Peter Evans-Greenwood
banner
peter.evans-greenwood.com
Peter Evans-Greenwood
@peter.evans-greenwood.com
Thinking out loud about hybrid agency, crooked paths & coordination without stories. Essays → thepuzzleanditspieces.substack.com
What Happens in the Gap
The material process of technological transformation—why it takes decades and what actually has to happen during that time.
thepuzzleanditspieces.substack.com/p/what-happe...
What Happens in the Gap
1875: Whiskey fire in Dublin. 1881: 23 textile mills burn from electrical wiring. 1920: Productivity boom. Between deployment and transformation lies a material process that takes decades. Here's the ...
thepuzzleanditspieces.substack.com
January 28, 2026 at 12:27 AM
LLMs Are Following the Expert Systems Playbook—But the Score Is Different
Why the 56% Wage Premium Goes to Navigators, Not Builders
LLMs Are Following the Expert Systems Playbook—But the Score Is Different
Your AI assistant spent three weeks helping plan your cloud migration. Then your credentials appeared on the dark web. The forensics revealed something strange: no single conversation showed the breac...
open.substack.com
January 20, 2026 at 12:28 AM
The Electrification Productivity Puzzle
Why it took thirty years for the lights to turn on for the global economy
thepuzzleanditspieces.substack.com/p/the-electr...
The Electrification Productivity Puzzle
Electric motors appeared in factories throughout the 1890s. Productivity didn't boom until the 1920s. The delay wasn't organisational learning—it was the 30 years it took to build the commercial infra...
thepuzzleanditspieces.substack.com
January 13, 2026 at 12:52 AM
The post's strength is how it grounds speculation in deployment realities: human support costs, safety regulations, capital requirements, and the basic physics of falling robots. It's engineering judgment applied to hype cycles.
January 8, 2026 at 10:15 AM
6. Quantum computing nuance: He distinguishes digital (error-prone, oversold) from analog quantum computers that actually model physical systems—arguing the latter will be the real winners.
January 8, 2026 at 10:15 AM
5. LLM boxing: His most important insight might be that the action isn't in bigger models but in the "boxing in" mechanisms—guardrails, explainability, evaluation systems. He sees the arms race shifting to control systems, not raw capability.
January 8, 2026 at 10:15 AM
4. EV adoption miss: He correctly notes that even his conservative 30%-by-2027 prediction now looks wildly optimistic—Q3 2025 hit only 10.5% US market share, and that was front-loaded before tax credits expired.
January 8, 2026 at 10:15 AM
3. Realism on humanoids: His takedown is devastating—the conjunction of 12 false premises that would need to ALL be true for current hype to materialise. Safety issues, pathetic dexterity, teleoperation dependencies, and absurd scaling claims are all meticulously documented.
January 8, 2026 at 10:15 AM
2. Definition-drift: He nails how "self-driving cars," "humanoid robots," & "flying cars" have all quietly had their definitions watered down to match what's achievable rather than original promises. Waymo's "remote operators" reveal the gap between marketed autonomy and reality.
January 8, 2026 at 10:15 AM
1. Measured self-assessment: He acknowledges being "a little too optimistic" overall, which is notable since many dismissed his 2018 predictions as overly pessimistic. This intellectual honesty is rare.
January 8, 2026 at 10:15 AM
Predictions Scorecard, 2026 January 01
"Nothing is ever as good as it first seems and nothing is ever as bad as it first seems."
buff.ly/LXrqKYY

A few key themes that stand.
Predictions Scorecard, 2026 January 01 – Rodney Brooks
Robots, AI, and other stuff
buff.ly
January 8, 2026 at 10:15 AM