Into the Gnar

What Tech Stack Would We Use for B2B SaaS in 2026? | Into the Gnar EP.3

Not sure if your team is actually ready for AI? We built a free 2-minute assessment that gives you a straight answer → ai-assessment.thegnar.com/ 
 
If your team is about to pick a tech stack and nobody can agree on what to build on, this one's for you.
 
We've been getting this question constantly since the AI coding wave hit. If you were starting a B2B SaaS company today, what would you actually build on? Not what sounds impressive. Not what worked five years ago. What would you ship on right now.
 
This week we talk about why the tool is almost never the problem, why boring technology is still the right call in 2026, and why the LLM your team codes with is starting to matter as much as the framework you pick. We also get into something founders don't talk about enough: the difference between a tech stack that gets you to launch and a stack that quietly kills you six months later.
 
We cover the AI coding tool wars — Claude Code, Cursor, Copilot, Codex — and what the data actually says about who's winning and why. 
 
Spoiler: developers find the better tool. They always do.
Here's what we got into:
The worst B2B SaaS tech stack decision we've ever seen and what it actually cost
Why 84% of developers use AI coding tools daily but only 29% trust what ships to production
Claude Code vs. Copilot vs. Cursor in 2026 — what we actually use and why
The exact tech stack we'd choose to build a B2B SaaS startup today
Why annual AI tool licenses are a trap and what to do instead
What "AI-ready architecture" actually means in practice
If you're sitting on a stack decision right now, or someone on your team is pushing for the shiny new framework, send this one to your CTO before your next architecture conversation.
New episodes every other Monday.
 
TIMESTAMPS:
00:34 Worst Tech Stack Decisions and What They Cost 
03:15 Claude Code, Cursor and Codex: Are They Merging Into One Stack? 
04:25 Should You Standardize on One LLM for Your Dev Team? 
06:46 Best Programming Languages for AI Coding Tools in 2026 
08:09 Annual vs Monthly AI Tool Licenses: What Founders Get Wrong 
10:15 How to Trust AI Generated Code in Production 
15:42 How to Evaluate an AI Dev Shop Before You Hire Them 
20:12 Claude Code vs GitHub Copilot: What the Adoption Data Says 
26:39 How to Choose a Tech Stack With Limited Runway 
29:52 Why Boring Technology Is Still the Right Call in 2026
32:46 What AI Ready Architecture Actually Means 
34:39 The Exact B2B SaaS Tech Stack We Use at NAR 
38:42 Key Takeaways: What to Build Your SaaS On Today

What is Vibe Coding and Why It's Becoming a Major Problem for Companies in 2026 | Into the Gnar EP.2

Not sure if your team is actually ready for AI? We built a free 2-minute assessment that gives you a straight answer → ai-assessment.thegnar.com/ 
 
If someone on your team just shipped something that now runs payroll reconciliation and nobody knows what's inside it, this one's for you.
We've been wanting to make this episode since we started to think about launching the show. Vibe coding. You've probably heard the term. You may already have an employee doing it. And if you haven't heard about it yet, you will soon, because the bill is coming due for a lot of companies.
This week we get into what vibe coding actually is, why it's showing up inside companies that never thought of themselves as tech companies, and what happens when that AI-generated code becomes load-bearing infrastructure with no one to maintain it. We also draw a line that we think matters a lot right now: vibe coding and agentic development are not the same thing. One is a tool. The other is a process. And the difference between them is the difference between a fun prototype and a ticking time bomb.
Gartner predicts that 75% of technology decision-makers will face serious technical debt by 2026. A new concept called comprehension debt is getting traction in the developer community, and honestly, it's scarier than technical debt because it means nobody in the building understands what the code does or how to fix it when it breaks. And if you're using tools like Cursor, Replit, or Lovable to build internal tools, this episode will tell you exactly what to watch out for.
In this episode, we cover:
What vibe coding actually is and why it's exploding in 2025 and 2026
The difference between vibe coding and agentic development
What comprehension debt is and why it's worse than technical debt
Why Gartner predicts a 2500% increase in software defects by 2028
What AI-generated code actually looks like when you open it up
What to do when a vibe coded app becomes a critical business infrastructure
If this one hits close to home, send it to your ops leader, your COO, or whoever on your team just said, "Look what I built in Cursor."
New episodes every other Monday.
TIMESTAMPS:00:58 — Nick vibe codes a full security camera system from scratch
03:43 — Can non-technical people actually do this?
05:20 — When vibe coding saves a client meeting
07:37 — The numbers: technical debt, defects, and Gartner's ugly predictions
13:38 — Vibe coding vs agentic development: what's the actual difference
16:33 — Comprehension debt: worse than technical debt and harder to fix
23:39 — The internal tool nobody owns that's running your business
26:30 — What companies should actually do about this
30:22 — Inside a vibe-coded codebase: what we actually see
36:15 — The doom loop: every new feature breaks something
41:13 — How Into the Gnar builds with AI differently
 

If You Run a Business and Software Keeps Going Wrong, Start Here | Into the Gnar EP.1

Not sure if your team is actually ready for AI? We built a free 2-minute assessment that gives you a straight answer → ai-assessment.thegnar.com/ 
 
If you run a real business and software keeps going wrong, you're not alone, and it's probably not the developers' fault.
 
Into the Gnar is where we talk about what actually happens when software gets built. The real decisions, the real consequences, and the moments where things went sideways and how they got back on track. No theory. No consultants. Just two people who've been in the trenches for ten years and have seen enough to know what nobody's telling you.
 
We're Mike Stone and Nick Maloney, co-founders of The Gnar Company, a software development agency based in Boston. We started this show because the conversations we keep having with operators, founders, and executives deserve to be louder. The gap between how software gets talked about and how it actually gets built is where most of the damage happens. We're here to close that gap.
In this first episode, we get into it: why Claude Code just displaced GitHub Copilot and what that means if you're hiring a dev partner right now, why MCP's 97 million installs don't mean your company should default to it, and why the thing that actually matters in software development was never the code to begin with.
 
Ten years of hard lessons. Forty-five minutes at a time. New episodes every other Monday. If this resonated with you, subscribe, share it with someone who needs to hear it, and leave us a comment telling us what you want us to tackle next. What decision is keeping you up at night? What question do you wish someone would just answer plainly? That's exactly why we're here.
 
 
TIMESTAMPS:00:00 Welcome to Into the Gnar
01:22 Team Autonomy Moment
03:26 AI Fluency Program
05:20 News Claude Code Wins
09:46 News MCP Hits Scale
15:04 Why We Started Thenar
18:14 Best Practices in AI Era
22:09 US-Based Product Shaping
26:41 Biggest Software Lies
31:24 Moving Fast with Guardrails
36:04 Who This Podcast Is For
38:38 Rapid Fire Questions
42:08 Why a Human Podcast
43:25 Wrap Up and Next Steps