Bricks And Bytes We Need a Moonshot" - Why AI in Construction Is Focused on the Wrong Problem
4 snips
Mar 24, 2026 Matt Gough, an experienced construction and innovation leader who advised NEOM and led programmes at MACE, argues that AI is being applied to the wrong problems in construction. He discusses why the industry's delivery model blocks transformation. Short, punchy takes cover virtual vertical integration, misaligned incentives across the supply chain, and what a true moonshot for construction might look like.
AI Snips
Chapters
Transcript
Episode notes
Empire State Building As AMoonshot Example
- Matt contrasts the 1930 Empire State Building delivery with modern practice: rapid on‑site coordination achieved without today's tech.
- He uses that story to ask why, despite more tools, productivity hasn't tracked the rise in capability.
Cost Of Intelligence Fell But Use Is Incremental
- The cost of intelligence is collapsing so AI enables huge information work in construction, but firms mostly use it to make existing tasks faster.
- Matt Gough notes GPT-era inference costs fell ~280x (Nov 2022–end 2024), yet adoption focuses on takeoffs and reporting, not rethinking delivery.
Build AI Muscle By Experimenting Now
- Increase AI budgets and build AI skills now; treat adoption like training at the gym to develop muscle.
- Matt recommends 'get in the arena': experiment, build internal muscle, and learn rather than waiting for perfect governance.