TechCentral
Read post

Why Discovery is going slow on AI coding agents

Discovery Group CIO Derek Wilcocks shares a measured take on AI coding agents in large regulated enterprises. He reports 20-25% efficiency gains from AI coding assistants, but notes developers only spend 25-30% of their time writing code, making the real-world impact smaller than headline numbers suggest. Three unsolved problems block fully agentic coding: explainability, efficiency, and maintainability of AI-generated code. Discovery constrains its agentic tools heavily to meet coding standards, security, and privacy requirements. Wilcocks argues the future winners won't be those with the best AI models, but those who build the scaffolding to deploy AI safely at scale. On hiring, he says coding alone is no longer sufficient — computer science fundamentals, communication, and continuous learning matter more. Notably, 50-60% of Discovery's IT staff completed at least one AWS or Google training programme when given free access, far exceeding the expected 3-4%.

    #productivity#ai-coding#ai-governance
Today•4m read time•From techcentral.co.za
Post cover image
Table of contents
Read: Discovery teams up with Google to launch AI-powered Vitality platform
34 Impressions
TechCentral's image
TechCentral

185 Followers

•

2.4K Upvotes

Would you recommend this post?

Copy link
WhatsApp
Facebook
X
New Squad
  • © 2026 Daily Dev Ltd.
  • Guidelines
  • Explore
  • Tags
  • Sources
  • Squads
  • Leaderboard