In A disastrous chatbot release in 2016 helps explain why Microsoft is trouncing Google in A.I. today, Steve Mollen tells us that in the current round of the Rock-Em-Sock-Em Robots Match, Microsoft is winning.
It is, as we all know, using technology from OpenAI. But that partnership came about because of Microsoft’s horrible chatbot release in 2016. The Tay chatbot, released on Twitter, learned the (deliberately) offensive behavior of other Twitter users in perhaps one of the earliest examples of prompt hacking. Its tweets were so offensive that Microsoft terminated it within a day. This lead Microsoft to a partnership with OpenAI and over three billion dollars in investment.
Google has a rival called Bard, but it has not yet integrated it into search results. It’s recent demo was marred by a factual error. But OpenAI came to be as the result of a mistake, so perhaps Google’s mistakes in this space will also end up being serendipitous.