Moving towards a more autonomous computer-oriented news curation, Microsoft this week laid off all their editors and writers for Microsoft News. Leo says that the news site isn't going away, Microsoft is just going to rely on an artificial intelligence algorithm to curate the stories it posts. Leo also adds that both Apple and Google have gone the other way, believing that human curation increases engagement. Leo also says that Facebook is also relying on an algorithm, and its news page is just terrible.
Concerned about how China could take advantage of it, the US is moving to limit exports of artificial intelligence research. But Leo says that China may be ahead of us.
Joseph says that AI can more accurately predict things than humans, so why should we trust things like political decisions, medical decisions, and even financial trading to anything but AI? Leo says that we're not there yet. But in the not too distant future, that may be the case. Look at self-driving cars: they have come a long way and someday, cars will be completely automated. But when that happens, millions of drivers will be out of work. Now multiply that by brokers, lawyers, doctors, etc.
Julian called in with a suggestion for for Larry in Prescott, AZ, who wanted to connect an external camera to his smartphone for use with an app called BeMyEyes that acts as a visual aid. Julian's idea is to either use an Android device for this, or to use a service called aira.io. This service works in conjunction with glasses that would be worn and identify what things it is seeing.
Jim wanted to bring up a movie from 1970 called Colossus: The Forbin Project, which dealt with the futuristic unintended consequences of true artificial intelligence. If it is true artificial intelligence, it will be smarter than humans, and it'll realize it shouldn't leave its future to humans. Back in 1970, that seemed pretty far-fetched because there was no internet and computers weren't as sophisticated. But now we are much closer to this potentially being a reality.
Google's latest artificial intelligence, AlphaGo Zero, now has the ability to teach itself how to master board games after only knowing the rules, and without any human intervention. While previous AI took months to beat the world champion Go player, this latest system was able to master these games in less than a day.
Robert wants a voice assistant that has the best artificial intelligence. He wants to be able to ask it a variety of different questions about the news. Leo says the Google Home is the winner in that category, but the Amazon Echo is the only device that lets you shop on Amazon. Amazon is also very good with music, and simple commands like setting a timer, and home automation tasks. There's also more than 50,000 third party skills on the Echo. The problem is that it's hard to find skills, and then he'd have to learn the specific syntax to use that skill.
Elon Musk is afraid that machines are getting smarter and smarter and he worries that they are an existential threat to humanity and may decide that they can rule over us or even worse, get rid of us. Leo thinks Musk has watched too many movies. Rodney Brooks, an artificial intelligence expert from MIT, said most people worried about that aren't working in AI, and he says that they aren't too worried because AI is too hard to perfect. Most are limited machines who can't move beyond their programming. It's a thin veneer of what looks like AI, but in reality, machines are pretty dumb.
The creators of Siri have moved on to create a new artificial intelligence platform called Viv. It is a new conversational interface, and it is brilliant. The original Siri developers say they believe it is better than Siri.