Join us tomorrow at 2 pm ET for AI Coffee Time as our senior devs talk AI workflows and harnesses. Sharing approaches are River Lynn Bailey, A.J. Hekman, Dave Mosher, Rick Reilly, and Jed Schneider.
Register: testdouble.com/debrief/ai-coffee-time-w...
Remember when web crawlers were called "spiders"? Bots were stitching the internet together from the beginning.
@cathycolliver.bsky.social says the fabric is worn and we're all restitching it. AI tools will help. Humans decide the pattern.
link.testdouble.com/141dd6
Results:
→ Aggressive deadlines met
→ Well-tested, secure code
→ Knowledge transfer to internal teams
"Kenneth delivered features quickly from week one. He also raised the bar on automated testing and proactively helped ramp up new team members."
— Marco Kitagawa Pedroso, Clever
EdTech at scale: Clever case study
Challenge: Two mission-critical projects
1. Automate user data syncing for 940+ school districts (12.6M daily API requests!)
2. Build self-service identity management system that's easy to use for school administrators
Thoughts from Principal Consultant Dave Mosher: AI lowered the cost of outputs but raised the cost of being wrong. The differentiator isn't the model anymore, it's the harness. Tests, boundaries, feedback loops, and decision records encode taste and judgment. Outputs are cheap. Outcomes never were.
Feeling overwhelmed trying to figure out which AI harness approach works best for you and your team?
Join us April 22 as some of our senior software developers share and discuss the AI harnesses and approaches that they're using in their work.
Register: link.testdouble.com/17e917
If you can't attend at that time, register anyway and we'll send you a recording link!
Healthcare teams spend more time managing tech than improving care. Our research reveals why healthcare tech and product teams need to focus on root causes, not symptoms.
https://link.testdouble.com/a8bdb4
Monoliths might actually be strategic business tools (not tech debt).
The uncomfortable reality: Most teams break apart monoliths to solve orgproblems, then spend years trying to coordinate the pieces they scattered.
Why? https://link.testdouble.com/52aa3c
Staff Consultant Steve Jackson talks about the value of software tests and automated test suites as onboarding tools for your new software engineers.
www.youtube.com/watch?v=WgT_...
🚨Too many cooks in the codebase! Adding more developers to solve velocity problems? You might be scaling your bottlenecks, not your productivity. Here's what actually slows teams down (and what to fix first). https://link.testdouble.com/904ed5
A 1,000-line setup guide made onboarding feel like navigating fog. So Jeremy Fairbank built a Claude skill to be the captain: a documentation companion that guides developers through setup step by step.
Six lessons from the build: https://link.testdouble.com/5c7ba7
Microservices are often just distributed dysfunction.
Your monolith's "forcing function" for org alignment might be exactly what's driving your business success.
The anvil of alignment principle: Sometimes constraints create clarity.
Whoops, the link shortener crapped out on this one, and Bluesky doesn't have edits, so here's the correct link: testdouble.com/debrief/ai-c...
More autonomy for AI agents doesn't mean better code. It often means worse code, faster. Test Double senior developers share how constraints and human judgment shape what AI actually produces. April 22, 2 pm ET.
Register: testdouble.com/debrief/ai-c...
Test Double co-founder @searls.bsky.social shared his perspective on AI impact on jobs for this New York Times op-ed.
What do you think?
www.nytimes.com/2026/04/08/opinion/ai-we...
Welcome to our newest doubt agent, Schwa! 🎉💚
Thought from @docondev.bsky.social on why "go faster" backfires: teams cut composition and validation corners, creating the brittleness that slows them down. Speed comes from flow: small batches, trust, systems built for adaptation. Not from pushing harder.
The best remote team building happens when you're NOT trying to build the team.
Unstructured "coffee and code" sessions > formal team meetings.
How it works: https://link.testdouble.com/220119
We inherit our constraints. Legacy systems, team expertise, deployment limits. HTMX turns these into advantages.
One script tag. Enhanced UX. No rewrite. Ship today.
Stop fighting your stack—enhance it. Your users are waiting.
Dave Mosher shows how:
https://link.testdouble.com/845798
AI doesn't need more freedom, it needs constraints. On April 22, Test Double senior devs share how encoding taste, judgment, and guardrails into agentic coding leads to better outcomes. Join us at 2 pm ET to find out more about what we're doing.
Register: https://link.testdouble.com/17e917
Your Python code doesn't need more abstractions. It needs fewer moving parts.
Staff Consultant @kadams54.bsky.social lays out seven tenets for functional simplicity in Python.
https://link.testdouble.com/0a1025
Talk to someone else who gets it. We offer free office hours to talk about challenging problems.
Pair with an experienced software developer or product manager who will ask you curious questions to help get you unstuck. https://link.testdouble.com/b0de11
Senior DevOps Consultant Joé Dupuis says stop spamming the permission button on Claude.
Learn how to set up Claude Code's sandbox so your AI agent works freely and securely without constant interruptions.
Get the illustrated guide and get in the box: https://link.testdouble.com/fef21c
Welcome to our newest double agent, Tamara Nicholl-Smith! 🎉💚
The quality of AI-generated code is informed by taste, judgment, and constraints. Our senior devs will share their agentic coding approaches April 22, 2 pm ET//11 am PT, plus discuss workflows and harnesses.
Register for the webinar: https://link.testdouble.com/17e917
Jonathon Baugh shares an idea that originated from his 1980s frustration with that annoying VCR blinking 12:00.
AI is reshaping the workforce, blurring roles, and breaking career ladders. Orgs can balance using a triad of people, technology, and business.
https://link.testdouble.com/b419b3
Staff Consultant Brian Hughes shares an 80/20 rule for test coverage in your software application, along with practical tips.
youtu.be/uD5VRFQvFPc?...
Neal Lindsay thinks of mutation testing as a more strict version of code coverage. And when you're trying to wrangle the profuse code generation an LLM gives you into code that actually does what you want, strictness is gold. Here's how he's applying that. https://link.testdouble.com/c8f60d
Last week Double Agents Anya Iverova, Cathy Colliver, and Christine McCallum-Randalls attended All Things AI. Here's a field report from the sessions they attended. https://link.testdouble.com/786ef9
