Gets NoticedThree to five real projects, presented so people actually read them
Clemence Ayekple·1d agoMuted, Blind or Deaf: Can Everyone Still Get Your Video?Captions, transcripts and audio description: what each one does, which ones WCAG requires, and how to add them to a video on your own site.
Clemence Ayekple·1d agoSquash the Belt into One Box: How reduce Turns an Array into a Single ValueThe mental model
Clemence Ayekple·1d agoWorse Together: Why One AI Agent per User Breaks Down on Shared ResourcesThe default vision of agentic AI is that everyone gets a personal agent. Anthropic’s newest alignment study tested that vision on shared…
Clemence Ayekple·1d agoFrom Analysis to PredictionWhere data science meets machine learning, with one worked example
Clemence Ayekple·3d agoSwap the Card, Change the Job: How Callbacks Tell a Function What to DoThe mental model
Clemence Ayekple·3d agoSeven Neurons, One Backdoor: Microsoft’s Lazarus Shows Model Weights Need a Supply ChainEdit seven neurons in a downloaded checkpoint, and an 8B coding agent will execute an attacker’s command on cue nearly every time. No…
Clemence Ayekple·3d agoLifeVal Is a Survey. Read It Like One.OpenAI’s new everyday-task benchmark is a welcome step away from trivia-style evals. But once you build a benchmark from population data…
Clemence Ayekple·3d agoYour Table Is Just a Grid of Boxes Until You Give It HeadersHow to mark up data tables so screen reader users hear “Monday, 13:00” and not just “MATH 205”.