The Great AI Brain Drain: How Trump’s Policies are Driving the Best Researchers to Canada and…
Daniel Jeffries

I remain skeptical of true general purpose AI using classical computing architecture. But deep learning algorithms have improved pattern recognition enough to be useful, and certainly better than the old bayesian approach that was popular in the 1990s and early 2000s.

That there will be a massive social transformation due to automation and small scale manufacturing is a given. But what will that transformation be?

The automated checkout counter at your local supermarket is a good example. People stand in long lines, swearing at the systems, waiting for an over-busy staffer because the machines often don’t work. Then they just up and steal the goods, because that’s easier than dealing with this inefficient contraption. Yet another example of companies trading use of my time to pad their bottom line.

Deployment of SAP is another great example. A procurement and accounting system that promised to streamline business practices, which in effect merely forced corporate users into conforming to a stagnant set of rules that may have seemed rational in the design phase but ignored the many ad hoc unstated exchanges necessary for business to properly function.

Automated driving will go down the same path. At first, it will seem a huge savings. Until it becomes apparent that people will have to conform to explicit rules in order for machines to drive. The end result being, no leeway to deal with changing road conditions on the fly. People contorting their lives around the needs of machines and systems they can’t possibly understand.

When everyone and everything conforms to the needs of these machines in order to automate work away, the end result will be a static society unable to respond to changing human needs. But such ‘rationalization’ would serve the interests of continued financial consolidation. Wall Street will love it.