Sign up to save your podcasts
Or
Share Build, Adapt, or Get Eaten by AI: Modern Software Engineering in 2025 (w/ Stephen Barr)
Share to email
Share to Facebook
Share to X
Share to email
Share to Facebook
Share to X
Or
Build, Adapt, or Get Eaten by AI: Modern Software Engineering in 2025 (w/ Stephen Barr)
When AI is moving this fast, you’ve got three choices: build, adapt, or get eaten.
In this week's Tech Talk with Stephen Barr, we unpack real-world lessons for modern software engineers, where “vibe coding” hits the wall, and the practical guardrails every engineer needs to ship safely in 2025.
You’ll hear wild stories from the field (like that time when a client's API keys were hard-coded in the frontend 🤯), why fundamentals still win, and the exact AI tool stack we use day-to-day to stay ahead.
In this episode, you’ll learn:
Why “build + adapt” beats waiting for the next model drop, and what the AWS Hero Summit reinforced about real engineering.
A vibe-coding post-mortem: exposed API keys, how it happened from vibe coding, and how to prevent it (even with AI in the loop).
“Vibe coding” vs. engineering: backups, PITR, DR, and the boring stuff that saves companies.
How deeper fundamentals (Rust/Haskell/typing) make your AI prompts—and outcomes—dramatically better.
Our current AI tool kit (Bedrock, OpenRouter, Perplexity, Claude Code, Kiro) and when to use each.
Key Moments (Timestamps):
0:00 – Intro + Welcome to Stephen Barr!
0:36 – Post-AWS Hero Summit debrief + why this matters in 2025
1:06 – From “new Hero” to lessons on tools, frameworks, and running a business
14:20 – Fundamentals still win (how expertise changes your AI prompting)
15:33 – Inside the Hero Summit: builder sessions, code walk-throughs, roadmap vibes
17:04 – Why “vibe coding” isn’t enough—and where real engineering begins
18:49 – Client story: the exposed API keys (what broke + how we fixed it)
20:25 – “The AI agent deleted the database”: backups, PITR, and DR in practice
27:54 – My AI tool stack: Bedrock, OpenRouter, Perplexity, Claude Code/Kira (and why)
29:18 – Ditch copy-paste loops: talk to your codebase, speed up delivery
32:00 – Takeaways: ship now, iterate forever
💼 RESOURCES:— Join Profit With Cloud®: https://dub.sh/PWC-Waitlist— Start 7 Days of Development: https://dub.sh/7DaysOfDev— Join our email list: https://dub.sh/TSP-Email— Start Coding in Python: https://dub.sh/TSP-Join-Udacity🌟 DEALS & DISCOUNTS:— My Favorite Tools: https://dub.sh/BHH-Tools-YT— My Workspace Essentials: https://dub.sh/BHH-Workspace-Kit— My Top Books: https://dub.sh/BHH-Tech-Books-Kit👨💻 TOP BUILDS:— Create Your Own LLM: https://youtu.be/Kq597DYdMEE— Build a Slack AI Chatbot: https://youtu.be/vtbtq7AgfIo— Analyze Stocks with Python: https://youtu.be/3q8FbiR-t8gWhat are you currently learning? Drop it in the comments 👇#ai #llm #coding #softwareengineers #coders #tech #vibecoding
View all episodes
By Brian H. Hough
5
2626 ratings
When AI is moving this fast, you’ve got three choices: build, adapt, or get eaten.
In this week's Tech Talk with Stephen Barr, we unpack real-world lessons for modern software engineers, where “vibe coding” hits the wall, and the practical guardrails every engineer needs to ship safely in 2025.
You’ll hear wild stories from the field (like that time when a client's API keys were hard-coded in the frontend 🤯), why fundamentals still win, and the exact AI tool stack we use day-to-day to stay ahead.
In this episode, you’ll learn:
Why “build + adapt” beats waiting for the next model drop, and what the AWS Hero Summit reinforced about real engineering.
A vibe-coding post-mortem: exposed API keys, how it happened from vibe coding, and how to prevent it (even with AI in the loop).
“Vibe coding” vs. engineering: backups, PITR, DR, and the boring stuff that saves companies.
How deeper fundamentals (Rust/Haskell/typing) make your AI prompts—and outcomes—dramatically better.
Our current AI tool kit (Bedrock, OpenRouter, Perplexity, Claude Code, Kiro) and when to use each.
Key Moments (Timestamps):
0:00 – Intro + Welcome to Stephen Barr!
0:36 – Post-AWS Hero Summit debrief + why this matters in 2025
1:06 – From “new Hero” to lessons on tools, frameworks, and running a business
14:20 – Fundamentals still win (how expertise changes your AI prompting)
15:33 – Inside the Hero Summit: builder sessions, code walk-throughs, roadmap vibes
17:04 – Why “vibe coding” isn’t enough—and where real engineering begins
18:49 – Client story: the exposed API keys (what broke + how we fixed it)
20:25 – “The AI agent deleted the database”: backups, PITR, and DR in practice
27:54 – My AI tool stack: Bedrock, OpenRouter, Perplexity, Claude Code/Kira (and why)
29:18 – Ditch copy-paste loops: talk to your codebase, speed up delivery
32:00 – Takeaways: ship now, iterate forever
💼 RESOURCES:— Join Profit With Cloud®: https://dub.sh/PWC-Waitlist— Start 7 Days of Development: https://dub.sh/7DaysOfDev— Join our email list: https://dub.sh/TSP-Email— Start Coding in Python: https://dub.sh/TSP-Join-Udacity🌟 DEALS & DISCOUNTS:— My Favorite Tools: https://dub.sh/BHH-Tools-YT— My Workspace Essentials: https://dub.sh/BHH-Workspace-Kit— My Top Books: https://dub.sh/BHH-Tech-Books-Kit👨💻 TOP BUILDS:— Create Your Own LLM: https://youtu.be/Kq597DYdMEE— Build a Slack AI Chatbot: https://youtu.be/vtbtq7AgfIo— Analyze Stocks with Python: https://youtu.be/3q8FbiR-t8gWhat are you currently learning? Drop it in the comments 👇#ai #llm #coding #softwareengineers #coders #tech #vibecoding