# Beat Map - Signal B
## Source: lean-ai.pages.dev - Chapter 1: Intelligence Brief
## Extracted: 2026-06-03

---

Talk Track B · This Week · The Manifesto Post

Naming the Category — Maximum-by-Default vs. Frugal AI

LinkedInLong-form · 90sSeries Anchor

"The enterprise AI market has a Maximum-by-Default problem. We're naming it. Naming it means owning the category that solves it."

Four independent research teams published quantization breakthroughs in the last 60 days. Every single one reached the same conclusion: **you don't need the biggest model. You need the right-sized model.**

Meanwhile, enterprise AI buyers are still defaulting to prestige model selection. GPT-4 on intake forms. Frontier models on structured data extraction that a fine-tuned 7B model handles at 1/20th the cost. Maximum capability on minimum-complexity tasks.

This is **Maximum-by-Default** — the prevailing mental model in enterprise AI. The assumption that deploying the most powerful available model is always the sophistication signal. It's the same mistake enterprise made with cloud: over-provisioned, under-optimized, and paying for capacity they never use.

**Frugal AI is the new game.** Not cost-cutting AI. Not "cheap AI." The discipline of deploying exactly the AI capability a workflow requires — no more, no less. Every deployment decision tied to cost per correct output. Every architecture designed for the intersection of determinism and efficiency. Every model selection made as an economic decision, not an optics decision.

We're building the practice. We're writing the framework. We're claiming the category. **And we're doing it before Gartner names it, before the Big 4 builds a slide deck about it, and before the vendors decide they invented it.**

→ Series: "The Frugal AI Stack" (Week 2) · "Right-Sizing Your AI Architecture" (Week 3) · "AI Economics: The CFO's Framework" (Week 4)

REL 94 · Score 80

Mentat · 175d