Shareable cards — Wednesday, July 15, 2026

Wednesday, July 15, 2026 — top 10 database + AI stories, one card each. Save the image, copy its caption, post.

Phone: tap Share image, pick Save Image in the sheet (LinkedIn's app won't take images from Safari directly), then tap the button again — it opens LinkedIn: start a post, add the photo on top, paste the caption. Desktop: Open → save the image, attach it in the composer, paste the caption.

Hand-drawn sketchnote — AWS: Build semantic ontologies for AI assistants using property graphs and vector indexing

Open image

AWS: Build semantic ontologies for AI assistants using property graphs and vector indexing

The day's edition as one sketchnote — full stories + sources:
https://offnetchat.com/news/2026-07-15.html?s=11e8426cbc4a0151#item-11e8426cbc4a0151

#TechNews #AI

AWS: Build semantic ontologies for AI assistants using property graphs and vector indexing

Open image

AWS: Build semantic ontologies for AI assistants using property graphs and vector indexing

▸ Use property graphs to explicitly model data relationships for AI context.
▸ Implement vector indexing to enable semantic search over structured data.

Summary + source: https://offnetchat.com/news/2026-07-15/s/11e8426cbc4a0151.html

#AI #TechNews

Open image

PostgreSQL 19 Beta 2 Released

▸ Beta 2 indicates the feature set is largely frozen, focusing now on bug fixes and stability.
▸ Testers should verify critical workloads against this build to identify regression issues.

Summary + source: https://offnetchat.com/news/2026-07-15/s/173152e7a63e085e.html

#Databases #TechNews

Distill LLaMA 3 into Efficient Cross-Encoder Reranker for RAG

Open image

Distill LLaMA 3 into Efficient Cross-Encoder Reranker for RAG

▸ Replaces expensive cross-encoders with a quantized 8B LLM for real-time RAG reranking.
▸ Achieves 14% gain in answer relevancy and 16% in context precision via distillation.

Summary + source: https://offnetchat.com/news/2026-07-15/s/c6ffb6b96b9f7890.html

#AI #TechNews

LLM Countdown Subcircuit Identified in Llama-3.1-70B

Open image

LLM Countdown Subcircuit Identified in Llama-3.1-70B

▸ LLMs use a reusable subcircuit for precise token counting tasks.
▸ Mechanism compares current position to a goal length estimate.

Summary + source: https://offnetchat.com/news/2026-07-15/s/af6ee43cb5ba2542.html

#LLM #TechNews

DoorDash AI Assistant Boosts Conversion with LLM-Tool Hybrid Architecture

Open image

DoorDash AI Assistant Boosts Conversion with LLM-Tool Hybrid Architecture

▸ Combines LLMs with specialized agents and MCP tooling for precise task execution.
▸ Persistent consumer memory and live backend data drive intent accuracy.

Summary + source: https://offnetchat.com/news/2026-07-15/s/2af73345a56a9552.html

#AgenticAI #TechNews

MySQL 9.7 LTS arrives post-8.0 EOL with stability focus

Open image

MySQL 9.7 LTS arrives post-8.0 EOL with stability focus

▸ MySQL 8.0 has reached End-of-Life, making 9.7 the recommended LTS path.
▸ Release focuses on enhanced stability for long-term production deployments.

Summary + source: https://offnetchat.com/news/2026-07-15/s/a28dbafaa1bd2d1d.html

#Databases #TechNews

LakeQuest Benchmark Tests End-to-End QA Over Heterogeneous Data Lakes

Open image

LakeQuest Benchmark Tests End-to-End QA Over Heterogeneous Data Lakes

▸ Evaluates end-to-end performance over heterogeneous, weakly structured data rather than clean schemas.
▸ Covers three distinct domains: AI/ML metadata, retail banking, and biomedical drug information.

Summary + source: https://offnetchat.com/news/2026-07-15/s/80efcf08174d74bb.html

#Databases #TechNews

Bonsai 27B: 27B-parameter model designed to run on mobile devices

Open image

Bonsai 27B: 27B-parameter model designed to run on mobile devices

▸ 27B parameter count is unusually large for on-device inference
▸ Enables local AI processing without cloud dependency

Summary + source: https://offnetchat.com/news/2026-07-15/s/449874491314fa9e.html

#LLM #TechNews

Gartner: Hybrid AI PCs curb runaway cloud token bills

Open image

Gartner: Hybrid AI PCs curb runaway cloud token bills

▸ Local inference reduces variable cloud token costs significantly
▸ Hybrid models balance latency needs with budget constraints

Summary + source: https://offnetchat.com/news/2026-07-15/s/9e2003119484ce1f.html

#LLM #TechNews

OpenAI encrypts Codex agent instructions, obscuring debug data

Open image

OpenAI encrypts Codex agent instructions, obscuring debug data

▸ Encrypted MultiAgentV2 messages block visibility into agent instructions
▸ Debugging agent errors becomes significantly harder without prompt access

Summary + source: https://offnetchat.com/news/2026-07-15/s/32f47d1988cacafb.html

#AgenticAI #TechNews