Skip to content

R0023/2026-03-25/Q002/H2

Statement

Most widely circulated prompt engineering advice comes from people without formal AI research training — marketers, content creators, or documentation writers — explaining the gap between popular advice and empirical evidence.

Status

Current: Partially supported

While the most influential original guides are researcher-authored, the vast majority of prompt engineering content circulating on the internet is produced by marketers, content creators, and social media influencers who lack research backgrounds. Vendor documentation (OpenAI, Google, Anthropic) is produced by documentation and product teams, not research teams. The sheer volume of derivative content from non-researchers overwhelms the researcher-authored originals.

Supporting Evidence

Evidence Summary
SRC03-E01 OpenAI guide: no individual authors, organizational product with commercial incentives
SRC04-E01 Anthropic guide: documentation writers, not clearly researchers

Contradicting Evidence

Evidence Summary
SRC01-E01 The most popular independent guide is researcher-authored
SRC02-E01 The second most popular guide is researcher-authored

Reasoning

H2 is partially true at the volume level (most content) but wrong at the influence level (the most cited guides). The picture is more nuanced than H2 suggests.

Relationship to Other Hypotheses

H2 captures the content creator / marketer layer that H3 describes as the secondary distribution mechanism. It is not wrong about the existence of this layer — it is wrong about which layer is most influential.