A trusted tool gets compromised… and it raises a bigger question: why do we trust anything at all?

The Axios npm hack wasn’t just another security incident — it was a reminder of how much trust we place in systems we don’t fully understand.

In this episode of Shop Talk, we start with the Axios supply chain attack and then dig into something deeper: why we trust things that “seem right,” even when they aren’t — from software to backups to classic cars.

This show is driven by real viewer questions from Dave’s Garage and Shop Talk on Dave’s Attic — especially the ones with question marks.

Topics in this episode:

The Axios npm compromise and what actually happened

Why “working” systems often fail when it matters

Trusting the cloud vs owning your data

AI that looks right… but isn’t

Why we still want things that make no logical sense

If you enjoy thoughtful engineering discussions, debugging philosophy, and real-world systems thinking — this is Shop Talk.

00:00 Intro

07:09 SYSTEMS — When “It Works” Isn’t True

07:50 Purpose vs Results

08:26 Notifications ≠ Success

08:44 Schrödinger’s Backup

09:11 Restore Failure Reality

10:35 Never Trust the Cloud

11:04 TRUST — Who Controls Your Data?

11:48 Data Ownership Risk

12:22 When the Cloud Fails

13:02 AI SLOP

15:58 AI Almost Gets It Right

17:18 AI Rewriting AI Code

20:28 Vibe Coding?

21:53 WTF Are Those Diagrams?

27:33 QOTW

27:37 Do They Make Sense?

28:08 Where’s the Walkaround?

30:13 Maybe Part 2?

31:21 Geo-location Reveal?

32:21 Today’s Equivalent?

34:21 Legacy in 50 Years?

36:53 No Brazil Reference?