Send us Fan Mail

AI and software development - the Real Problem with AI-Driven Software Engineering.
AI is dramatically accelerating software delivery — but speed alone is not the answer.

In this episode of Serverless CrAIc, we explore how AI is reshaping software engineering, platform engineering, architecture, and organisational design.

As code generation becomes commoditised, the real differentiator is no longer how fast teams can build software — it’s whether they are building the right thing.

We discuss:

why clarity of purpose matters more than ever
how AI amplifies both good and bad engineering practices
the growing importance of socio-technical systems
platform engineering and cognitive load
why North Star metrics still matter
how engineering leaders should think about AI adoption
the risks of accelerating poor organisational decision making

If your organisation is adopting AI into software delivery, this conversation is essential listening.

Chapters

00:00 Introduction
01:42 AI is changing software engineering
05:18 Why building faster is not enough
09:34 The danger of accelerating bad decisions
14:27 Why clarity of purpose matters
18:40 AI as a commodity vs differentiator
24:05 Platform engineering and cognitive load
30:12 Socio-technical systems in the AI era

Resources

🌐 Website: The Serverless Edge https://theserverlessedge.com/
📘 The Value Flywheel Effect: https://itrevolution.com/product/the-value-flywheel-effect/#o5a04b7992465
🎧 Podcast Playlist: Serverless CrAIc Playlist https://open.spotify.com/show/5LvFaitkSkg2q5MWqKLrXu
📰 Newsletter: The Serverless Edge on LinkedIN https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7066788643985596416

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

Why Team Topologies Matters More Than Ever in the AI Era.
Are AI agents changing how software teams should be structured?

In this episode of Serverless CrAIc, David Anderson, Mark McCann, and Michael O’Reilly explore one of the biggest questions emerging in the AI era:

👉 Does Team Topologies still matter when AI agents can generate code, tests, and workflows at incredible speed?

The discussion dives deep into:

Cognitive load in AI-driven engineering teams
Socio-technical systems and AI adoption
Why human collaboration still matters
Stream-aligned teams in an agentic world
The evolving role of platform teams
Why enabling teams are more important than ever
AI agents as “team members” — myth or reality?
How engineering organisations scale safely with AI
Why guardrails, standards, and architecture matter more now
The balance between autonomy and control in AI-enabled organisations

One key theme runs throughout the conversation:

AI may accelerate software delivery — but the human systems around software are still critical.

As development speeds increase, organisations must rethink:

collaboration
communication
cognitive load
organisational design
engineering enablement
platform strategy
operational excellence

This is a must-watch discussion for engineering leaders, architects, platform teams, and anyone building AI-enabled software organisations.

Chapters

00:00 – Introduction
00:23 – AI, socio-technical systems, and Team Topologies
01:02 – Why cognitive load matters more in the AI era
02:07 – Drinking from the AI fire hose
03:20 – Shifting cognition from code to outcomes
04:32 – Why engineers are moving higher up the value chain
05:48 – DP1 vs DP2 organisational design principles
07:15 – Autonomy, mastery, and purpose in AI teams
08:50 – Are AI agents team members?
10:45 – Agent orchestration and organisational principles
11:44 – Why AI is not truly a “team member”
13:09 – Can you really pair program with AI?
13:52 – Stream-aligned teams in an AI world
15:34 – Jevons Paradox and accelerating software delivery
17:11 – The changing role of platform teams
18:46 – Security, governance, and AI platforms
20:31 – Why platform teams must stay ahead
21:08 – The critical role of enabling teams
22:32 – Coaching engineers to work effectively with agents
23:23 – AI anti-patterns and “We Jimmy” chaos engineering
24:54 – Complicated subsystem teams and deep expertise
27:20 – Does Team Topologies still matter?
28:06 – Constraints, guardrails, and organisational design
28:39 – Closing thoughts

Resources & References

📘 Books & Concepts Mentioned

Team Topologies — Matthew Skelton & Manuel Pais
Cognitive Load Theory
Socio-Technical Systems
Team Design Interaction Modes
Stream-Aligned Teams
Platform Teams
Enabling Teams
Complicated Subsystem Teams
Cynefin Framework
Jevons Paradox
Well-Architected Systems
AI Agent Orchestration

📚 Key Themes

AI engineering teams
Organisational design
AI agents and workflows
Platform engineering
Developer productivity
AI adoption
Engineering leadership
Team structures in AI
Guardrails and governance
Human + AI collaboration

🌐 Learn more:
https://theserverlessedge.com

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

Is AI-generated code creating more value — or more liability?

In this episode of Serverless Craic, David Anderson, Mark McCann, and Michael O’Reilly explore why one of software engineering’s oldest principles is suddenly more relevant than ever in the age of AI:

“Code is a liability. The system is the asset.”

As agentic AI and code generation tools accelerate development, teams are producing more code, more tests, and more complexity than ever before.

But:

Does more code actually mean better outcomes?
Are organisations creating massive technical debt without realising it?
What happens when AI accelerates poor engineering practices?
And how do you maintain confidence, security, and quality in probabilistic systems?

This episode explores:

AI-generated code and technical debt
Validation, verification, and testing strategies
Observability and evaluation frameworks
Security vulnerabilities and unmanaged code
Critical thinking in modern software engineering
Why “lines of code” ≠ business value
The return of XP and foundational engineering principles
Chapters

00:00 – Introduction
00:24 – Old engineering principles returning in the AI era
00:51 – The return of Extreme Programming (XP)
01:43 – “Code is a liability” explained
02:46 – AI-generated code and growing technical debt
03:32 – Why engineers must review AI-generated code carefully
05:16 – The history of generated code and technical debt
06:28 – Why more code doesn’t mean more value
07:12 – AI hype, supply chains, and unmanaged complexity
08:30 – AI accelerates weak engineering practices
09:02 – Why teams still struggle with testing strategies
10:39 – Observability and deploying with confidence
11:54 – Evaluation frameworks for probabilistic systems
12:55 – System boundaries and verification
13:11 – Engineers are still accountable for AI-generated code
13:50 – Critical thinking in probabilistic systems
14:44 – Security vulnerabilities and unmanaged legacy code
16:27 – Commodity systems vs unnecessary custom code
17:25 – AI models finding security vulnerabilities
18:38 – Exploration, testing, and security charters
20:02 – Why code liability matters more than ever
20:24 – Engineering excellence as competitive advantage
20:44 – Final thoughts

Resources & References

📘 Concepts & People Mentioned

Ward Cunningham — Technical Debt
Kent Beck — Extreme Programming (XP)
Dave Farley — Continuous Delivery & modifiable systems
Dan North — Engineering practices & architecture
Elizabeth Hendrickson — “Testing = Checking + Exploring”

📚 Topics Discussed

Technical debt
AI-generated code
Agentic AI workflows
Evaluation frameworks (evals)
Observability
Continuous verification
Security scanning
Probabilistic systems
Platform engineering
Serverless architecture
Engineering excellence

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

Psychological Safety in the AI Era: AI is moving so fast it’s not just changing how we build software — it’s changing how teams think, learn, and work together.

But there’s a problem no one is talking about enough:

What happens to psychological safety when everything is changing at once?

In this episode of Serverless CrAIc, Dave Anderson, Mark McCann, and Michael O’Reilly explore the human side of the AI revolution — from hype cycles and uncertainty to leadership, learning, and team dynamics.

Because while AI is accelerating engineering, it’s also:

Creating pressure to “keep up”
Challenging confidence and expertise
Shifting how teams collaborate and make decisions

And without psychological safety, teams won’t question, won’t challenge — and won’t build well.

“It’s psychologically exhausting trying to keep up with the pace of change.”

This is a conversation about what it really takes to build high-performing, resilient teams in the AI era.

Chapters

00:00 – Welcome to Serverless CrAIc
AI hype, rapid change, and keeping up

00:31 – Why psychological safety matters in the AI era
The difficulty of challenging AI in organisations

02:02 – The most aggressive hype cycle we’ve seen?
Comparing AI to cloud and previous tech shifts

03:25 – The turning point in AI capability
From hype to real engineering impact

04:17 – The psychological impact on engineers
Why the pace of change is exhausting

04:49 – Innovation vs standards
Why too much structure too early can slow teams down

05:37 – The four stages of psychological safety
From inclusion to challenger safety

07:01 – The capacity problem
Why senior engineers are struggling to mentor while learning themselves

07:38 – Sense-making in fast-moving environments
How experienced engineers are adapting

09:00 – What skills matter now?
Growth mindset, experimentation, and adaptability

10:45 – Bias for action
Why experimenting with AI tools is critical

11:59 – Vulnerability, empathy, and humility
Key leadership traits in uncertain times

13:23 – Confidence in core engineering skills
Why experience still matters

13:58 – Demand isn’t slowing down
Why engineers are busier than ever

15:15 – AI and engineering standards
Applying world-class practices faster than ever

16:09 – Final thoughts
Psychological safety as a leadership priority

Key Themes
Psychological safety in high-change environments
AI hype vs reality in engineering teams
The impact of rapid change on confidence and learning
Leadership challenges in AI-driven organisations
Growth mindset, experimentation, and vulnerability
Applying high engineering standards with AI
Resources & References

Concepts and ideas mentioned in the discussion:

Psychological Safety (Amy Edmondson – The Fearless Organization)
Four Stages of Psychological Safety (Mutual respect → Challenger safety)
Growth vs Fixed Mindset (Carol Dweck)
Bias for Action (engineering and product principle)
Well-Architected Frameworks (cloud and serverless design principles)
Event-driven and serverless architectures

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

AI is dramatically increasing the speed at which teams can build software. But if you can ship features in hours instead of months, a new problem emerges:
How do you know you’re building the right thing?

In this episode of Serverless CrAIc, Dave Anderson, Mark McCann, and Michael O’Reilly explore why clarity of purpose and a strong North Star are more important than ever in an AI-accelerated world.

As AI tools and agentic systems remove friction from development, teams can prototype, build, and deploy faster than ever before. But without clear direction, that speed can quickly turn into chaos, feature overload, and wasted effort.

We discuss:
Why the North Star framework still matters in the AI era
The importance of leading vs lagging metrics
How observability and telemetry support decision-making
Why product management and engineering roles are shifting
The growing need for product-oriented engineering teams
If AI increases your delivery velocity, your strategy and decision-making must evolve just as quickly.

Chapters

00:00 – Welcome to Serverless Craic
AI everywhere and the coming singularity (maybe).
00:31 – Does the North Star still matter in the AI era?
Why clarity of purpose becomes even more critical when you can build faster.
01:30 – Why speed without direction is dangerous
How AI can lead teams to build the wrong things faster.
02:20 – Experience as an advantage in the AI era
Why experienced engineers ask better questions of AI systems.
03:06 – The first North Star question: What game are you playing?
Defining your problem space before building anything.
04:29 – Rapid experimentation with AI prototypes
Using AI-driven prototyping to discover meaningful product signals.
05:41 – Observability hasn’t changed
Why understanding what to measure is still the hardest problem.
06:32 – Leading vs lagging metrics
How telemetry and instrumentation help teams track progress.
07:53 – The shift toward systems thinking
Why engineers increasingly need a systems engineering mindset.
08:29 – Product management pressure in the AI era
The growing importance of solving real customer problems.
09:27 – Wardley mapping, user needs, and rapid iteration
Why product strategy becomes more important as teams move faster.
10:38 – The danger of overwhelming users with features
Understanding organisational and user adoption limits.
11:01 – Decision-making speed in large organisations
Why strategic decisions must flow faster through organisations.
11:55 – Engineering teams becoming product teams
Autonomy and product ownership in high-velocity environments.
12:48 – Making good decisions against the North Star
Why strong leadership and judgement still matter.

Resources & References
North Star Framework – aligning teams around a single product metric
Leading vs Lagging Metrics – measuring immediate vs long-term outcomes
Observability in modern systems – instrumentation and telemetry
The Build Trap (concept discussed by product leadership thinkers)
Wardley Mapping – understanding user needs and strategic positioning
DORA Metrics – measuring engineering delivery performance

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

🎙 AI: Commodity or Differentiator? | The Value Flywheel in 2026

AI has taken a step change. Over the past few months, adoption has accelerated dramatically — but are organisations applying it with clarity, or just chasing hype?

In this episode of The Serverless Edge, we unpack:

AI as a commodity vs differentiator
Why clarity of purpose matters more than ever
The risks of “vibe coding” critical systems
Security, blast radius, and agent containment
Accelerating your Value Flywheel safely with AI
Why you cannot outsource critical thinking to an LLM

If you're a CTO, architect, or engineering leader trying to navigate AI adoption without introducing systemic risk — this conversation is for you.

⏱ Chapter Markers

00:00 – The AI Step Change: What Happened in Late 2025?
02:00 – AI as Commodity vs Differentiator
03:10 – The Cost of Building What’s Already a Commodity
04:20 – Internal Acceleration vs Product Features
05:30 – Training Data, Sovereignty & Enterprise Risk
06:45 – Where AI Becomes Dangerous in Your Organisation
08:00 – Agentic AI & Blast Radius: Why Containment Matters
09:20 – Competing With the Platform Providers
11:10 – SaaS Killed by AI? The AWS Reinvent Effect
13:00 – “Vibe Coding” Core Business Systems (And Why That’s Madness)
15:00 – Where You Shouldn’t Experiment With AI
16:00 – Faster Feedback Loops & Engineering Throughput
17:30 – Discipline, Metrics & the North Star
18:20 – Using AI to Improve Your Own Thinking
19:00 – Context Is Everything (And Harder Than You Think)
20:10 – Organisational Design for Humans and Agents
21:00 – Turning the Flywheel Before Adding AI
21:50 – Why Sitting on the Sidelines Isn’t an Option

🔎 Key Themes
1. Clarity Before Capability

Most organisations should consume AI, not build foundational models. The differentiator is rarely the LLM itself — it’s how clearly you understand:

Your user needs
Your value chain
Your supply chain dependencies
Your regulatory boundaries
Without that clarity, AI simply accelerates confusion.

2. Blast Radius & Containment

Agentic systems introduce a new risk profile.
If you deploy AI into:
Poorly governed SDLC environments
Weakly defined security domains
Legacy operational processes
You expand blast radius unintentionally.
Think SaaS isolation. Think sandboxing. Think containment by design.

3. Speed Changes Everything

If AI compresses delivery cycles from weeks to hours:
Your product discovery loop must accelerate
Decision-making must tighten
Metrics must be explicit
Engineering must sit inside the feedback loop
AI increases velocity — but velocity without direction is chaos.

4. You Cannot Outsource Critical Thinking

AI can help:
Refine your North Star
Improve impact mapping
Sharpen KPIs
Draft Wardley Maps
Analyse value chains
But it cannot replace:
Context
Judgement
Organisational alignment
Strategic trade-offs

You still need to do the hard yards.

📚 Related Concepts & Resources

The Value Flywheel Effect
Wardley Mapping
North Star Framework
Impact Mapping
SDLC Acceleration
Agentic AI Architecture
Blast Radius & SaaS Isolation Models

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

AI Myths in Software Engineering. AI is colliding with software engineering at full speed — and a lot of myths are emerging along the way.

In this episode of Serverless Craic, Dave Anderson, Mark McCann, and Michael O’Reilly unpack how AI, GenAI, and agentic systems intersect with the ideas behind The Value Flywheel Effect. Rather than hype or fear, this is a grounded engineering discussion about quality, responsibility and long-term value.

We explore six common myths about AI and software engineering — and why good engineering judgement, domain knowledge, and clarity of purpose matter more than ever.

If you care about building sustainable systems, not just shipping demos, this one’s for you.

Chapters

00:00 – Welcome & context
Why AI + serverless + the Value Flywheel is colliding right now

01:50 – Myth 1: “Software engineering is dead”
Why engineering skills are more valuable, not less

07:06 – Myth 2: “My skills will become irrelevant”
Moving up the value chain, domain expertise, and growth mindset

13:40 – Myth 3: “The quality isn’t good enough”
Standards, constraints, and why worst it’ll ever be is today

18:44 – Myth 4: “The model understands the problem”
Pattern matching vs understanding, context, and critical thinking

24:20 – Myth 5: “I’ll be forced to use AI”
Workflows, guardrails, security, and excessive privileges

31:54 – Myth 6: “We’ll need fewer engineers”
Jevons Paradox, lowered barriers, and the coming demand explosion

34:22 – Closing thoughts
AI, velocity, and the future of sustainable software engineering

Key Themes Discussed

AI as an abstraction layer, not a replacement for engineering
Why standards, constraints, and operability still matter
Domain-Driven Design as AI-amplifying, not obsolete
Agentic systems, skills, prompts, and containment
Security risks: excessive privileges & supply-chain concerns
Velocity vs sustainability in AI-assisted development

Resources & References

The Value Flywheel Effect – principles referenced throughout
Wardley Mapping & situational awareness
Domain-Driven Design (DDD)
OWASP Top 10 for LLMs (excessive privileges, agent risks)
Jevons Paradox (efficiency driving increased demand)
Early cloud cost & governance parallels
Threat modelling for AI and agentic systems

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube

Send us Fan Mail

In the first Serverless Craic episode of 2026, Dave Anderson, Mark McCann, and Michael O’Reilly reflect on a five-year journey that began in early 2021 with the idea for The Value Flywheel Effect.

This episode closes out the book series by looking back—warts and all—at what it really took to write, publish, promote, and apply the ideas in practice. The conversation spans writing fatigue, editing realities, imposter syndrome, enterprise adoption, and why the flywheel is arguably more relevant than ever in an AI-first world.

If you care about modern software delivery, cloud strategy, serverless-first thinking, and leading technology change, this one is for you.

⏱️ Chapters

00:00 – Welcome & context
Episode 79, first show of 2026, and closing out the book series

01:20 – How the book started (2021 → 2026)
From an idea to a five-year journey

01:35 – Did we enjoy writing the book?
Ideation, Guinness-fuelled drafts, and the reality of writing

02:30 – Shaping the narrative
Why writing is harder than it looks, and why shared context doesn’t scale

04:10 – Atomic essays & capturing thinking early
GitHub, short-form writing, and building habits

05:00 – Would this book have helped us 15 years ago?
Modernisation gaps, agile limits, and why the flywheel mattered

06:00 – The editing process (and thick skin)
What professional editors really do to your manuscript

07:40 – Feedback, criticism, and author psychology
Why the one negative comment sticks

09:15 – Has the book made an impact?
Enterprises, conferences, and unexpected adoption stories

12:40 – Applying the flywheel in real organisations
North Stars, Team Topologies, serverless-first in practice

14:30 – The hardest part of the whole journey
Finishing, introductions, and the truth about selling a book

17:30 – Promotion, modesty, and imposter syndrome
Why marketing a book is a full-time job

19:00 – Influences & supporters
Kent Beck, Adrian Cockcroft, Simon Wardley, and standing on shoulders

21:45 – Is the book still relevant in the age of GenAI?
Why the flywheel + AI is a force multiplier

23:00 – AI, context engineering, and agentic systems
Using codified principles to guide AI effectively

25:30 – Lowering the barrier to good practice
How AI helps teams apply architecture, security, and governance

29:00 – Business strategy vs technical strategy
Is the divide finally disappearing?

31:45 – The emerging “builder” persona
Shifting left, shifting right, and new SDLC realities

34:50 – Closing thoughts & what’s next
AISDLCs, Brownfield challenges, and future episodes

📚 Resources & Links

📘 The Value Flywheel Effect — principles for modern cloud and serverless transformation

🌐 The Serverless Edge: https://theserverlessedge.com

🎥 Subscribe on YouTube for weekly Serverless Crack episodes

💬 Follow the conversation on LinkedIn

💡 Key Takeaways

Writing a book is much harder than most engineers expect

The flywheel was never about tech or business—it’s about both

AI makes codified principles (North Stars, well-architected practices) more valuable, not less

Critical thinking remains non-negotiable, even with powerful models

Context is now a first-class architectural concern

👍 If you found this useful, like, subscribe, and share with your team.
💬 Let us know in the comments how you are applying the flywheel—or where it’s challenged you.

Cheers,
The Serverless Edge

Serverless CrAIc from The Serverless Edge
Check out our book The Value Flywheel Effect
Follow us on X @ServerlessEdge
Follow us on LinkedIn
Subscribe on YouTube