It's the start of the year and so the internet is full of predictions for 2017. Instead of trying to come up with my own set of predictions, I wanted to think about some highly-hyped areas from 2016 that under-delivered.

AI

AI was everywhere in 2016 - probably the biggest hype story of the year. In many ways AI stepped up and delivered real progress in 2016, but there was still a lot of smoke, mirrors and over promising. Hopefully in 2017 we'll see fewer companies trying to associate themselves with AI when they don't actually have anything to do with it.

Blockchain

It's supposed to be the miracle cure for big financial services, but I feel that 2016 under-delivered in terms of real world deployment. Blockchain is another cool label that fin-tech start-ups slap on the pitch deck, but the differentiation usually comes from more mundane things like good UI and coding.

VR and AR

Pokemon Go was one of the biggest mobile stories of the year, but turned out to be a flash in the pan. I think the hardware costs for VR are still too high for mass adoption, and there still hasn't been a sustainable block-buster use case.

Smart watches

My wrist is still stubbornly analogue. Smart watches just don't seem to have connected with consumers in the same way as smart phones have. That being said we're still effectively on the first generation of the smart watch and the ecosystem of apps and use cases are still developing. I think smart watches really need a compelling use case that isn't just being a wrist mounted smart phone, to catch hold of consumers imagination.

What about 2017?

As well as all the exciting predictions of new technology that are being made, I think ReCode's below asks some of the better questions. The same old themes will still make a lot of noise. There are still plenty of industries run off-line or on old school enterprise software, and the internet and cloud will keep knocking them off one by one.

In general people tend to overestimate the pace of short term change, but underestimate long term change. So maybe all those predictions for 2015 or 2016 will finally turn up this year.