Q001 — ACH Matrix¶


Research	R0041 — Enterprise Sycophancy
Run	2026-04-01
Query	Q001

Matrix¶

	H1: Enterprise products exist	H2: Research progress, no products	H3: No meaningful progress
SRC01-E01: OpenAI postmortem, general fixes only	--	++	-
SRC01-E02: Lambert says RLHF sycophancy unsolvable	--	+	+
SRC02-E01: Anthropic 70-85% reduction, no enterprise features	-	++	--
SRC04-E01: Bloom eval tool across 16 models	-	++	--
SRC05-E01: Constitutional framework, not product	-	+	-
SRC06-E01: Gemini 3 reduction, independent benchmark confirms	-	++	--
SRC07-E01: Multiple independent benchmarks emerging	-	+	--
SRC03-E01: Sycophancy inherent to RLHF	--	+	N/A

Legend:

++ Strongly supports
+ Supports
-- Strongly contradicts
- Contradicts
N/A Not applicable to this hypothesis

Diagnosticity Analysis¶

Most Diagnostic Evidence¶

Evidence	Why Diagnostic
SRC01-E01	OpenAI's postmortem is the most detailed vendor disclosure on sycophancy, and the fixes being general (not enterprise) strongly discriminates H1 from H2
SRC06-E01	Independent benchmark confirmation of Google's claims discriminates H2 from H3

Least Diagnostic Evidence¶

Evidence	Why Non-Diagnostic
SRC05-E01	Constitutional framework provides weak discrimination -- it is philosophical context rather than evidence for or against specific hypotheses

Outcome¶

Hypothesis supported: H2 — All evidence consistently shows active vendor research and measurable progress but zero enterprise-differentiated products

Hypotheses eliminated: H1 — No evidence of any enterprise-specific product, API parameter, or configuration across any vendor; H3 — Independent benchmarks and detailed technical work demonstrate genuine progress, not just marketing

Hypotheses inconclusive: None