Tools for thought should be evaluated in the context of intrinsically meaningful purposes

We don’t evaluate Mathematica by asking how many students use it, by asking about its impact on test scores, or by measuring how much faster it helps people solve problems. Those are all relevant proxies, to varying degrees, but they’re not the tool’s actual purpose. Mathematica was created as part of Wolfram’s research into symbolic manipulation and automata, then eventually expanded to empower mathematicians and scientists in general. Its utility function is, roughly: how many powerful insights does it enable on the margin?

Relatedly, it’s tempting to evaluate the Mnemonic medium by asking how much people remember after reading a text. But its real purpose is to enable people to do more of whatever they find most meaningful, to help them become more themselves. We’ll need to find some way to evaluate that metric.

If we want to offer some kind of secular Transcendental narrative, we must ultimately evaluate all our work relative to that—relative to giving meaning, relating people to the world, organizing how people live, etc. Sure, we’ll need proxy metrics, but that’s all they’ll be. A focus on this purpose erases distinctions between “content” and “tools” and pushes us to pursue ideas that will better deliver meaning.

Better framings include:

In this sense, “Tools for thought” is a somewhat distorting term. It pushes us to think about “thought” as the end-goal. We’re really trying to construct “contexts for meaning.” Note too that a “context” could be a plain old-fashioned essay!


  • What is the highest-growth environment for individuals and groups? What does this project look like, conceived of as an (extremely unusual) online class?
  • Is it possible to help create super-empowered subversive individuals who believe in their own ability to have a meaningful impact on important problems?
  • What do MOOCs really want to be? + How to reinvent Universities? What does University++ look like? + What does the Primer++ or Dynabook++ look like? + How many Sistine Chapel Ceilings per lifetime?

Kay, A. C. (1972). A Personal Computer for Children of All Ages. Proceedings of the ACM Annual Conference - Volume 1.