Goodhart’s Law: “When a measure becomes a target, it ceases to be a good measure.“
AI usage is a terrible metric. Using AI for what? We end up with initiatives that are effectively “AI-washing”. Investment in AI projects has to be aligned with company goals, not vanity metrics.
Gall’s Law: “Complex systems that work invariably evolve from simple systems that worked.”
Most AI projects fail because companies try to implement complex end to end AI solutions instead of focusing on narrow, well-defined and measurable problems. It is possible to build complex AI systems, but their success is predicated on simple foundations.
Jevons’ Paradox: “Technological progress that increases efficiency tends to increase rather than decrease total consumption”
Inference will become cheaper, on-device models will become more capable. It makes sense to assume broad, cheap, and widely available AI capabilities when planning for the next 3-5 years.
Amara’s Law: “We overestimate technology’s impact on the short term and underestimate it in the long term”
Gartner is already stating AI is in the “Trough of Disillusionment.” Studies claiming 95% of AI projects fail go viral . However, we are less than 3 years out from when ChatGPT first went live. We may never get to AGI, but imagine showing Claude Code to a developer in 2020…
Sagan’s Standard: “Extraordinary claims require extraordinary evidence”
It’s worth questioning the motives of leaders who claim the AI-mediated collapse of the knowledge economy is coming. Or those that welcome a “Gentle Singularity”, or perhaps warn of imminent mass extinction. These claims are often presented as quasi-religious arguments with scant evidence.
Bonus – The Lindy Effect: “The longer something has survived, the longer it will continue to survive”
Pattern matching, networking, and mentoring were how successful careers were made since the time the wheel was cutting edge technology. Yes, AI is amazing, but technology is transient, soft-skills endure..