I've just spent two days at The Future World Symposium organised by the NMI (trade association for the UK Electronic Systems, Microelectronics and Semiconductor Communities). With a broad agenda covering the Connected Home, Connected Intelligence, the Internet of Things, and Autonomous Systems, the standard of presentations was high. I thought the following were the most significant:
- James Younger, Parliamentary Under Secretary of State for Intellectual Property at the Department for Business, Innovation & Skills. This was significant because of his appearance rather than what he said. Several years of effort by the NMI have resulted in the UK government recognising the existence of our important industry sector. It was also interesting to hear a Conservative minister talking about Industrial Strategy.
- Hossein Yassaie, CEO Imagination. Interesting analysis comparing the UK and South Korea, and a restatement of his views (expounded in the ESCO report) on the importance of developing global brands.
- Keith Robertson, Linn. Argued that as things in the home become connected, opportunities arise to rethink (reimagine) how products and services are constructed. Keith looked at the recored music industry and how Linn has evolved its thinking. He gave a very nice example of product personalisation where the acoustic properties of individual speaker drivers are measured at assembly time and the digital processing within that speaker is tweak to optimise playback. Being Linn, it's not "mass customisation" but it points the way.
- Wally Rhines, Mentor. Interesting analysis of cost structures in microelectronics and the end of Moore's Law. To me he seemed to end up saying "but it will be alright, something will happen to make sure we continue of the learning curve". I have some sympathy for this view, I've seen the industry overcome enough "fundamental barriers" to have hope but.....
- Simon Knowles, Xmos. Set out his view that machine intelligence was a field ripe for commercial implementation if only we could do it economically. At the moment we are one or two orders of magnitude short of computational power, and with the death of Dennard scaling and the decline of Moore's law, we're not going to get their with conventional computer. Furthermore, some of the techniques we use to gain performance in conventional computers and GPUs (caches, vector units) don't work for machine intelligence applications. We need new, parallel architectures, to deliver machine intelligence at the consumer scale. Simon is positioning this as the largest opportunity since the computer.
- Finella Frost. Explained why I need to spend £2000 to improve the lighting in my home. Not many people I spoke to appeared to be convinced by Lighting 2.0.
- Anthony Baxendale, MIRA. Useful insight into some of the problems of increasing the amount of automation in road vehicles. Enjoyed the characterisation of different types of automation as "hands off" (e.g. parking assist) and "foot off" (adaptive cruise control) with "brain off" being the challenging step.
Stepping back from the details, there were a number of themes that came through. Everything will be connected; there will be disasters along the way. Autonomy will come to vehicles (and aircraft) in time, but but may take generational change for it to become acceptable; "Generation Y considers driving to be a distraction to texting".