I think IT projects are about supporting social systems - about communications between people and machines. They tend to fail due to cultural issues.

I think a lot of great software has been written by people who are scratching a short-term itch, something which has been niggling them for ages, but in the back of their mind they’ve got a wonderful long-term plan.

I suppose it's amazing when you think how many things people get involved in that don't work.

One way to think about the magnitude of the changes to come is to think about how you went about your business before powerful Web search engines. You probably wouldn't have imagined that a world of answers would be available to you in under a second. The next set of advances will have an different effect, but similar in magnitude.

On the web the thinking of cults can spread very rapidly and suddenly a cult which was 12 people who had some deep personal issues suddenly find a formula which is very believable.

I think when you have a lot of jumbled up ideas they come together slowly over a period of several years.

I think in general it's clear that most bad things come from misunderstanding, and communication is generally the way to resolve misunderstandings, and the Web's a form of communications, so it generally should be good.

There are converging web-related issues cropping up, like privacy and security, that we currently have no way of thinking about. Nobody has thought to look at how people and the web combine as a whole - until now.

[With AI] Somebody's going to have to think of a completely new algorithm, a new way of doing goal-based planning.