It’s always good to keep an eye on what Chamath Palihapitiya is up to. He has been building a team of ex-Googlers, and is supposed to be after the next generation of computing. A $10 million startup, to be sure!
This could signal a lot of things, depending which pair of lenses you put on. It probably has a lot to do with AI – Artificial Intelligence. For instance he hired away eight of the ten people at Goggle working on a secret project involving a chip with AI. He has poked fun at Watson, the IBM cloud-based machine learning application.
Watson, as you might be aware turned tables on Jeopardy and Go (the 2,500-year-old game), but has machine learning entrenched in many sectors from genomics to industrial safety. Google’s machine learning project, known by its bland name, Tensor Processing Unit (TPU), is underway.
Pahlihapitiya talks of ‘probabilistic‘ software that is changing how we depend on devices – a great shift from ‘deterministic’ software based on “if-the” sequences. Watch how he explains how machine learning and II is transforming, and will up-end computing. I bet Watson took in every word of this.
Back to AI. That famous ‘machine’ known as Watson, which beat humans in that game show Jeopardy, was able to search a massive databases and respond faster than Ken Jennings and Brad Rutter. But that was not all. It also outsmarted them in strategy – that is in picking the categories that would win bigger. It is eerie to watch those rounds and see how a computer sitting in between two humans looks like. (It sounds human too, as it calmly picks a category such as ‘chicks dig me’ to the nervous laughter of the live audience.)
I was intrigued to read about how JWT, the agency that handled IBM and this show, was briefed on how to present Watson. At one time, the inventor behind it, specifically asked that Watson should not bear any resemblance to…HAL. If you know Stanley Kubrik’s and Arthur Clarke’s