Avoiding Too Many Abstraction Layers

Senior engineers don’t “wing it.” They use mental models to turn plans into production—faster, safer, and with less drama. Danny Thompson, Leon Noel, and Kent C. Dodds break down the exact models they use: decision docs, second-order thinking, reducing cognitive load, Occam’s Razor, leaky abstractions, feature flags vs. staging, chaos engineering, AI context windows (and rot), MCP, onboarding docs, blind code reviews, and more. If you’re pushing from mid → senior (or trying to sound senior in interviews), steal these.

Guest: Kent C. Dodds!

https://kentcdodds.com/

https://x.com/kentcdodds

@KentCDodds-vids

SITE https://www.programmingpodcast.com/

Stay in Touch:

📧 Have ideas or questions for the show? Or are you a business that wants to talk business?

Email us at dannyandleonspodcast@gmail.com!

Danny Thompson

https://x.com/DThompsonDev

https://www.linkedin.com/in/DThompsonDev

www.DThompsonDev.com

Leon Noel

https://x.com/leonnoel

https://www.linkedin.com/in/leonnoel/

https://100devs.org/

📧 Have ideas or questions for the show? Or are you a business that wants to talk business?

Email us at dannyandleonspodcast@gmail.com!

You’ll learn

→ How to communicate invisible work like a senior (impact vs activity)

→ Decision documents: making tradeoffs explicit

→ Second-order thinking to prevent “future bugs”

→ When to ship fast with flags vs. slow with staging

→ Designing for onboarding—humans and AI assistants

→ Using AI without nuking your circle of competence

→ Local job tactics (Memphis example), recruiters, and “entry-level” framing

If this helped, hit 👍 and subscribe. Drop your favorite mental model in the comments

Chapters

00:00 Cold open — “When plans break in prod”

00:36 Why mental models matter for seniors

02:45 Define “mental model” (map vs. territory)

05:22 Intros + episode setup

07:05 Decision documents & guiding principles

10:48 Coaching moments: easy vs. right solutions

12:56 Second-order thinking (caching, short links, spam)

15:58 Occam’s Razor in real engineering decisions

18:12 Onboarding for Future-You (and your AI)

20:44 Tests as guardrails for humans + LLMs

22:18 AI context windows, “context rot,” and structure

24:55 Circle of competence in the AI era

27:18 Should juniors use AI? The real risks

30:06 Shipping fast vs. learning deep: when to limit AI

31:55 Interviews are changing: system design vs LeetCode

33:42 Confidence traps: getting gaslit by your prompts

36:00 Too many abstractions → leaky abstractions

38:40 Blind code reviews & cross-team learning

41:20 Google-style anonymous reviews (tradeoffs)

44:00 Feature flags vs. staging (the spicy debate)

46:52 Chaos engineering & safe rollouts in prod

49:10 Pragmatism over dogma: what actually ships

50:58 Kent’s Epic AI cohort & MCP primer

54:20 Listener Q&A: local vs. remote, .NET vs. Node

1:00:45 Ask Danny, Leon, And Kent!