Mixing oil & water?
That’s the challenge of modern #SoftwareArchitecture: Deterministic Systems must coexist with Non-Deterministic AI.
In this #InfoQ article, Shweta Vohra introduces the Architect’s V Impact Canvas- a framework designed to keep human trust at the center of AI integration.
The 3 core layers: 🎯 Architectural Intent ⚖️ Design Governance 📈 Impact & Value
➡️ Read the deep dive here: https://bit.ly/4blYGkR
Hot take: If 80% of a system consists of “special cases”, they are not special cases, they _are_ the system, and it could be much improved if the people working on it acknowledged that fact and acted accordingly.
What does it really mean to embrace #AI in #SoftwareEngineering?
In this #InfoQ #podcast, Ben Greene explores why technical excellence is no longer enough!
The new priority: deeply understanding business context & prioritizing human empathy in increasingly automated systems.
🎧 Listen now: https://bit.ly/4qKxrWJ
📄 #transcript included
The Semantic Irony of SAFe - a thread 🧵
There is a profound irony in the name "SAFe" (Scaled Agile Framework). In a truly agile system, the first thing you have to accept is that nothing is safe. Agility is our response to uncertainty, complexity, and risk.
#WorkFeedbackLoop #Agile #SAFe #SystemThinking #SoftwareEngineering #Management
(1/6)
We need to talk about the cult of "staying current."
I often see teams pushing for major framework or library upgrades without being able to name a single friction point, security risk, or performance bottleneck they are actually solving. When asked for the "Why," the answer is usually a shrug: "It’s the latest version."
A thread 🧵
#WorkFeedbackLoop #SystemThinking #SoftwareEngineering #Actionism #TechDebt #EngineeringManagement
(1/4)
At what point does automation stop being an advantage and start exposing your system's deepest flaws?
Right now, we are obsessed with "efficiency" in the wrong places. We automate reports, summaries, and dashboards, thinking we are accelerating our work.
But we might just be making our stagnation more visible.
A thread 🧵
#WorkFeedbackLoop #SystemThinking #DecisionLatency #Agile #SoftwareEngineering #Governance (1/7)
I learned the hard way that words can be weapons.
(No trauma dump, you're good.)
Language is powerful.
Or as I like to say:
Language is load-bearing.
Words have real somatic impact.
They change nervous system state.
And yet tech treats it as.. vibes.
(Words make the body go brrr.)
What happens when language is treated with the same rigor as code?
When nervous system cost is not an abstract idea but part of the equation?
Which communicational failure modes become myths of the past?
(Static types for human language when?)
Imagine people asking:
"Why do you care about language so much?"
And the answer being:
"Because it works."
systemic.engineering explores what happens
when language stops being vibes
and is treated as infrastructure.
Because that's what language is:
The transport layer between divergent realities.
(Be they human or agentic.)
Intruiged?
Members get practice-oriented advice.
No bullshit. Unsubscribe any time.
👉 https://systemic.engineering/welcome
#SystemicEngineering #SRE #SREforHumans #Language #LanguageAsInfrastructure #Communication #CommunicationProblems #FailureModes #SystemThinking #EpistemicAbuse #Leadership #Regulation
"A “user” is, reductively, someone who uses. An “actor” is someone who acts on a system, or, even better, interacts with that system."
“We already have feedback.”
That’s often where learning stops.
Most orgs don’t lack feedback — they lack the ability to act on it. Surveys, metrics, incidents… and no change.
Feedback is only feedback if it changes the next piece of work. If change is slow, risky, or expensive, systems ignore signals.
Feedback problems are work problems.
Work–Feedback Loop:
Learning speed = response speed.
InfoQ
Sebastian Hans
Eckhard Lohse – Durchdenker
Thomas
Jason CranfordTeague