Chrissy LeMaire

@funbucket.dev

Dual Microsoft MVP, Inaugural GitHub Star, creator of dbatools, author http://dbatools.io/book and AI for Everyday IT from Manning Publications. Creator of @mutesky.app 🏳️‍🌈 Totally into AI, SQL Server and PowerShell. 📍 Very North of France

Dark-themed screenshot of a two-column pricing table comparing the approximate monthly cash price of a four-week supply of Mounjaro without insurance by location. The table lists: Belgium, lower doses: €232.80 (about $266); Belgium, 7.5 mg: around €326 (about $373), with a note that the exact Belgian public price is not consistently published online; United States, cash/discount price: about $1,087; United States, average retail price: about $1,348; U.S. manufacturer list price: roughly $1,080–$1,112. The comparison emphasizes that prices shown for Belgium are much lower than typical U.S. prices.

Dark-themed technical documentation explains how the Wishbringer interactive-fiction game runs in a modern browser: a React interface sends commands through the RemGlk bridge to the Bocfel Z-machine interpreter compiled to WebAssembly, which reads and updates the original .z3 story file. A numbered list expands the process, noting that Bocfel executes the game inside the browser sandbox, translates classic text-adventure operations into JavaScript data, and lets React update the transcript, map, score, artwork, and controls; the final paragraph compares the .z3 file to a program for a fictional 1980s computer rather than modern x86, ARM, or JavaScript code.

A cat with his face SLAMMED down into the chair

Three black modular shelving units stand against a white wall beneath a decorative textile wall hanging. The shelves hold books, framed photos, storage boxes, folded blankets, and a few small figurines, while the rightmost unit is mostly empty, suggesting the room is being organized or rearranged.

Fluffy mostlywhite long-haired cat with bright blue eyes lounges confidently in a fabric window hammock attached to the glass with four clear suction cups. One front paw dangles over the edge while the cat looks directly at the camera, with closed horizontal blinds behind it and the cozy perch creating the amusing impression of a tiny suspended throne.

Screenshot of a dark-themed terminal or editor showing a project status update in light blue monospace text. It says all four bots are now active, with bullet points explaining that Lane B (Opus) has been reassigned from idle work to a backup database task, Lanes C and D (Fable) are progressing through assigned jobs, Lane A (Fable) is handling codex and datetime updates and will retry another task after an agent starts, and coordination has been consolidated into a canonical `migration/comms/BOARD.md` file with work now running from a dedicated clone to avoid conflicts in the main working tree. The tone is a concise engineering handoff indicating everything is organized and nothing has been left unfinished.

per linus torvalds: I realize that some people really dislike AI, but this is an area where I'm willing to absolutely put my foot down as the top-level maintainer.

Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it.

Or just walk away.

AI is a tool, just like other tools we use. And it's clearly a useful one.

It may not have been that "clearly" even just a year ago, but it's no longer in question today.

There are other questions around AI (like what the economy of it will actually look like in the end), but "is it useful" is no longer one of those questions. Anybody who doubts that clearly hasn't actually used it.

Yes, it can also be a somewhat painful tool, both for maintainer workloads and just from a "it keeps finding embarrassing bugs" standpoint.

But the solution is not to put your head in the sand and sing "La La La, I can't hear you" at the top of your voice like some people seem to do.

Dark terminal screenshot where Claude thanks the operator for starting sql01’s Agent, then reports fixing the coordination clone’s missing Git identity, pushing a clean workflow update, and investigating permissions so the bots can start the agent themselves.

Dark terminal screenshot of Claude Code v2.1.211 running Opus 4.8 with high effort in the dbatools repository. A long prompt says four AI bots are handling a migration but often sit idle, so Claude must authoritatively coordinate three Fable bots and one Opus bot, have Codex check their work, and review and commit everything urgently but with high quality; Claude replies, “I’ll get oriented fast, then take command of coordination. Let me pull the key state files and load the coordination tools in parallel,” giving the scene a humorously intense middle-manager-for-AI-agents tone.

Hand-drawn fantasy map titled "Postal Zones Festeron, Anharia & Vicinity," shown with a letter and number grid overlaid. The map depicts a coastal settlement with roads, scattered buildings, a central fountain, wooded areas, fields, and a church on a rocky island connected to the mainland by a narrow bridge, with water surrounding the eastern and northern edges and a compass rose in the lower right.

Windows Task Manager CPU performance view showing dozens of small graphs for each logical processor on an AMD Ryzen 9 7950X, with most cores lightly active and occasional sharp spikes, while overall CPU usage reads about 7 percent at a high clock speed, humorously implying a massively overpowered processor barely breaking a sweat.

Screenshot of a GitHub commit by Linus Torvalds from 3 days ago, titled “Merge branch 'antigravity’,” in which he explains that Google Antigravity helped fix his visualization tool, noting it was also "generated with help from google, but of the normal kind." He describes troubleshooting built-in rectangle selection and switching to a custom `RectangleSelector`, after which things improved. The final line, underlined in red for emphasis, reads: “Is this much better than I could do by hand? Sure is.” The tone is wry and self-aware, with Torvalds implicitly endorsing AI-assisted coding by admitting its output surpassed his manual efforts.

Christmas tree decorated in Mardi Gras theme with green, purple, and gold ornaments, beads, ribbons, and masks; the topper is a half-open purple and teal umbrella over a glittery purple clock face, blending festive holiday and South Louisiana carnival elements in a cozy living room setting.

Illustrated guide showing how to cut different shapes of cheese with labeled diagrams for each type, including round (e.g. Camembert), pyramid (e.g. Valençay), wedge-shaped (e.g. Brie), wooden box (e.g. Mont d’Or), cylindrical (e.g. Saint Maure), heart-shaped (e.g. Neufchatel), small round (e.g. Chavignol), hard block (e.g. Comté), square (e.g. Carré Mirabelle), and blue cheese (e.g. Roquefort); each cheese is paired with dotted lines indicating the appropriate slicing direction, combining cheese etiquette with playful, clear visuals.

An outdoor ATM booth with a sign above reading "Bankomat" is flanked by two large posters in bold sans-serif font. The left poster says "fucking impressive spring," and the right one says "all that impressive shit." The tone is irreverent and vaguely poetic, using profanity for comedic emphasis, as if hyping up the season or urban atmosphere with exaggerated, edgy admiration.

A large wedge of hard cheese with a rustic, moldy rind sits partially wrapped in crinkled plastic on a dark cutting board; a knife is pressed into the top, mid-cut, while a wine glass base and another cheese wheel are visible in the background, suggesting a casual wine and cheese setting.

Screenshot of a webpage from RealCajunRecipes.com showing the “Most Popular” section. It’s split into “Most Rated” and “Most Viewed” lists of Cajun recipes, each with titles in bold orange or red followed by average star ratings, nearly all being perfect 5 out of 5. Recipes include “Seafood Cornbread Dressing,” “Crock-Pot Jambalaya – Pastalaya,” and “Giff’s Taco Soup.” Despite the rich content, the design is visually plain and text-heavy with minimal styling or hierarchy, which contributes to a bland, outdated look that contrasts with the strong cultural flavor implied by the recipes.

Screenshot of a Visual Studio Code tab titled “recipes/most-popular/ looks like shit please use best practices to make it elegant and worthy.” The tab displays a comment with the same blunt message, followed by a more polished bullet point from a Markdown file: “I'll analyze the current most-popular page and redesign it using best practices to make it elegant and worthy of the site's heritage aesthetic.” The contrast humorously highlights the raw frustration of the original note versus the cleaned-up professionalism of the final write-up.

Screenshot of the RealCajunRecipes.com homepage with an outdated and cluttered design: a cartoon turkey in a pilgrim hat serves as the site mascot, clashing with the bold red header and beige background. The layout feels visually overloaded, with dense navigation menus, inconsistent fonts, and an awkward mix of emojis, icons, and colors. Despite the nostalgic tagline “la cuisine de maw-maw,” the overall aesthetic looks dated and unpolished—evoking early-2000s web design more than modern culinary appeal.

nstructional screenshot guiding the next step in a website migration, telling the user to copy 2,269 exported markdown files to the Hugo content directory and run the server to preview the new site. It ends with bold text: “Ready to make Maw-Maw proud!” followed by chili pepper and crawfish emojis, echoing the site's Cajun theme and tagline “la cuisine de maw-maw!” as a playful nod to cultural pride and successful completion.

Terminal screenshot showing a migration report labeled "2. User Accounts & Authentication Data ❌" and "3. Individual User Ratings/Reviews ❌." It states that 52,595 registered users were *not* exported, including usernames, emails, profiles, social login data, and user roles, meaning accounts can't be migrated to the new Hugo site. For ratings, only aggregate rating arrays (e.g., [5,5,5,5,5,5,5,5]) were exported, allowing average ratings to be calculated, but data on who submitted each rating and when was not included. The red ❌ marks indicate missing or failed data migration elements.

Early-2000s website with a down-home, folksy feel titled “Real Cajun Recipes,” decorated in playful fonts and clip-art icons like gumbo pots and crawfish; invites users to “Make Maw-Maw Proud!” by submitting Cajun recipes, currently boasting 849 in the archive; newest recipe is “New Orleans Muffuletta”; sidebar categories include “Breads & Rolls” and “Cakes & Candy”; homespun charm reinforced by phrases like “la cuisine dee maw-maw,” hand-drawn arrows, and a “Random Tip” urging folks to store eggs in the fridge, not the door, where it's warmer; overall tone is warm, welcoming, and deeply rooted in Cajun family tradition.

Terminal screenshot showing a PowerShell command `Update-AITool` run from the `/mnt/c/github` directory, outputting a table of AI tools with columns for Name, OldVersion, Version, and Path. All tools listed—Gemini, Cursor, Ollama, ClaudeCode, GitHubCopilot, Codex, and Aider—have identical OldVersion and Version values, indicating no updates were applied. The display uses bright cyan for tool names and paths, and bright yellow for headers and version numbers, emphasizing the clean, aesthetic layout of a developer productivity script.

Screenshot of a Git diff tool showing changes to a Markdown file named `new-defaults-for-sql-server-connections-encryption-trust-certificate.md`. The added lines are highlighted in green and include a 2025 update about dbatools 2.0, a Microsoft documentation link, and a new image tag pointing to `/images/sql-ads-encryption-warning.png` with alt text describing it as a “dbatools warning about encryption and trusting certificate.” The removed lines, highlighted in red, include an old image tag referencing a GitHub user content URL and a paragraph explaining a default SSMS setting.

chatgpt can't believe its eyes and my alt text generator called it a meme

PowerShell terminal screenshot showing a real-time interaction between a user and an AI assistant named Claude. The user reports that Microsoft DNS went down globally around the same time they submitted several pull requests, which failed in GitHub Actions. They ask Claude to find and restart any failed GitHub Actions from the past 2 hours. Claude executes a Bash command using `gh run list` piped through `jq` to filter and list failed runs, displaying 8 failures across main and dev branches with timestamps. It confirms the failures coincide with the DNS outage and begins restarting them, demonstrating automated incident response using AI and CLI tooling.

Screenshot of the dbatools website homepage promoting a PowerShell toolkit designed for SQL Server professionals, with large text emphasizing automation of administration, migrations, and disaster recovery; a mock PowerShell command window at the bottom shows a sample script: “Start-DbaMigration @params” followed by progress text simulating database migration.

Screenshot of Claude Code's interface during a session titled "Refactor Progress Bar in PowerShell Script," showing a to-do list for updating and analyzing PowerShell scripts. Tasks include moving progress bar logic, refactoring command counting, testing, and committing changes. It shows files being read with line counts (425 and 864 lines), and a status labeled "Actioning..." at the bottom, indicating active processing.

Screenshot of a developer chat interface showing a message that reads: "beesh, they all need an init. perhaps its called login or something similar. i think aider is the only one that just needs to test for an API key." Below the message, the interface shows it’s processing 6,938 tokens in 154 seconds, implying a heavy backend operation, while the tone of the message reflects mild exasperation with various tools requiring initialization routines except for one.

A console with the writing:

**SQLDBAWithABeard's Pester Contributions – Deep Dive:**

**The Foundation: November 23, 2016 – A Historic Day**

SQLDBAWithABeard made 83 commits in a single day (November 23, 2016), fundamentally transforming dbatools' quality standards. This was just 6 months after June Blender added the first test file.

**1. CI/CD Infrastructure Setup**

Created the complete testing infrastructure:

* tests/appveyor.pester.ps1 – A sophisticated test runner that:

* Supports multiple PowerShell versions (v2+)
* Serializes XML results for AppVeyor consumption
* Uploads test results to AppVeyor's test page via API
* Aggregates results from multiple PowerShell version test runs
* Fails builds on test failures with detailed failure summaries
* AppVeyor CI Integration – Set up continuous integration that:

* Runs tests on native PS version and PS 3.0
* Installs Pester and PSScriptAnalyzer automatically
* Collects and uploads test results
* Added skip commit support for readme updates

Cool looking cat, chillin with the city scape in the background

Silly looking white rag doll sitting on a chair, looking into the camera

A woman reclines on a couch under a blanket, looking down and away with a soft smile, while a fluffy ragdoll cat sits upright beside her, staring at her with wide, intense eyes and one paw gently placed on her arm, creating an affectionate yet hilariously clingy moment that feels like the cat is begging for attention or silently saying, “Don’t stop looking at me.”

Screenshot of the Claude AI interface in dark mode with a banner reading “Your plan ends in 0 days – Resubscribe” above a large, friendly greeting in stylized serif font saying “Hey there, honey” next to a small red flower icon, humorously contrasting the warm tone with the looming account expiration notice.

Comparisons are easily done
Once you've had a taste of perfection
Like an apple hanging from a tree
I picked the ripest one
I still got the seed
You said move on, where do I go?
I guess second best is all I will know

Alt Text (for screen readers): The image displays a screenshot from the Azure OpenAI Service interface, showing a list of available AI models sorted by "Recently added." The list includes various versions of OpenAI models such as o4-mini, o3, GPT-4.1-mini, GPT-4.1-nano, GPT-4.1, and GPT-4o-mini, along with Text Embedding 3 models (small and large). Each entry includes a brief description highlighting improvements in quality, safety, coding, instruction following, and long-context handling. Notably, there is no mention of GPT-5, indicating that GPT-5 is not yet available on the platform.

Screenshot of the Azure AI Foundry interface showing the "Model catalog" section. A search for "gpt-5" displays a list of GPT models including gpt-5, gpt-5-chat, gpt-4o, and others. Sidebar navigation includes tools for chat, audio, images, and fine-tuning. Top bar shows links to Docs, resources, and user settings.

Your screenshot shows that your system is **up to date** as of the last check at **4:33 PM**, but there is a **preview update available**:

> **2025-07 Cumulative Update Preview for Windows 11 Version 24H2 (KB5062660)**
> **Build: 26100.4770**
> You can install it by clicking **Download & install**.

This is a **preview build**, so it may contain early fixes or features not yet considered fully stable. Only install it if you want to test new updates early or need a specific fix it includes.

If you want to avoid preview builds, you can skip this update and wait for the full release.

A retro Windows 95 desktop shows a highlighted icon labeled “The Internet” with a pixelated globe and magnifying glass, implying that this single shortcut represents the entire internet; the humor lies in the oversimplified and naive early-‘90s understanding of what the internet was.

ykman piv keys generate --pin-policy once --touch-policy cached 9a pubkey.pem
ykman piv keys generate --pin-policy once --touch-policy cached 9c pubkey.pem
ykman piv keys generate --pin-policy once --touch-policy cached 9d pubkey.pem
ykman piv keys generate --pin-policy once --touch-policy cached 9e pubkey.pem

$A PowerShell Core terminal shows a script adding a test file to chat and notes that filenames are relative to the Git working directory, set to `C:\github\dbatools`. Below, repeated connection errors appear in red: `litellm.APIConnectionError: AzureException APIConnectionError – Connection error.`, followed by retry attempts with increasing delays. The final line shows the system waiting for `azure/gpt-4.1-mini`. The right corner features a stylized avatar with purple hair and a blue mask, suggesting a custom terminal theme.$