R0023/2026-03-25/Q002/H1¶
Statement¶
Most widely cited prompt engineering guides are written by researchers with relevant AI/NLP credentials, and the advice reflects empirical evidence.
Status¶
Current: Partially supported
The two most influential independent guides (promptingguide.ai by Saravia, LearnPrompting by Schulhoff) are indeed written by people with legitimate NLP research backgrounds. However, vendor guides (OpenAI, Google) do not identify individual authors, making credential verification impossible. The broader ecosystem of guides and tutorials is dominated by content creators and marketers who lack research backgrounds.
Supporting Evidence¶
| Evidence | Summary |
|---|---|
| SRC01-E01 | Saravia: PhD in NLP, co-creator of Galactica, Meta AI veteran |
| SRC02-E01 | Schulhoff: UMD NLP researcher, lead author of The Prompt Report |
Contradicting Evidence¶
| Evidence | Summary |
|---|---|
| SRC03-E01 | OpenAI guide has no individual authors — credential verification impossible |
| SRC04-E01 | Anthropic guide partially attributed but authors' research credentials unclear |
Reasoning¶
H1 is true for the top two independent guides but not for vendor guides. The distinction between "researcher-authored" and "empirically validated" is important: both Saravia and Schulhoff are researchers, but their guides were not originally published as peer-reviewed research. The Prompt Report (Schulhoff) is the exception — it is a systematic survey.
Relationship to Other Hypotheses¶
H1 captures one layer of the picture (the top guides), while H3 captures the complete landscape (researcher origins distorted by commercial and content ecosystems).