The Talent Pipeline Flush | Don't Panic Labs Uncover the truth behind AI-driven layoffs: it's about skill transformation, not massive workforce reduction.

Check out Chad Michel’s post where he breaks down what's really driving the headlines of mass layoffs.

Most workplace conversations about supporting parents stop at policy. Lindsey Hruby-Yardley goes further.

In her latest post, she writes about what it actually took to meet her goals as a nursing mom returning to work, and what made the difference.

The LLM Is the New Runtime | Don't Panic Labs Understand the shift in AI tool design: moving away from bundled models to leveraging existing subscriptions, driven by new standards like MCP.

Most developer tools that include AI are quietly reselling API access. Chad Michel noticed something different in a tool called Pencil, which doesn't bundle its own model at all. It connects to the one you already have.

That's not a workaround. It's a sign of where the whole category is heading.

Building with the GitHub Copilot SDK, Part 4 – Building Mermaid.js Diagrams | Don't Panic Labs Learn how the Copilot SDK empowers you to generate live, editable Mermaid.js diagrams from plain language descriptions, enhancing your documentation workflow.

If your architecture only lives in your head or in a file nobody can find, it might as well not exist.

Chad Michel's latest in his GitHub Copilot SDK series shows how he embedded the SDK into a Mermaid.js diagram builder.

Don’t Panic, Adapt | Don't Panic Labs Worried AI is coming for your job? Discover the real impact of AI on white-collar roles, emphasizing augmentation over automation and how to adapt effectively.

Spreadsheets didn't eliminate accountants. CAD didn't kill drafters. So why are we so sure AI is coming for every desk job?

Chad Michel looks at what real usage data says about AI and white-collar work, including a Microsoft study of 200,000+ Copilot conversations.

Building with the GitHub Copilot SDK, Part 3 – Tracking Project Progress | Don't Panic Labs In Part 3 of Chad's series, discover how to leverage GitHub Copilot SDK to manage project designs and track progress with Earned Value charts.

Most project status updates are theater. Everything is "green" until suddenly it isn't.

In the latest installment of his GitHub Copilot SDK series, Chad Michel tackles that problem head-on.

AI Replacing Software Developers? What the Latest Research Actually Shows | Don't Panic Labs The truth about AI in software development: smarter workflows, new roles, and a profession that adapts—not disappears.

Chad Michel dug into what the research actually shows about AI and developer productivity, who's benefiting, who's falling behind, and what the data says about the future of the profession.

Building with the GitHub Copilot SDK, Part 2 – Wiring Up an Electron + Angular Desktop App | Don't Panic Labs Learn how to integrate the official GitHub Copilot SDK into your Electron and Angular app with a robust, secure architecture that supports real-time streaming.

In the second post of his series, Chad Michel walks through how to wire it up inside an Electron + Angular desktop app with streaming responses, per-workspace sessions, and support for both native GitHub auth and bring-your-own-key setups.

Building with the GitHub Copilot SDK, Part 1 – Getting Started | Don't Panic Labs Explore the practical setup and innovative features of Microsoft's new Copilot SDK, transforming how integrated AI enhances developer workflows.

Chad Michel spent a weekend with the new GitHub Copilot SDK, and now he has a new series that kicks off today.

In Part 1, Chad walks through getting the SDK running, explains why the authentication approach is a standout for developers, and covers how to wire Copilot directly to your codebase.

Specialized Models Beat General LLMs (Sometimes) | Don't Panic Labs Learn why domain-specific models for JS/TS, trained on curated data, lead to better results and resource efficiency compared to larger, general models.

Can a smaller, specialized model for JavaScript and TypeScript outperform a flagship general-purpose LLM at a fraction of the cost?

For a lot of that kind of work in early 2026, the answer is yes. Chad Michel breaks down why, and when you still want a generalist in your corner.

Pulling Data from Microsoft Access to CSV: What I Built and Why | Don't Panic Labs Learn to extract tables from Access databases (.mdb and .accdb) into CSVs on Mac, combining mdbtools and UCanAccess for reliable, swift data processing.

Chad Michel wrote about a challenge many teams face: getting data out of legacy Microsoft Access databases.

Putting a UI on My Access-to-CSV Converter | Don't Panic Labs Discover how to transform a command-line tool into a user-friendly web app using Flask and Bootstrap, streamlining data export from Access files.

Sharing command-line tools with others can often lead to complicated setup steps.

Chad Michel explains how he moved from a terminal-only Access-to-CSV converter to a lightweight Flask web app.

Agent Chat using LangChain Part 3 – Fun Examples | Don't Panic Labs In the final installment of his LangChain series, Chad Michel shows how natural language interactions can replace tedious UI workflows with powerful agents.

Natural language beats navigating menus every time.

Chad Michel's final post in his LangChain series shows how AI agents can search contacts, create events, and even generate test data, all from simple requests. The best part? Tools combine to create thousands of possible workflows.

Agent Chat using LangChain Part 2 – Token Streaming with WebSockets | Don't Panic Labs Discover how adding token streaming over WebSockets transforms AI chat responsiveness, creating a seamless, real-time user experience.

“Adding proper token streaming completely changed how the agent feels—from “kinda slow but functional” to “this is actually fun to use.””

That is how Chad Michel described his agent after updating it so users see responses appear word by word.

Agent Chat using LangChain, Part 1 – Tools and Memory | Don't Panic Labs Chad Michel walks through integrating LangChain tools and memory into a NestJS backend to create an actionable AI agent.

Read how Chad Michel is building a system that enables users interact with it using natural language.

#softwareengineering #softwaredevelopment #AI #LangChain

We discuss Volatility-Based Decomposition as a way to group systems by where change occurs, usually before a project begins.

In his latest post, Chad explores a data-driven approach to this process using what he calls "Combat Agents".

dontpaniclabs.com/blog/post/20...

We love our community!

Throughout November, we collected $1264.01 and over 70 lbs of food donations for the Food Bank of Lincoln to support our neighbors in need.

Thank you for your generosity!

The Price Per Token Is Dropping. Will It Stay That Way? | Don't Panic Labs Chad Michel explores the sustainability of dropping LLM token prices against the backdrop of rising AI training costs.

Will prices eventually reflect the true cost of training?

Chad Michel analyzes the economics of the current AI movement. While data from @epochai.bsky.social shows the price of tokens falling over time, the cost to train a model is doubling every eight months.

Putting GitHub’s New AI Agent to the Test | Don't Panic Labs Chad Michel tests GitHub's new AI Agent tasks. See how AI-powered issue completion and repo initialization performed in a real-world workflow.

What happens when you let GitHub's new AI agent take over a task? Chad Michel recently ran the experiment.

From Guesswork to Structured Context: How Figma MCP Changed My Dev Workflow | Don't Panic Labs Jack Nolley dives into Figma's Dev Mode MCP. Learn how to stop guessing at pixels and feed AI structured design data , letting it generate code that understands your design system.

In his first post for the Don't Panic Labs blog, Jack Nolley tackles a common friction point in the developer-designer handoff: the gap between pixels and design intent.

Read how he moved away from relying on screenshots and manual "pixel-peeping" to utilizing Figma’s Dev Mode MCP server.

Harnessing the Hype: A Pragmatic Guide to Using AI in Development | Don't Panic Labs Chad Michel outlines a framework for using AI in development. Learn productive practices and essential guardrails to harness AI's power with discipline and intent.

AI is a powerful tool, but it's not a shortcut to bypass critical thinking. Using it with discipline and intent is the key to avoiding a mess.

In his latest post, Chad Michel shares our current framework for harnessing AI effectively.

Finding Joy in the Work and in What Comes Next | Don't Panic Labs Doug Durham shares words of wisdom as he announces his stepping back from daily operations at Don't Panic Labs.

We have some big news to share today.

After 15 years, Don’t Panic Labs Managing Partner Doug Durham is stepping away from daily operations.

In his latest post, Doug reflects on his career and shares his perspective on what's next.

Beyond the Booth: How We Talk About Coaching and Growth in Engineering (and Everywhere) | Don't Panic Labs At the 2025 Nebraska.Code() conference, we asked engineers about coaching and growth. Here’s what they shared and what it means for the future of development teams.

At @nebraskacode.bsky.social this summer, we wanted to do something different at our booth. Instead of distributing handouts or giving demos, we asked conference attendees three questions - and the responses were eye-opening.

Vectors, Python, AI, and Your Face | Don't Panic Labs Everett Rhodes dives into Python, AI, and vector databases showing how a prototype facial recognition system took shape with Cursor.

What happens when you let AI help you build a facial recognition system from scratch?

That’s the experiment Everett Rhodes recently ran. Armed with Python, a vector database, and Cursor, he set out to see how far AI-assisted coding could go in tackling a real (but lightweight) prototype.

Kansas City, here we come!

We’re proud to be back again as a sponsor for the Kansas City Developer Conference (@kcdc.bsky.social). This year, we’re bringing big ideas, a brand-new AI booth activity, and a lineup of talks from Bill Udell, Brian Zimmer, and Doug Durham.

See you at #KCDC2025!

Understanding window.open() Behavior on iOS Safari | Don't Panic Labs Read how Sarah dealt with the strict pop-up rules enforced by Safari on iOS.

Read how Sarah dealt with the strict pop-up rules enforced by Safari on iOS.

Next Tuesday August 5th we will be at Salt Mine City Brewing Co. in the Haymarket from 6:30 to 8:30pm!

MCP Magic: AI-Powered Azure SQL Analysis in Plain English | Don't Panic Labs Branden Barber demonstrates how to make data analysis accessible to both technical and non-technical users.

Imagine being able to ask your database questions in plain English and getting visualizations quickly.

In his latest post, Branden Barber shows us how to make data analysis possible for technical and non-technical users alike.

#SoftwareDevelopment #SoftwareEngineering #AI

Nebraska.Code() Full-Day Workshops are this Wednesday, July 23rd!

nebraskacode.amegala.com

#ProductDiscovery #OKRs #Personas #Storymapping #InteractiveMapping #JavaScript #React #Kubernetes #ProblemSolvingToolbox #RAGSolution #AWS #CSharp #AsynchronousProgramming #SustainableIT #Tech #Nebraska

Nebraska startup community continuing to grow, add connections Nebraska’s startup community has been working to create more ways for new businesses to connect and find resources.

Nebraska's startup scene is growing and changing. Organizations like 1 Million Cups, @dontpaniclabs.com and Open Range are working to create community and resources for the new companies.

journalstar.com/news/local/b...