---
id: "contrarian-marginal-improvements-invisible"
type: "contrarian-insight"
source_timestamps: ["§ The U.S. Strategy: Compete on Capability"]
tags: ["contrarian-insight", "rd-strategy", "consumer-behavior"]
related: ["concept-capability-competition"]
challenges: "The belief that superior technical benchmarks and model capabilities are the primary drivers of consumer adoption and market dominance."
sources: ["attention"]
sourceVaultSlug: "hbr-seg-attention"
originDay: 4
articleStem: "hbr-tier2-07-chinese-ai-firms-habits"
sourceUrl: "https://hbr.org/2026/06/lessons-from-chinese-ai-firms-on-owning-customers-habits"
sourceTitle: "Lessons from Chinese AI Firms on Owning Customers’ Habits"
---
# Contrarian: Marginal AI capability improvements are invisible to consumers

## Contrarian Insight — Marginal capability is invisible to consumers

**Conventional view (challenged):** Pushing the frontier of model capabilities (e.g., improving benchmark accuracy from **90% to 93%**) is the key to winning market share.

**The authors' counter:** Once models cross a **"good enough" threshold** — competently answering questions, writing emails, generating code — further marginal improvements become **entirely invisible to ordinary consumers**. The consumer's choice of ChatGPT vs. Google is driven by **habit and friction**, not a 3% difference in benchmark accuracy.

This is the psychological engine behind the critique of [[concept-capability-competition]] and the case for the [[concept-habit-moat]].

**Enrichment / counter-perspective:** The claim is context-dependent. In **enterprise/professional settings** (legal, medical, financial), marginal accuracy, reliability, and safety gains *do* matter and *are* visible because errors are costly. A balanced reading is that capability and habit are **complementary** — strong capabilities earn the trust required before users will routinize AI, after which habit and friction dominate.