Building Fundamental Redundancy in Industrial AI

Why do I keep saying that we have a long way to go before we need to worry about fine-tuning “Responsible AI” so that it doesn’t hurt humanity?

Because the current capabilities have rough edges that need to be smoothed before we get to the point of serious governance, and smoothing most of the rough edges requires human intellect and common sense.

This one falls within the common sense bucket.

I was surprised that Waymo’s infrastructure did not have built-in redundancy for a risk that anyone can imagine—a loss of power.

I keep talking about the long-tail scenario of driverless cars, but that always pertained to the sh*t that can happen on the road. This one was scary in terms of oversight.

We discuss possible remedies in Episode 29 of “Maverick Mondays”.

#AI #ArtificialIntelligence #Tech #Technology


Leave a comment