How to get 80% of GPT-4’s value for 20% of the cost (5 tips for reducing AI spend)
AI model costs can explode fast – and overpaying is common. The reality is, you don’t need GPT-4 for most business tasks like customer support, summarization, or data extraction. Switching to smaller AI models, smart routing, and techniques like RAG, c …
You Can’t Trust “Edit Query Text” in SSMS’s Execution Plans.
When there are IF branches or conditional logic, “Edit Query Text” only shows you the branches that actually executed. It’s easy to miss whole swaths of code that didn’t happen to execute one particular time. To demo it, let’s create a simple stored pr …
What is fillfactor in PostgreSQL – and when should you adjust it?
fillfactor in PostgreSQL controls how full table or index pages are. The lower the fillfactor, the better the performance on update-heavy tables. In this article, learn everything you need to know about fillfactor in PostgreSQL – from what it is, to ex …
What are the top database platforms in 2026? A look at the latest data
The database market in 2026 remains stable at the top, with AWS, Microsoft, Oracle, and Google Cloud Platform leading by revenue. However, open-source and cloud-native platforms like MongoDB, Snowflake, and Databricks are gaining popularity – albeit ve …
There Are Days When I Feel Like Giving Up on the Plan Cache and Query Store.
In theory, SQL Server performance monitoring is pretty simple: Review the server’s top wait types Find the queries causing those wait types Fix those queries, or improve the way the server reacts to them (indexes, settings, etc.) But in practice, step …
Vibe coding and databases: the hidden risks of AI-generated database code
“Vibe coding” with AI can introduce hidden risks in database development, including security vulnerabilities, broken query logic, and data integrity issues. In this article, Chisom Kanu explores how Andrej Karpathy’s “vibe coding” trend reached databas …
Contest Winners: Make the Comments Look Like My ChatGPT History.
I asked y’all to make the comments on this post look like my ChatGPT history, and y’all understood the assignment. The answers were all over the place from accurate to humorous. Here are the 5 runners-up first, each of whom won a Fundamentals Bundle fo …
Gouvernance Fabric : Autoriser des items Fabric spécifiques sans ouvrir toute la plateforme ?
« Nous voulons que les utilisateurs puissent créer des Data Agents, tout en gardant un contrôle strict sur les permissions de création dans Microsoft Fabric » a { text-decoration: none; color: #464feb; } tr th, tr td { border: 1px solid #e6e6e6; } tr t …
How to handle large datasets and parallel processing in esProc SPL
As data volumes continue to grow exponentially, the ability to efficiently process and analyze large datasets has become a critical skill for data professionals. In this article, part of the ‘Moving from Python to esProc SPL’ series, learn how to tackl …
SQL ConstantCare® Population Report: Spring 2026
It’s time for our quarterly update of our SQL ConstantCare® population report, showing how quickly (or slowly) folks adopt new versions of SQL Server. There were only slight 1-2% shifts, nothing big: SQL Server 2025: 1% SQL Server 2022: 31%, up 2% from …
There’s a Bug in sys.dm_exec_query_plan_stats.
When you turn on last actual plans in SQL Server 2019 and newer: ALTER DATABASE SCOPED CONFIGURATION SET LAST_QUERY_PLAN_STATS = ON; The system function sys.dm_exec_query_plan_stats is supposed to show you the last actual query plan for a query. I’ve h …
Local vs remote MCP servers – which should you choose?
MCP (Model Context Protocol) servers let AI apps access real-time data and external tools, improving accuracy beyond static model knowledge. Choosing between local and remote MCP servers affects security, scalability, and integration. In this article, …
Building My Dev/Prod Demon Hunters Session, Part 2: The Tactics
This week at the SQLBits conference in Newport, I’m teaching an all-new all-day session called Dev-Prod Demon Hunters: Finding the Real Cause of Production Slowness. It filled up initially, but they’ve opened up more seats for it, so if you’re interest …
When, and when not, to use LLMs in your data pipeline
“Let’s add LLMs to the pipeline” has become a familiar refrain in modern data teams, but turning that idea into production reality is where things often go wrong. While large language models can unlock powerful capabilities, they’re just as likely to i …
Updated First Responder Kit and Consultant Toolkit for April 2026
If there was ever a release you should skip for reliability reasons, it’s this one. There’s a lot of AI-edited code in here. I’m really only putting it out there because I want to let early adopters start playing around with a few new things: sp_BlitzI …