In 2016, a music critic described Tool’s decade without a new album as “a sick and twisted PR campaign, but a damn effective one.” The band released nothing. No new material — maybe a one-off show. The critic’s explanation: “While demand for new music is high, Tool are savvy enough not to dilute their brand.”
Nine years later, JetBrains researchers published a study on why AI coding agents choke on their own context. The diagnosis: “agent-generated context actually quickly turns into noise instead of being useful information.” The fix they propose is compression — stripping accumulated material back to what carries weight.
The vocabulary differs: “brand dilution” in one register, “suboptimal return on investment per token” in the other. The structural claim is the same — a system that keeps adding to its own output gets buried by it. The intervention that works is subtraction.
The nine-year gap is diagnostic. In 2016, a music journalist noticed that one band’s stubbornness happened to work and called it hype. By 2025, the same observation had become a research problem with empirical studies and a proposed hybrid architecture. JetBrains called it efficiency.
Tool never published a paper on strategic absence. The researchers probably didn’t read Consequence of Sound. The pattern just arrived in each domain on its own schedule, wearing the local uniform.