Detecting coordinated financial narratives with embeddings and AVX2
Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.
Want to reach these readers? Sponsor an issue.
I built an open-source system called Horaculo that analyzes coordination and divergence across financial news sources. The goal is to quantify narrative alignment, entropy shifts, and historical source reliability. Pipeline Fetch 50–100 articles (NewsAPI) Extract claims (NLP preprocessing) Generate sentence embeddings (HuggingFace) Compute cosine similarity in C++ (AVX2 + INT8 quantization) Cluster narratives Compute entropy + coordination metrics Weight results using historical source credibility Output structured JSON signals Example Output (query: “oil”) Json Copiar código { "verdict": { "winner_source": "Reuters", "intensity": 0.85, "entropy": 1.92 }, "psychology": { "mood": "Fear", "is_trap": true, "coordination_score": 0.72 } } What it measures Intensity → narrative divergence Entropy → informational disorder Coordination score → cross-source alignment Credibility weighting → historical consensus accuracy per source Performance 1.4s per query (~10 sources) ~100 queries/min ~150MB memory footprint Python-only version was ~12s C++ optimizations: INT8 embedding quantization (4x size reduction) AVX2 SIMD vectorized cosine similarity PyBind11 integration layer Storage SQLite (local memory) Optional Postgres Each source builds a rolling credibility profile: Json Copiar código { "source": "Reuters", "total_scans": 342, "consensus_hits": 289, "credibility": 0.85 } Open Source (MIT) GitHub: [https://github.com/ANTONIO34346/HORACULO] I'm particularly interested in feedback on: The entropy modeling approach Coordination detection methodology Whether FAISS would be a better fit than the current SIMD engine Scalability strategies for 100k+ embeddings
HN · launched Feb 12, 2026 · 5 upvotes (top 89%) · 0 views
Niche: stock and market analysis software · 366 launches, 44 in the last 7 days ▲ +7% · Explore this niche →
Category and tags
- Fintech
- Analytics & BI
- B2B
- AI-native
- Hobby / open-source project
More like this: all 366 launches in stock and market analysis software →
Could you build this?
Partial Fetching news and calling embedding APIs is simple, but building high-throughput AVX2-accelerated similarity matrix computation and specialized narrative entropy algorithms requires low-level optimization.
What it would actually take: The system architecture pairs an ingestion pipeline (Python/NewsAPI) with a low-level compute engine (C/C++ or Rust utilizing SIMD/AVX2 intrinsics) for ultra-fast pairwise vector distance and clustering. Hard parts include implementing vectorized dot products, numerical stability in entropy calculations, and real-time narrative drift detection across temporal graphs. Requires systems programming skills alongside applied NLP/information theory domain knowledge.
Discussion
1 comment analyzed.
Competitors
Other products that read as similar to this one : 174 launches clear the similarity bar, closest 8 shown.
Attention rank: #150 of 175 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).
Launched 99 days after the earliest competitor.
- Rhetoric Audit · ph · 2026-10-01 · 2 upvotes · similarity 0.44
- Kanon 2 Enricher · hn · 2026-03-03 · 10 upvotes · similarity 0.43
- Project AELLA · hn · 2025-11-11 · 6 upvotes · similarity 0.42
- We fingerprinted 178 AI models' writing styles and similarity clusters · hn · 2026-04-08 · 78 upvotes · similarity 0.41
- MAGA or Not? Political alignment scores for people and companies · hn · 2026-03-22 · 7 upvotes · similarity 0.40
- genpark-complex-financial-formula-audit-validator-skill · github · 2026-09-27 · 7 upvotes · similarity 0.40
- Irpapers · hn · 2026-02-23 · 5 upvotes · similarity 0.40
- Sup AI, a confidence-weighted ensemble (52.15% on Humanity's Last Exam) · hn · 2026-03-26 · 26 upvotes · similarity 0.39
The niche, labels, vibe code rating and competitor comparison on this page are analysis by Nicheloom. If you use them, please credit nicheloom.com. How you may use our data.