Back to browse
GitHub Repository
3 stars

A closed source engine that stops hallucinations deterministically

by MattijsMoens·Feb 26, 2026·2 points·3 comments

AI Analysis

Pass

Marketing-heavy claims with zero auditable proof, no code, no reproducible benchmarks.

Strengths
  • Addresses real problem: RLHF/filtering is statistically fallible and provably bypassable.
Weaknesses
  • 'Deterministic' AI safety is contradictory—LLMs are inherently probabilistic; rebranding doesn't change physics.
  • 30-min Loom 'proof' is anecdotal theater; no peer review, no open methodology, no reproducible attack suite.
  • Closed source + extraordinary claims = impossible to verify; smells like fundraising narrative.
Category
Target Audience

AI safety researchers, enterprise security teams

Similar Projects

AI/ML●●Solid

Kremis – Deterministic memory graph for AI agents (Rust)

Deterministic graph memory vs. embeddings is clever, but fragmentation of AI memory ecosystems is already painful.

Big BrainWizardry
M2Dr3g0n
104mo ago