Joined October 2011
- Tweets 1,405
- Following 221
- Followers 2,401
- Likes 637
330 Photos and videos
Jun 12
In a conversation with Traingent CTO Wayne Marley about building Ride with Dave, Principal Consultant @dmosher talked about what's happening by shortening the loop in product development with agentic coding.
youtube.com/watch?v=w_JFsFZ3…
1
78
Jun 11
Software becomes legacy by succeeding long enough to accumulate the weight of every decision and shortcut made along the way. AI is changing legacy modernization, but not everything is getting better.
Read the full article from @toddkaufman link.testdouble.com/caa595
1
44
Jun 10
Teams actually shipping AI products build a system that keeps agents honest with humans in charge. Principal Consultant Dave Mosher talked with Traingent CTO Wayne Marley about how this shaped the way Ride with Dave was built.
link.testdouble.com/7c712c
53
Jun 9
Facing a lack of standards and capabilities in the industry, Pelagic Research Services developed mission-critical emergency software for the ocean's final frontier in partnership with Test Double.
Read the full case study: link.testdouble.com/efe707
40
Jun 8
There are a lot of good ways to bring research, planning, and implementation structure to AI coding tools. Han was built by River Bailey for people who would rather pick their own path than ride someone else's track.
link.testdouble.com/6ebbb9
86
Jun 3
The real prerequisite for effective AI at work isn't more autonomy. It's organizational observability: making intent visible enough for humans and agents to act well. link.testdouble.com/48c645
42
Jun 2
We're hiring a Principal Product Consultant at Test Double! We're seeking an experienced product leader who can partner with clients on complex product challenges, shape product strategy, and drive meaningful business outcomes.
Apply: testdouble.hire.trakstar.com…
55
May 13
Claude Code permission fatigue is real, but sandboxing can help. A practical (and illustrated) guide from @_joedupuis to Seatbelt on macOS and Bubblewrap on Linux, plus an allow/kill/push heuristic for the prompts that remain: link.testdouble.com/243e75
1
135
May 12
A 1000-line, 55K-character setup guide loaded as one chunk = instant context compaction and a hallucinating wizard. Six lessons from @elpapapollo while building a Claude skill that walks devs through onboarding without burning tokens:
link.testdouble.com/9be652
97
May 12
Software becomes legacy by succeeding long enough to accumulate the weight of every decision and shortcut made along the way. AI is changing legacy modernization, but not everything is getting better.
Read the full article from @toddkaufman link.testdouble.com/caa595
1
223
May 11
AI doesn't have an accuracy problem. It has a target problem. Given/When/Then ACs are the clearest target you can hand a coding agent. Andy Vida adapted Three Amigos for solo devs working with AI: two design docs, one structured conversation, then code:
link.testdouble.com/a04ba1
1
131
May 7
You committed the two-line fix. The 40 minutes of investigation that produced it went nowhere. The next person (or agent) pays the same cost again. A TDD-shaped loop from Rick Reilly for capturing context across agent sessions: link.testdouble.com/11acb4
69
May 6
The real prerequisite for effective AI at work isn't more autonomy. It's organizational observability: making intent visible enough for humans and agents to act well. link.testdouble.com/48c645
1
65
May 5
Every abstraction we've built since assembly was a subsidy paid to human cognition, funded by hardware. Take humans out of the inner loop and the rationale weakens.
@docondev shares a prediction about what languages and frameworks look like next:
link.testdouble.com/ae23c6
1
1
189
May 4
Cyclomatic complexity, method length, naming conventions: every one of those metrics was a workaround for human cognitive limits. Agents don't have those limits. So what survives, and what gets renegotiated? @docondev shares reflections on craft:
link.testdouble.com/fdff50
1
2
207
Apr 30
PC shift: ~10 years.
Web shift: ~5 years.
Agentic coding: 2-3, possibly less.
The first two expanded the profession. This one compresses it. @docondev shares thoughts on the pattern, and where the time actually goes: link.testdouble.com/aafa83
1
2
215
Apr 28
AI changed the cost structure of software. Outputs may be cheap, but outcomes are still valuable. @dmosher says leverage has moved to the harness: lint rules, types, tests, ADRs, feedback loops. link.testdouble.com/ecfcc5
1
1
130
Apr 27
Code coverage tells you a line ran. Mutation testing tells you whether your tests would notice if it stopped working. With agent-generated code, that distinction matters more than ever. @neall shows how Stryker keeps coding agents honest: link.testdouble.com/d689d0
1
126
Apr 23
If you missed our Coffee Time session yesterday about AI harnesses and workflows you can catch up with the recording. Pour yourself a cup of coffee and listen in.
youtu.be/grPAIsIgTQY
2
157
5 Dec 2025
Stuck on a tough challenge?
➡️ Debugging a gnarly codebase?
➡️ Wondering whether to do a rewrite or refactor your legacy system?
➡️ Evaluating if your product idea is solving the right problem?
We've got you covered: free, no strings attached office hours
link.testdouble.com/58w
1
243