Engineer In the Looop

What does it really take to turn bold ideas into real impact inside one of the world’s largest technology companies?

In this episode of Engineering in the Loop, Alec Harrison sits down with Taylor Black, Director of AI & Venture Ecosystems in the Office of the CTO at Microsoft, to unpack how internal incubators actually work — and why most innovation efforts fail before they ever ship.

Taylor leads Microsoft’s internal incubation studio, where early-stage, high-risk ideas are tested, validated, and scaled into products capable of generating hundreds of millions — and eventually billions — in revenue. Unlike traditional startups, these ventures must meet Microsoft-scale expectations while navigating enterprise constraints, long buying cycles, and strategic alignment across product groups.

In this conversation, we cover:

• What makes an idea “Microsoft-sized” (and why most aren’t)

• How internal incubators de-risk innovation before product teams invest

• Why $1B in revenue within five years is the bar — not the exception

• The biggest mistakes founders make when starting companies

• When not to take venture capital (and why most founders do it too early)

• How AI agents will reshape work far beyond chat interfaces

• Why the future may include one-person billion-dollar companies

Whether you’re an engineer, founder, product leader, or innovation executive, this episode offers a rare, inside look at how venture-style thinking works inside a global enterprise — and what you can learn from it.

In this episode Alec and Brian talk about AI, turkey, and AI turkey?

Alec welcomes Sakari Nahi, CEO of Zure, for a fun and thoughtful discussion that spans 25 years of tech evolution. Sakari shares how a single C# book jump-started his career, why he left a job he didn’t love to found a cloud-native consultancy, and what it’s like building a people-first engineering culture across multiple countries.

The two dig into real AI use cases that actually work today—vector search, customer service automation, field-tech knowledge retrieval—and explore how spec-driven development and tools like GitHub Copilot are transforming the way teams build software. They also get honest about shadow IT, geopolitics affecting cloud decisions, the future of Power Platform, and why AI feels “magical” even without AGI.

Whether you’re a developer, leader, or just AI-curious, this episode is packed with relatable stories and practical perspectives.

Cory House (Pluralsight/DomeTrain author and principal at ReactJS Consulting) shares the story of going “all-in” on JavaScript/React and how that focus grew into a successful independent consulting and training career. We dig into the tradeoffs of deep specialization vs breadth, how to spot real opportunities, and the “two-way door” idea for tech career moves. Cory also walks through his current pivot: using AI as a developer accelerator (how teams use it, where it helps most, what to watch out for) and how experimentation today — while tooling is cheap and rapidly evolving — is valuable. Along the way we surface mindset lessons (Cal Newport, Carol Dweck), how to balance giving away content vs paid courses, and practical tips for auditors/consultants trying to scale their impact.

Guest: Cory House — https://www.bitnative.com/

· Consulting & training: https://www.reactjsconsulting.com/

· Courses: Dometrain (TypeScript: Getting Started / Deep Dive) · YouTube: https://www.youtube.com/@housecor

· X: https://x.com/housecor

· GitHub: https://github.com/coryhouse

· DevOpsDays Des Moines (speaker): https://devopsdays.org/events/2025-des-moines/welcome/

· Podcast: https://eitl.ai/podcast/

· Books: So Good They Can’t Ignore You (Cal Newport), Mindset (Carol Dweck)

Alec Harrison and Brian Gorman return from a short break to talk all things certifications, Copilot, and the curious evolution of learning with AI. Alec shares his experience taking Microsoft’s new Applied Skills exams for Copilot Studio, while Brian gives some veteran insight into two decades of Microsoft certifications and how the new role-based system compares.

They debate whether AI tools are replacing junior engineers, discuss what makes modular Infrastructure as Code essential, and riff on the future of “speech-to-IaC” — where voice meets automation. Plus, Brian shares his own upcoming video course and gives pragmatic advice for anyone chasing their next cert.

👉 Explore Microsoft Learn’s Applied Skills here: https://learn.microsoft.com/credentials/applied-skills
🎧 Listen, comment, and tell us: Is modular Bicep overkill, or best practice?

#EngineerInTheLoop #Azure #AI #Copilot #MicrosoftLearn #Bicep #Terraform #Certifications

Founder of .NET Conf and 20-year Microsoft MVP Javier Lozano joins Engineer in the Loop to talk about what’s changed—and what hasn’t—in web development. We riff on MVP categories, why demos are easy and production is hard, and how AI is a force multiplier that still demands human judgment. Javier shares “SIMON” (Simplified Minutia and Operational Nonsense), his vision for agents that run conference ops, plus practical takes on trust, determinism, and the trade-off between faster output and more defects. We hit identity’s growing role, WebAssembly’s promise, and why tools (the picks and shovels) often win the gold rush. If you build .NET apps, run cloud workloads, or just want a grounded view of AI’s near future, this one’s packed with hard-earned lessons and optimistic realism.

Tech news for Oct 3rd, 2025https://github.com/github/copilot-clihttps://devblogs.microsoft.com/azure-sdk/azure-sdk-release-september-2025/https://techcommunity.microsoft.com/blog/securitycopilotblog/from-idea-to-security-copilot-agent-create-customize-and-deploy/4458516

From Exchange 2003 to Azure landing zones, Elliott shares 15 years of lessons on moving beyond VMs, building platform ops that devs actually love, and keeping cloud costs sane.

Key Takeaways

• Platform engineering > “just cloud”: Guardrails, vending, and templates win hearts (and roadmaps).

• FinOps framing: Translate infra to “cost per sale/visit” and budgets to anomaly alerts.

• PaaS by default: Avoid VMs unless there’s a clear reason.

• DX = adoption: The easier you make it, the faster devs ship safely.

• AI is force-multiplier, not autopilot: Keep humans in the loop for security, design, and intent.

Chapters

• 00:00 – Cold Open & Setup
  Welcome to Engineering the Loop. Alec introduces Elliott Leighton-Woodruff, Principal Architect at Synextra, and shouts out his pro studio setup.

• 01:18 – Origin Story: From Helldesk to Architect
  Elliott’s path through managed services, Exchange migrations, and into Azure—why his “O365 is only for SMBs” take aged badly and what that taught him about platform shifts.

• 03:00 – Cloud Era Pivot
  From AD/Exchange to Azure/Entra, early lift-and-shift vs. today’s PaaS-first mindset: “How do we avoid servers?” and drive ROI/security simultaneously.

• 04:32 – Compliance Across the Pond
  GDPR, UK/EU data residency, and pragmatic risk: why proper data handling matters more than headlines—and the nuance of Microsoft’s US metadata access.

• 09:08 – SaaS/PaaS > IaaS & Vendor Lock Worries
  Why “vendor lock-in” is often less risky than running your own racks—and how integrated ecosystems (Microsoft, Apple) win on user experience.

• 14:48 – IaC Choices in the Wild
  ARM → Terraform → Bicep (and back). When Terraform’s versatility wins, when Bicep is “good enough,” and how state files can be a superpower for diffing real changes.

• 17:49 – Standards & Modules
  Minimal code, shared module repos, and composable templates; using internal modules to enforce good defaults and speed delivery across clients.

• 21:10 – Developer Experience (DevEx) as the Sell
  Resource vending, guardrails, and starter pipelines so devs ship .NET without touching VNet/front door/APIM—but still stay within policy.

• 26:40 – Cost Control & Landing Zones 101
  What an Azure landing zone really is (governance + network + RBAC + policy). Budgets, quotas, and anomaly alerts to prevent “surprise bills.”

• 32:59 – Real Billing War Stories
  Costly misconfigs (AKS + Log Analytics, Custom Neural Voice hosting), forgiveness policies, and AWS vs. Azure leniency, plus why alerts matter.

• 38:19 – Spot & Batch = Cheap Compute
  Spot VMs and batch patterns for big workloads; tradeoffs and when to queue jobs for 80%+ savings.

• 41:45 – FinOps Mindset Shift
  Talk in cost per X (visit/transaction/customer) instead of monthly totals; why scaling beats fixed VMs when revenue is on the line.

• 49:50 – Agents, “Vibe Coding,” and Reality Checks
  AI can ship features, but humans still set direction and prevent face-palm security mistakes (like leaking waitlist emails via DevTools).

• 52:19 – The Junior Talent Question
  If agents do the grunt work, where do juniors learn? Potential futures for hiring, training, and the skills that stick.

• 55:49 – Hybrid, Edge, and HCI Use Cases
  Azure Stack/HCI examples (manufacturing/food QA) and the appeal of local inference with cloud aggregation.

• 57:39 – The (Maybe) Dystopian Future
  Meetings → summaries → agents; what stays human, what becomes automated—and why good platform ops multiply teams.

• 58:08 – How to Reach Elliott & Synextra
  Where to follow Elliott and Synextra; why they give away “do-it-yourself” playbooks and when to call them for the hard stuff.

• Platform engineering > “just cloud”: Guardrails, vending, and templates win hearts (and roadmaps).

• FinOps framing: Translate infra to “cost per sale/visit” and budgets to anomaly alerts.

• PaaS by default: Avoid VMs unless there’s a clear reason.

• DX = adoption: The easier you make it, the faster devs ship safely

• AI is force-multiplier, not autopilot: Keep humans in the loop for security, design, and intent

AI News For October 2nd

Sora 2 is out!

Perplexity Acquires AI Video Company

Microsoft Agent Framework

Here's a little life update and AI/Tech news for October 1st 2025

[Anthropic Joins Microsoft 365 Copilot](https://www.microsoft.com/en-us/microsoft-365/blog/2025/09/24/expanding-model-choice-in-microsoft-365-copilot/)

[Claude Sonnet 4.5 Is Out!](https://www.techradar.com/ai-platforms-assistants/claude/anthropics-claude-sonnet-4-5-is-available-now-the-best-ai-model-in-the-world-for-real-world-agents-coding-and-computer-use)

[Amazon unveils new generation of AI-powered Kindle and other devices](https://apnews.com/article/amazon-ai-alexa-kindle-echo-4e2983e166b8b62b2a3e51fc30a4a46c)