clarity about problem

Loading complete world state into context is how you operationalize 'clarity about what problem you're solving'

MCP requires explicit specification of tool capabilities, parameters, and data schemas—forcing clarity about what context/information each tool needs.

CLAUDE.md as a mechanism for achieving clarity on organizational requirements and workflows

Requiring users to set explicit completion conditions forces clarity about what problem they're solving—directly supporting the thesis that clarity is the bottleneck.

The technique directly implements 'clarity' by making implicit assumptions explicit. Enumerating 'must be true' beliefs is the operational definition of clarity.

Kerr's observation that testing burden increased to 99% implies teams must define success criteria explicitly before agents execute. This is clarity about the problem made concrete.

High token waste suggests organizations lack clarity about which context is actually relevant for each use case

Agents that want to add lack clarity about the core problem. Systems that 'push back' and remove are expressing clarity about scope.

Tweet's claim that 'good instructions have never been higher' directly validates the thesis that problem clarity is the bottleneck, not model quality

Implicit lesson: teams that didn't have clarity about which agent pattern solves their problem defaulted to LangChain without evaluating fitness. Clarity about agent pattern needed to come first.

Tweet acknowledges the problem (having tool without clear questions) but doesn't validate that clarity is THE bottleneck—it just jokes about it. No evidence provided that clarity would solve the state