Progress follows observation and tinkering, not abstract reasoning first; the “linear theory of innovation” and waterfall model are both forms of “thinkism.”
Key Takeaways
“Thinkism” (Kevin Kelly’s term) is the belief that hard thinking alone solves problems; it works in school and bureaucracy but fails in R&D.
Historical example: the pendulum clock (1656) preceded Hooke and Newton formalizing mechanics (1660-1666), not the reverse.
Waterfall software design and UML-first “analyst” roles are institutional thinkism; real engineering requires incomplete-understanding-first iteration.
Recipe for breakthroughs: maximize observation and experimentation, minimize upfront abstract reasoning.
The AI-will-solve-everything argument is a form of thinkism; no matter how much an AI reads, its knowledge remains insufficient for novel discovery.
Hacker News Comment Review
Commenters pushed back on the school exam analogy: at work, engineers can run experiments and do research, making the comparison to timed exams structurally unfair.
Several commenters found “thinkism” a genuinely useful new term; one noted it reframes AGI doom scenarios (“AGI thinks 10,000x faster”) as thinkism by another name.
The repair/debugging framing resonated: fixing broken systems forces tight feedback loops between mental models and physical reality, which is the opposite of thinkism.
Notable Comments
@joshuahedlund: argues AGI singularity fears are thinkism – assumes raw thinking speed equals problem-solving power.
@JSR_FDED: repairing physical things as the ideal hybrid of experiential learning and model-building, feeding each into the other.
