Hangar DX Podcast

“AI is producing more than we could ever imagine before, but we’re still defining confidence in the same old way that we’ve been doing it for decades. And it’s causing a lot of problems.” 

On HangarDX, Ankit Jain talks with Jay Paulson, Director of Engineering at Symetra, about engineering confidence in the AI era and why the business still sees engineering as a black box. They talk about: 

• Why AI raised expectations while the confidence behind delivery is still based on belief, not evidence, and
• Why a team told it’s going too fast should slow down, and what to do with the extra capacity, from juniors writing code by hand to giving people time back, and
• How asking for delivery evidence and forward visibility changes the conversation with leadership

  Chapters
  00:00 Introduction to the episode and guest
  01:24 Jay Paulson's background and focus on engineering confidence
  03:25 What is engineering confidence? Definitions and examples
  05:01 The role of metrics and belief in confidence
  06:20 AI's influence on productivity and confidence
  08:48 Grounded confidence versus belief-based confidence
  11:15 When does engineering confidence meet expectations?
  14:52 The importance of evidence in planning and estimation
  17:40 The pressure to do more with AI and its effects
  20:23 The idea of slowing down to improve quality
  26:52 The risks of belief-based planning and confidence
  28:36 Building more grounded confidence through evidence
  32:18 Using visibility to manage delivery and expectations

  
  Replace code reviews with verified intent
  AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

  📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

  ✏️ Subscribe for more videos: @Aviator-Co

  🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

"I feel this is a non-solution to a non-problem that is being given so much credence for very dubious reasons, and in general, it will be harmful.”

On HangarDX, Ankit Jain talks with Ray Myers, a software developer and coding-agent expert, about why the push toward dark software factories, where agents write the code and no one reads it, is the wrong goal. They talk about:

- Why dark factories are a reboot of vibe coding that still breaks down under real scale, risk, and change over time, and
- Why you can get most of the value of dark factories from continuous delivery practices teams have taught for a decade, and
- Why people who are telling you not to read code are actually reading the code

Chapters

00:00 Introduction to Ray Myers and the concept of dark factories
04:09 What is a dark factory and its hypothetical nature
07:12 Scale, risk, and change in enterprise software
11:48 Why Ray Myers is anti-dark factory
13:27 The importance of reading code and guardrails
17:23 Leveraging guardrails and automated checks
22:25 The role of verification systems and revealing code
33:21 The limits of automation and the analogy of self-driving cars
36:51 The future of AI, automation, and human oversight in software development

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

"We just pretend to do code reviews. I'm wading through this large PR, and I try to understand what's going on and be smart, but the fact is that I don't really know what's going on."

In this episode of the HangarDX podcast, Ankit Jain, co-founder and CEO of Aviator, talks to Dr. Michaela Greiler, software engineering researcher and consultant, about:

What code review surrender looks like on teams that never actually decided to stop reviewing
How her SCOPE model moves oversight earlier, into planning, and scales it by risk, alignment, and understanding, and
Why an engineer saying "I'm shipping code I don't understand" is a green flag for a team?

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

"My working hypothesis is that human beings are too slow when dealing with code. If we're going to move at the speed of the agents, we're all going to have to back off a little bit and maintain high-level supervision."

On this episode of HangarDX, Ankit Jain sits down with Robert C. Martin (Uncle Bob), co-author of the Agile Manifesto and author of Clean Code, to talk about:

Why he stopped reading code, and the constraints he uses instead: unit tests, CRAP scores, mutation testing, and Gherkin,
Why code review is slowing teams down, and why only humans can be responsible for what agents produce, and
How a single engineer may now do the work of a team, and why Clean Code principles still apply.

Note: At 17:05 Uncle Bob starts sharing his screen, so we recommend watching a video version of this episode to tget the full experience: https://youtu.be/5kCISBJwoZo

Chapters
00:00 Introduction
01:48 How Uncle Bob codes today
03:12 What changed his mind about AI coding
04:52 Three constraints: unit tests, CRAP, and mutation testing
08:06 Aligning tests with intent
09:38 Gherkin and human-authored specs
11:00 Human judgment as the final arbiter
12:52 Small functions, readability, and complexity
13:20 Relaxing constraints as agents improve
13:54 Code written for agents, not humans
15:10 The future of code review and software teams
17:55 A dynamic UML tool for reviewing agent output
21:09 Architecture diagrams as a review surface
23:05 Disciplines every team needs
25:25 Why harness engineering is overrated
26:37 From harness to a single trusted agent
27:59 Are software factories doomed?
28:30 The AI exponential curve and its physical limits
29:30 Future optimizations and quantum computing
29:58 Responsibility and ownership of AI-generated code
32:26 Rewriting Clean Code for the AI era
33:30 Why Clean Code principles still apply

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

David Curlewis, Senior Engineering Manager at Canva, joins Ankit Jain on HangarDX to talk about what happens to platform teams when AI gives every developer a bulldozer.

Why platform teams should go down a layer and offer primitives instead of full highways, and
How inner sourcing and a graduation model help platform teams scale without fighting their own developers, about
What happens when vibe-coded dashboards go mission-critical overnight, and why outcome metrics beat volume?

Chapters

00:00 Welcome to Hangar DX
01:03 David’s platform engineering background
02:13 Why AI is changing platform teams
04:04 The golden path versus AI-built roads
07:40 Platform primitives: observability, security, and MCPs
11:46 Why platform engineering is becoming a core company function
14:56 Centralization, autonomy, and maturity
20:51 Inner source as a model for shared ownership
23:17 Designing contribution guidelines for AI agents
26:50 Intent debt, cognitive debt, and keeping specifications aligned
31:00 Why companies are building their own AI platforms
32:25 Moving from code review to reliable validation
34:50 Measuring AI productivity beyond lines of code
40:06 The return of generalist engineers

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

"The tools are so powerful that people actually want to manage them themselves. Everything is just moving up a tier of thinking.'"

Chris Hodges and Kevin Altschuler from Dropbox's Agentic Experience team join Ankit Jain on the HangarDX podcast to talk about why they built Nova, their internal platform for AI coding agents, instead of buying off-the-shelf. 

They also discuss:
Why owning the harness gives you leverage over cost, model selection, and the total developer experience, 
What makes Nova a platform rather than a point solution?
What it takes to get security and privacy teams on board when agents touch your most sensitive data
Where is AI engineering headed next? Agents that watch deployments, monitor logs, and message you when something's wrong

00:00 Introduction to Developer Experience and Nova
01:57 Overview of Nova: A New Development Platform
03:57 Challenges Faced by Developers Before Nova
08:34 Metrics for Success: Tracking Nova's Adoption
10:00 Capabilities of Nova: Tackling Toil and Flaky Tests
13:25 Comparing Nova with Existing Cloud-Based Tools
16:28 Nova as a Platform vs. Point Solutions
18:36 Experimentation with Goose and Future Directions
26:35 Building Reliable AI Systems
31:32 The Role of Orchestrators in AI Development
35:07 Future Challenges in AI Development

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

"It's not 'Do I read or not read the code.' It's "Does my attention get directed to where it's actually useful?'" 

David Poll, head of GitHub's Code & Review organization, joins Ankit Jain on the HangarDX podcast to talk about what code review actually does and why it was never really about catching bugs.

Why code review is really a taste problem, not a correctness problem, and
How PRDs, architecture docs, and design reviews are collapsing into the code review process,
Why we might be headed toward a world with 20x more unmerged PRs,
How teams can extract review decisions into institutional memory to beat AI slop with AI, about
The sliding scale of not reading code and what determines where your team should land

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

"As an industry, I feel like we've lost our minds. Teams are throwing away everything they've learned over the last many decades about how to effectively measure our engineering experiences. And we're not even just throwing it away. We're making negative progress."

Brian Houck, applied scientist at DX and co-author of the SPACE framework, joins Ankit Jain on the HangarDX podcast to break down what's actually happening when engineering organizations try to measure AI's impact.

Why token usage is the modern equivalent of measuring lines of code, and
What his research across eight dimensions of code quality actually found,
How calendar cleanup beat AI in PR throughput gains by 2x
Why culture—not tech stack—predicts which organizations get the most from AI, about
The case for "agent experience" as the next frontier of developer productivity

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

Vanitha Kumar, Market Technology Director at Thoughtworks, joins Ankit Jain to talk about how code review holds up when agents are writing most of the code.

In this episode, they get into:

Why pre-integration is no longer the only place review should happen, and
How teams decide what's even worth reviewing when a feature spans twenty markdown files, and
What shifting review "left" into a teaching moment looks like in practice, plus

Chapters

00:00 Introduction to Code Reviews
02:50 Evolution of Code Review Practices
05:46 The Role of Collaboration in Code Reviews
08:24 Cultural Shifts in Code Review Practices
11:12 Harness Engineering and AI in Code Reviews
16:37 The Future of Code Reviews and AI Integration
22:05 Platform Engineering in the AI Era

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/

Frances Coronel is a senior software engineer on the agentic SDLC team at Slack, where she builds AI-powered developer tooling and helps enable engineers and non-engineers alike in the agentic era.

In this episode of The Hangar DX podcast,, Ankit Jain, CEO of Aviator, talks to Frances about :
How Slack is shifting from solo, one-to-one development to multiplayer collaboration with agents
Why overindexing on sharing custom setups is more valuable than formal enablement programs
The PR review crisis that comes when your merge volume doubles and your review process doesn't change

Chapters

00:00 Introduction to DevXP at Slack
01:29 Understanding DevXP and Its Evolution
06:43 The Shift to Multiplayer Collaboration
09:57 Internal Workshops and Upskilling for AI Tools
14:03 The Dashboard for Tool Discovery and Usage
18:51 Tokenmaxxing and Its Implications
21:11 Democratizing Tool Building and Standardization
25:57 Challenges in PR Review Processes
30:59 Future Focus: Enablement and System Rethinking

📫 Sign up for our email list for more podcasts, articles, events, and other updates: https://www.aviator.co/podcast

✏️ Subscribe for more videos: @Aviator-Co

🙌 Join a curated community of senior engineers and engineering leaders focused on developer experience and solving productivity challenges at scale! Check out our upcoming off-the-record online sessions where vetted, experienced professionals can exchange ideas and share hard-earned wisdom: https://dx.community/

Replace code reviews with verified intent
AI writes code faster than humans can review it. Aviator Verify provides compliance-grade verification through spec-driven development. Ship faster with complete audit trails. https://verify.aviator.co/