All Data and AI Weekly #224-12 Jan 2026
( AI, Data, NiFi, Iceberg, Polaris, Streamlit, Flink, Kafka, Python, Java, SQL, MCP, LLM, RAG, Cortex AI, AISQL, Search, Unstructured Data )
๐ The Lead Story: Snowflake Acquires Observe
"Observability is a Data Problem"
In a major move to own the entire data lifecycle, Snowflake has announced its intent to acquire Observe for approximately $1 Billion. This acquisition signals a shift away from siloed monitoring tools towards "AI-powered observability" where telemetry data (logs, metrics, traces) lives directly within the Data Cloud.
- Why it matters: It promises to break down the wall between "business data" and "operational data," enabling AI agents to troubleshoot infrastructure using the same context they use for analytics.
- Key Stat: Snowflake aims to lower observability costs by allowing enterprises to ingest and retain 100% of telemetry data, rather than sampling it due to cost constraints.
- Press Release | Observe Announcement
โ๏ธ Snowflake & Data Engineering
Technical deep dives for the modern data stack.
โก Low-Latency CDC & OpenFlow
The lines between transactional and analytical databases are blurring. A new "Lift and Shift" capability allows native Postgres tables to run directly within Snowflake, paired with OpenFlow (Snowflake's managed NiFi service) to build low-latency CDC pipelines. This is critical for feeding real-time AI models without fragile ETL glue code.
- Low Latency CDC for Real-Time AI
- Syncing Postgres to Native Tables with OpenFlow
- Turbocharge Pipelines with OpenFlow
๐ง Iceberg & Unity Catalog
Interoperability continues to be a major theme. New patterns allow connecting Databricks Unity Catalog directly to Snowflake via the Iceberg REST catalog, reinforcing the "open table format" war is ending in a truce of compatibility.
๐ค The AI Agent Ecosystem
Autonomous coding, Red Teaming, and Orchestration.
๐ ๏ธ Builders & Orchestrators
- Auto-Claude: An autonomous coding agent framework that goes beyond simple completion. It uses a multi-agent approach to plan, execute, and debug code.
- Gas Town: Steve Yeggeโs latest project. Itโs a multi-agent orchestration system (using tmux) designed to manage fleets of Claude Code instances.
- VideoRAG: A new framework that allows LLMs to "reason" over long-form videos by retrieving specific temporal segments rather than processing the whole file.
๐ก๏ธ AI Security (Red Teaming)
- Strix: An open-source AI agent designed for automated penetration testing. It mimics human hacker behavior to find vulnerabilities in your stack.
- Cortex Analyzers: For those using TheHive for incident response, these analyzers are essential for enriching threat data.
๐ Security Corner
OSINT and Threat Intel tools for the paranoid.
- Web-Check: The ultimate "check this website" tool. It runs dozens of checks (DNS, SSL, headers, server info) in one dashboard.
- IntelOwl: A scalable API for threat intelligence management. It orchestrates various analyzers to generate a single "threat report" for an observable.
- DataSploit: An OSINT framework to perform reconnaissance on domain names, emails, and usernames.
๐งฐ Developer Toolkit
Utilities to speed up your workflow.
- VSCodium: A clean, telemetry-free build of VS Code. If you want the editor without the tracking, this is the standard.
- YouTube Playlist Downloader: A simple, robust script for archiving playlists locally.
- Hardwood: Interesting work from the data engineering community (likely related to durable execution or storage engines).
๐ Full Resource List
- 2025 Databases Retrospective (Pavlo)
- From ETL to Autonomy (The New Stack)
- Snowflake Regulated Workloads & Backups
- Streamlit: 30 Days of AI
- LLMRouter: Route queries to the best LLM
- TruLens: Evaluating Agent Workflows
๐ญ 2026: SNACK-AI
As we look forward in 2026, we are introducing SNACK-AI. This architectural pattern is designed to power AI applications at any scaleโfrom mobile phones and robots to massive enterprise clusters.
SNACK-AI integrates:
- Snowflake Platform
- NiFi (Apache)
- Apache Iceberg (Apache Iceberg)
- Cortex AI (Agents/Search/MCP)
- Kafka (Apache Kafka)
Deep Dive into SNACK-AI:
- Read: SNACK-AI: The 2026 Pattern (Medium)
- Code: Official SNACK-AI GitHub Repository
- Weekly Context: SNACK-AI Overview
Thanks
https://github.com/timothyspann
ยฉ 2020-2026 Tim Spann https://www.youtube.com/@FLaNK-Stack
Top comments (0)