Build vs. Buy

Mang-Git sits down with Matthias Wagner, founder of Flux.ai, to explore how AI is finally making hardware design as accessible as software development. Matthias shares how Flux uses a combination of LLMs and proprietary spatial models to let anyone design PCB boards through a simple chat interface – and why LLMs alone fall flat when it comes to physical, geometric reasoning. The conversation digs deep into the build vs. buy dilemma in an era of rapidly shifting AI tooling, from when to roll your own stack to when to trust a third-party abstraction. Matthias also makes a case for using high-level AI agents like Devin to stop chasing the latest models and start compounding on integrations instead.

1:07 — Intro & Guest Background

4:30 — The Origin of Flux.ai

6:51 — Build vs. Buy: The AI Stack Reality

9:54 — Abstraction Layers & Developer Tooling

13:34 — How Flux.ai Uses AI to Design Hardware

15:53 — Where LLMs Fail: Spatial & Physical Reasoning

17:30 — Feedback Loops as the Real Moat

26:37 — Deterministic Tools vs. ML Models: How to Decide

31:27 — Lightning Round & Wrap-Up

This episode was brought to you by Anvil [https://useanvil.com].

Mang-Git interviews Danil Kolesnikov, previously a founding engineer at Shepherd, a commercial insurance MGA for construction that underwrites on behalf of carriers and distributes through brokers. He explains Shepherd’s thesis that tech-enabled contractors are better risks and should receive better pricing and coverage. The core build-vs-buy decision was whether to buy a policy administration system or build one; Shepherd chose to build due to vendor pricing that takes a percentage of premium (a lasting “tax”) and the need for flexibility and speed. Danil says owning the platform improved onboarding and day-one productivity across underwriting, claims, and compliance, and helped Shepherd deliver quotes in under 24 hours versus weeks for legacy carriers. They also discuss the complexity of commercial lines, a painful Salesforce integration, and how AI shifts the bottleneck from coding to domain understanding and workflow automation. 

00:00 Introductions 

02:38 What an MGA Does 

04:30 Shepherd’s Tech Advantage 

06:10 The Build vs Buy Call 

11:45 Evaluating PAS Vendors 

13:17 Why Owning the Stack Wins 

16:14 Org Design and Automation 

18:59 Vertical Integration and Salesforce Pain 

22:56 Secret Sauce Not for Sale 

24:12 Lessons and AI Prompting 

26:22 Lightning Round and Wrap Up

This episode was brought to you by Anvil [https://useanvil.com].

Rowland Hobbs, founder and CEO of Stake, joins the show to talk about his startup background and Stake’s mission to help renters build wealth amid rising rents. Stake offers cash-back and financial services across a network of about 550,000 homes, including free checking that rewards rent payments and credit building by reporting on-time rent to the three bureaus, benefiting renters while reducing landlord delinquencies. Rowland explains Stake’s model, funded by landlords for leasing plus banking, brand, property, and government partners, including Colorado Renter Rewards. For a build-vs-buy decision, Stake tested distributing renter incentives through rental brokers, saw higher lease conversions, and acquired the largest Texas rental brokerage, Yu Move Free, to accelerate go-to-market and onboard renters into Stake’s broader platform.

00:00 Welcome and Introductions

01:02 Roland’s Startup Journey

02:13 Photo Startup Lessons

03:56 What Stake Does

05:06 Business Model and Network

08:24 Build Versus Buy Setup

08:49 Testing Brokers Then Buying

10:49 Rental Brokers Explained

14:01 Why Acquire Not Build

19:08 Costco Model for Renters

22:07 Deposits and Trust

24:05 Favorite Recent Purchases

26:46 Personal Growth and Yoga

28:18 Closing and Farewell

This episode was brought to you by Anvil [https://useanvil.com].

Michael Grinich, founder of WorkOS, joins Mang-Git to discuss challenges and insights from his career, including his journey from founding Nylas to creating WorkOS. They delve into the critical considerations for build versus buy decisions, the impact of AI on software development, and the importance of leveraging stable infrastructure. The episode highlights how modern AI tools are accelerating enterprise readiness and shaping the future of software development. 

00:00 Introduction 

01:51 Michael's Early Career

02:59 Challenges with Enterprise Adoption 

04:23 Founding WorkOS 

04:59 The Shift in Enterprise Software 

08:03 The Rise of AI and Enterprise Adoption 

08:38 The Importance of Enterprise Capabilities 

09:40 Build vs. Buy: The Time Machine Argument 

14:06 AI's Role in Modern Development 

16:15 The Future of Development and Infrastructure 

24:58 The Hidden Infrastructure of New York City 

25:38 Frameworks for Build vs. Buy Decisions 

26:28 The Value of Time in Business Decisions 

28:31 Challenges with Vendor Lock-In 

29:10 The Future of Internal Tools and AI 

34:18 Personalized Software and Its Implications 

40:18 Exciting Projects and Personal Favorites 

44:23 Conclusion and Final Thoughts

This episode was brought to you by Anvil [https://useanvil.com].

In this episode of Build vs. Buy, Mang-Git hosts Alex Klarfeld, founder and CEO of Supergood, previously co-founder of Divvy Homes. They discuss Alex's career journey, the business model of Supergood, which creates unofficial APIs for enterprise portals using automated reverse engineering, and the company's growth strategy. The conversation highlights a build versus buy decision Alex faced with integrating Temporal for job execution monitoring, ultimately leading them to inadvertently develop their own solution. 

00:00 Introduction

01:38 Alex's Background and Supergood Overview

02:36 Supergood's Unique Approach to APIs

04:07 Challenges and Solutions in API Integration

06:42 Technical Insights into Supergood's Operations

10:53 The Evolution of Supergood

13:43 Build Versus Buy: A Deep Dive

21:25 Challenges with Temporal 

23:39 Exploring Alternative Vendors

24:51 Choosing Masra for Better Support

28:30 Building In-House Solutions

30:58 Reflections on Vendor Relationships

33:51 Exciting Projects and Future Plans

37:07 Product Recommendations and Closing Remarks

This episode was brought to you by Anvil [https://useanvil.com].

In episode 23, Mang-Git chats with Chris Dean, Co-Founder and CEO of Treasury Prime. Chris shares his journey from machine learning researcher to serial entrepreneur, culminating in the founding of Treasury Prime. The discussion covers Treasury Prime's unique business model, which focuses on providing APIs to banks for managing FinTech relationships safely. They delve into the complexities of the American banking system, the challenges of ensuring safe money management, and the merits of building versus buying solutions. The conversation also touches on the use of AI in enhancing productivity and the future of FinTech integrations. 

00:00 Introduction  

02:03 Chris Dean's Background and Journey 

05:17 Founding Treasury Prime 

07:37 Treasury Prime's Unique Business Model

15:18 Challenges and Decisions in FinTech

20:39 Building a Robust Workflow System 

23:45 Handling System Failures and Retries 

25:32 Reconciling Financial Discrepancies 

26:03 AI in Banking Operations 

28:22 Regulatory Challenges and Opportunities 

32:09 The Importance of Optionality 

32:42 Personal Insights and AI Tools 

38:50 Conclusion and Future Outlook

This episode was brought to you by Anvil [https://useanvil.com].

John-Denis Greze joins Mang-Git Ng to dive into JDG's extensive career in tech, including his tenure as CTO at Plaid. JDG shares his decision-making process on build vs. buy, highlighting examples from his time at Plaid. He also discusses his current venture, Town AI, which aims to automate small business tax prep through AI. The conversation touches on the challenges and joys of being a founder, including management philosophies and leveraging modern AI tools. 

00:00 Introduction and Guest Welcome 

01:33 JD's Background and Career Journey 

02:25 Joining Dropbox and Career Growth 

04:59 Transition to Plaid and Leadership Roles 

06:51 Challenges and Reflections on Management 

08:52 Networking and Fairness in Tech 

14:20 Management Lessons and CTO vs VP of Engineering 

23:18 Starting Town and the SMB Tax Space 

26:32 Building Products in Existing Markets 

28:26 The Build vs. Buy Dilemma 

30:58 Strategic Decisions in Product Development 

38:17 AI and Rapid Prototyping 

43:53 Personal Projects and Tools 

47:14 Final Thoughts and Recommendations

This episode was brought to you by Anvil [https://useanvil.com].

Tom Elliot, founder of Ocuroot, joins Mang-Git to discuss Tom's background in the tech industry, his experiences at companies like VMware and Yext, and the inception of Ocuroot. The conversation digs into build vs buy decisions, emphasizing the case study of monitoring tool Checkly. Tom shares insights on the complexities of self-built solutions versus external services and how to present these decisions to senior leadership. They also explore the benefits and challenges of various AI development tools.

00:00 Introduction and Guest Welcome

01:33 Tom Elliot's Career Journey

02:50 Challenges and Innovations at Yext

03:55 The Concept and Benefits of OccuRoot

11:27 Build vs. Buy: A Real-World Example

15:04 The Decision-Making Process

21:12 Reflections on Build vs. Buy

28:42 Fun Questions and Closing Remarks

This episode was brought to you by Anvil [https://useanvil.com].

Pulkit Agrawal, co-founder and CEO of Chameleon, joins Mang-Git Ng to discuss Pulkit's journey founding Chameleon, a tool that simplifies user onboarding and adoption for software applications. Pulkit shares insights into the importance of personalized user experiences and the intricacies of build versus buy decisions. He elaborates on his company's switch from custom-built UI components to using Radix UI, the role of trust in choosing software, and the tension between self-service growth and sales-assisted models. 

00:00 Introduction 

02:06 Pulkit's Background and Chameleon Origin 

03:35 Challenges in User Onboarding 

05:57 The Impact of AI on UI Development 

09:45 Build vs Buy: A Personal Experience 

13:00 Choosing Radix UI for Speed and Trust 

16:15 Factors Influencing Build vs Buy Decisions 

22:04 Chameleon's Go-to-Market Strategy 

26:07 Personalization in Onboarding and Campaigns 

31:31 Reflecting on the Entrepreneurial Journey 

32:56 Exciting New Projects and Personal Updates 

34:11 Product Recommendations and Closing Remarks

This episode was brought to you by Anvil [https://useanvil.com].

On this episode of the Build vs. Buy podcast, Mang-Git talks with Sumeet Vaidya, the founder and CEO of Crafting. Sumeet shares his career journey, which includes roles at Facebook, Uber, and Discord, and his current company, Crafting, which focuses on developing enterprise-ready cloud dev environments. They discuss the evolution of software development from local applications to complex cloud-based environments, the challenges of developer collaboration, and the costs associated with running multiple services. They also explore the importance of trust and reliability in enterprise products, particularly in the context of building vs. buying internal tools. The conversation wraps up with a discussion on the future potential of AI tools and personal anecdotes from both Mang-Git and Sumeet.

00:00 Introduction 

01:30 Sumeet's Background 

02:59 The Evolution of Development Environments 

04:51 Challenges in Scaling Development 

08:46 The Future of Crafting 

11:42 In-House Solutions vs. Off-the-Shelf 

17:29 Trust in Internal Tools 

19:41 Challenging 'Move Fast and Break Things' in Enterprise 

21:21 Building Trust in Enterprise Applications 

24:03 Addressing Developer Control and Customization 

26:14 Deploying in Customer Clouds 

28:27 Reflections on Internal Chat Applications 

30:40 Excitement for AI and Personal Life Updates 

33:03 Fun Questions and Product Recommendations

This episode was brought to you by Anvil [https://useanvil.com].