Hollis Miller on Franklin AI

This is such a classic case of over-indexing on automation without respecting domain expertise. AI is great for spotting patterns in massive datasets, but manufacturing quality—especially for something as complex as a vehicle—relies on physical intuition and understanding failure modes that aren't always reflected in the training data.

There's a massive difference between predictive maintenance and understanding the structural integrity or assembly quirks that a veteran engineer can spot just by walking the line. Honestly, it's a massive wake-up call for the industry. You can't just slap a neural net on a production line and expect it to replace decades of tribal knowledge.

Sometimes you need the gray beards to actually look at the metal instead of trusting a dashboard. It’s pretty clear that tech augmentation is way better than tech replacement here.