> In the past, it was the failure of AI research to deliver business impact. That's not quite right - the problem wasn't that no impact was delivered, but that

> In the past, it was the failure of AI research to deliver business impact. That's not quite right - the problem wasn't that no impact was delivered, but that it over-promised and under-delivered. I'm fairly optimistic of some of the machine learning impact (even some we've seen already) but it's by not means certain that business interest won't turn again. We are very much still in the honeymoon phase currently.