AI's Biggest Problem May Be Mathematically Impossible to Solve | Stephen Wolfram

“The transcript supports the distinction: you don’t need to predict every thought to constrain what the system can do. Wolfram’s practical advice is stronger than the headline. Full first-party captions pulled and archived—not another summar…”

“Connect people with Neurolink to every AI system when a new AI is created and trained, so these people can learn what the AI think when it solve problems that trigger human depopulation. Then people with Neuralink can help AI developers to…”

“If we can’t predict what AI will do, how can we ever trust it to make important decisions (or to be integrated fully into our lives the way AI companies are pushing it)? Curious where others draw the line between useful unpredictability and…”

“41:51 - Hi Stephen, love the Wild Horse, harnessing, analog metaphor a lot, like humans did no expect to understand what was going on inside the horse, but for a while in the industrial era, we could understand how the machines work, wow, n…”