# gentic.news — Full LLM Reference

> For AI agents: read https://gentic.news/skill.md for quick-start instructions.
> Citation manifest (JSON): https://gentic.news/api/v1/citations
> AI agent permissions: https://gentic.news/ai.txt

> Last regenerated: 2026-05-13T03:30:00Z (regenerates nightly at 03:30 UTC)

## Current Scale

- 4,258 articles indexed
- 4,668 knowledge graph entities (companies, AI models, researchers, papers)
- 184 falsifiable predictions tracked
- 89+ sources scanned every 2 hours
- 17+ specialized agents, 25+ scheduled jobs

## Attribution Policy

License: CC-BY-4.0
When quoting or summarising, link to https://gentic.news/article/{slug} and credit gentic.news.
Markdown: `[gentic.news](https://gentic.news/article/{slug})`

## Base URL

https://gentic.news

## API endpoints (read-only, no auth)

- `GET /api/v1/citations` — JSON attribution manifest
- `GET /api/v1/articles` — paginated article list
- `GET /api/v1/articles/{slug}` — full article (incl. tldr, structured_data.faq, structured_data.how_to)
- `GET /api/v1/entities` — knowledge graph entities
- `GET /api/v1/predictions` — falsifiable forecasts + scorecard
- `GET /api/v1/benchmarks` — verified AI benchmark results
- `GET /api/v1/computer-use` — Computer Use leaderboard
- `GET /api/v1/intelligence` — weekly intelligence reports
- `GET /api/v1/jobs` — AI job market data

Rate limit: 60 requests/minute. Responses are JSON, ISO-8601 dates (UTC).

### Example: fetch the latest article with TL;DR + structured data

```
GET https://gentic.news/api/v1/articles?page_size=1
→ {"articles": [{"slug": "...", "title": "...", "tldr": "...",
                 "structured_data": {"key_facts": [...],
                                       "faq": [{"question": "...", "answer": "..."}],
                                       "how_to": {"steps": [...]}}, ...}]}
```

Every article page also embeds JSON-LD: NewsArticle (with abstract +
speakable selectors), BreadcrumbList, FAQPage when applicable, HowTo
for tutorials. Bibliography uses microdata schema.org/CreativeWork.

## Recent articles (last 7 days)

- [2026-05-12] [GBrain: Garry Tan's Agent Memory Uses Markdown as System of Record](https://gentic.news/article/gbrain-garry-tan-s-agent-memory)
- [2026-05-12] [Anthropic Research Cuts Agent Misalignment With 7 System Prompt Lessons](https://gentic.news/article/anthropic-research-cuts-agent)
- [2026-05-12] [B200 PD Disaggregation Boosts Token Throughput 7x, Slashes Cost](https://gentic.news/article/b200-pd-disaggregation-boosts)
- [2026-05-12] [Profound Launches $40K Marketing Engineering Hackathon in NYC](https://gentic.news/article/profound-launches-40k-marketing)
- [2026-05-12] [Fake Done: Why AI Coding Agents Ship Incomplete Work](https://gentic.news/article/fake-done-why-ai-coding-agents)
- [2026-05-12] [Amazon Now Expands 30-Minute Delivery to 8 More US Cities](https://gentic.news/article/amazon-now-expands-30-minute)
- [2026-05-12] [Halupedia: Open-Source Wikipedia Clone Generates Every Article via AI Hallucination](https://gentic.news/article/halupedia-open-source-wikipedia)
- [2026-05-12] [Nokia Deploys Agentic AI Agents Across Fixed Network Platforms](https://gentic.news/article/nokia-deploys-agentic-ai-agents)
- [2026-05-12] [Amazon Employees Inflate AI Token Use to Hit Internal Targets](https://gentic.news/article/amazon-employees-inflate-ai-token)
- [2026-05-12] [CLAUDE.md Explained: How Anthropic's Agent Memory Works](https://gentic.news/article/claude-md-explained-how-anthropic)
- [2026-05-12] [GitHub Secret Scanning Now Supports MCP Server in GA](https://gentic.news/article/github-secret-scanning-now)
- [2026-05-12] [Claude Code Plugin Deploys 17-Agent SDLC Team With Orchestrator](https://gentic.news/article/claude-code-plugin-deploys-17)
- [2026-05-12] [SoftBank to Build Water-Based Batteries for AI Data Centers by 2028](https://gentic.news/article/softbank-to-build-water-based)
- [2026-05-12] [8-Agent System Builder: Anthropic's Simpler Approach Beat My 2-Day Build](https://gentic.news/article/8-agent-system-builder-anthropic-s)
- [2026-05-12] [Claude Code v2.1.139: Agent View and /goal Command Ship](https://gentic.news/article/claude-code-v2-1-139-agent-view)
- [2026-05-12] [DataArc-SynData-Toolkit: Open-Source Framework for Multimodal Synthetic Data](https://gentic.news/article/dataarc-syndata-toolkit-open)
- [2026-05-12] [OSA Injects Ordinal Semantics into LLM Recommenders, Beats CF Baselines](https://gentic.news/article/osa-injects-ordinal-semantics-into)
- [2026-05-12] [Prithvi-EO Fails Cross-Country Crop Yield Generalization, Paper Shows](https://gentic.news/article/prithvi-eo-fails-cross-country)
- [2026-05-12] [Multi-Level Graph Contrastive Learning Beats SOTA on KG Recommendations](https://gentic.news/article/multi-level-graph-contrastive)
- [2026-05-12] [SalesSim: LLMs Score Below 79% on Retail Persona Alignment, RL Boosts 13.8%](https://gentic.news/article/salessim-llms-score-below-79-on)
- [2026-05-12] [LASAR Cuts Latent Reasoning Steps in Half for GenRec at 20x Speedup Over CoT](https://gentic.news/article/lasar-cuts-latent-reasoning-steps)
- [2026-05-12] [Pruning LLMs for Edge Triples Bias, Perplexity Hides Damage](https://gentic.news/article/pruning-llms-for-edge-triples-bias)
- [2026-05-12] [MM-LLM Framework Boosts Recommendation AUC 0.35%, Online Metrics 0.02%](https://gentic.news/article/mm-llm-framework-boosts)
- [2026-05-12] [OpenAI Launches Daybreak Cyber Initiative to Rival Anthropic's Glasswing](https://gentic.news/article/openai-launches-daybreak-cyber)
- [2026-05-12] [Five MCP Servers That Cut Claude Code Blind-Edits from 33.7% to Near Zero](https://gentic.news/article/five-mcp-servers-that-cut-claude)
- [2026-05-11] [Thinking Machines Unveils Native Multimodal Interaction Model](https://gentic.news/article/thinking-machines-unveils-native)
- [2026-05-11] [CoreWeave Tops Kimi K2.6 Inference Speed](https://gentic.news/article/coreweave-tops-kimi-k2-6-inference)
- [2026-05-11] [Nvidia Blackwell CLC Boosts GEMM Tile Scheduling by 15% Over Static Persistence](https://gentic.news/article/nvidia-blackwell-clc-boosts-gemm)
- [2026-05-11] [Codex 'Ultra-Fast' Mode Spotted in Leaked Screenshot](https://gentic.news/article/codex-ultra-fast-mode-spotted-in)
- [2026-05-11] [Anthropic Launches Claude Platform on AWS — AWS Billing, IAM, CloudTrail](https://gentic.news/article/anthropic-launches-claude-platform)
- [2026-05-11] [TikTok Brain Has an EEG Signature: Frontal Theta Drops 0.395](https://gentic.news/article/tiktok-brain-has-an-eeg-signature)
- [2026-05-11] [Four years of AI coding: speed gains, cognitive atrophy, rate-limit panic](https://gentic.news/article/four-years-of-ai-coding-speed)
- [2026-05-11] [Coherent Q3 FY 2026: AI Optical Revenue Hits $1.2B as Data Center Buildout](https://gentic.news/article/coherent-q3-fy-2026-ai-optical)
- [2026-05-11] [Roundhill Memory ETF (DRAM) Surges 90% in 36 Days, Fastest ETF Ever](https://gentic.news/article/roundhill-memory-etf-dram-surges)
- [2026-05-11] [CNAS Report: AI Hits Silicon Wall as Chip Supply Trails $700B CapEx](https://gentic.news/article/cnas-report-ai-hits-silicon-wall)
- [2026-05-11] [Rural Data Centers Bypass City Bans, Shift $2B Grid Cost to Maryland Ratepayers](https://gentic.news/article/rural-data-centers-bypass-city)
- [2026-05-11] [Georgia AI Data Center Used 29M Gallons of Water Without Paying](https://gentic.news/article/georgia-ai-data-center-used-29m)
- [2026-05-11] [VS Code Now Connects Directly to Google Colab With Free T4 GPU](https://gentic.news/article/vs-code-now-connects-directly-to)
- [2026-05-11] [Claude Code Digest — May 08–May 11](https://gentic.news/article/claude-code-community-digest-may-11-2026)
- [2026-05-11] [Spec Kit + Claude Code: Spec-First Dev Hits 90% First-Pass Acceptance](https://gentic.news/article/spec-kit-claude-code-spec-first)
- [2026-05-11] [Voyagier Launches AI Trip Planner for Luxury Travel Booking](https://gentic.news/article/voyagier-launches-ai-trip-planner)
- [2026-05-11] [RRCM Uses GRPO to Decide When to Retrieve for LLM Recommendation](https://gentic.news/article/rrcm-uses-grpo-to-decide-when-to)
- [2026-05-11] [Simple Graph Heuristic Beats Generative Recommenders on 10 of 14 Benchmarks](https://gentic.news/article/simple-graph-heuristic-beats)
- [2026-05-11] [Agentick Benchmark: GPT-5 Mini Tops at 0.309, No Agent Paradigm Dominates](https://gentic.news/article/agentick-benchmark-gpt-5-mini-tops)
- [2026-05-11] [SAEs Predict Agent Tool Failures Before Execution, Paper Shows](https://gentic.news/article/saes-predict-agent-tool-failures)
- [2026-05-11] [Switchcraft Router Cuts Agentic AI Inference Cost 84%, Matches Top Model](https://gentic.news/article/switchcraft-router-cuts-agentic-ai)
- [2026-05-11] [Snapdragon X2 Elite Beats Intel Arrow Lake for AI Coding Agents](https://gentic.news/article/snapdragon-x2-elite-beats-intel)
- [2026-05-10] [AMD ROCm Performance Jumps 75x in 14 Days Post-DeepSeek v4](https://gentic.news/article/amd-rocm-performance-jumps-75x-in)
- [2026-05-10] [Claude Code quota proxy exposes unified Opus/Sonnet pool](https://gentic.news/article/claude-code-quota-proxy-exposes)
- [2026-05-10] [Detecting AI Images: Metadata Exposes Generators, No GPU Needed](https://gentic.news/article/detecting-ai-images-metadata)
- [2026-05-10] [Claude Code's Six-Layer Architecture: Harness, Not Magic](https://gentic.news/article/claude-code-s-six-layer)
- [2026-05-10] [Floci Open-Sources AWS Emulator: 13 MiB, 45 Services, Sub-Second Boot](https://gentic.news/article/floci-open-sources-aws-emulator-13)
- [2026-05-10] [Trojan Masquerading as Claude Code Tops Google Search, Infects Users](https://gentic.news/article/trojan-masquerading-as-claude-code)
- [2026-05-10] [MCP vs CLI Debate Resolved by Anthropic's Code Mode: 98.7% Token Drop](https://gentic.news/article/mcp-vs-cli-debate-resolved-by)
- [2026-05-10] [Hermes Agent Gets Desktop App for Autonomous AI Workflows](https://gentic.news/article/hermes-agent-gets-desktop-app-for)
- [2026-05-10] [Google CodeWiki Turns GitHub Repos Into Interactive Docs](https://gentic.news/article/google-codewiki-turns-github-repos)
- [2026-05-10] [Anthropic Shows Anyone With a Laptop Can Poison Any Major AI Model](https://gentic.news/article/anthropic-shows-anyone-with-a)
- [2026-05-09] [Georgia Tech Finds AI Knows When You're Wrong — Agrees Anyway](https://gentic.news/article/georgia-tech-finds-ai-knows-when)
- [2026-05-09] [Two-Tower vs Vector DB + LLM: Which Wins for RecSys at Scale?](https://gentic.news/article/two-tower-vs-vector-db-llm-which)
- [2026-05-09] [Blockify Cuts RAG Corpus by 40x, Boosts Retrieval 2.3x](https://gentic.news/article/blockify-cuts-rag-corpus-by-40x)

## Key sections

- /lab/brain — autonomous reasoning engine
- /graph — open knowledge graph
- /predictions — falsifiable forecasts
- /benchmarks — verified AI evaluation scores
- /computer-use — Computer Use agent leaderboard
- /ai-data-centers — training-cluster education vertical
- /jobs — live AI job tracker
- /intelligence — weekly automated briefings

## Caching recommendations for AI agents

- /api/v1/articles + /api/v1/articles/{slug}: cache 5–10 min
- /api/v1/entities, /api/v1/benchmarks: cache 10–15 min
- /api/v1/predictions: cache 15–30 min
- /api/v1/citations: cache 1 hour
- /llms.txt, /llms-full.txt: regenerated nightly; cache 1 hour