Robotics Funding

It's nice to see that robotics research in the UK getting a little more funding, although having seen some examples of government funding in the past I'm rather cynical about how much of the extra money will go to engineers and research equipment. Usually the majority is absorbed by managers, accountants and rather useless consultants.

To expand a little on the comment from Noel Sharkey, practical self-driving vehicles which an average person might expect to buy or use are perhaps still a decade away. Although much of the self-driving problem is now well characterized in terms of the underlying research the subsequent transition from research prototype to practical product is always the hardest part, and in general road driving there will be lots of corner cases. Assuming that most of those issues are solvable there will then be a new wave of technological unemployment which will be the biggest for one or two generations. Today many people are either directly or indirectly employed to drive. If you drive on a motorway just count the number of HGVs and delivery vans you see. It's typically a significant fraction of the vehicles on the road. In future decades all those jobs will be gone. Some new ones will be created in maintaining and servicing automated vehicles, but very few. Mostly the existing automotive engineers will just learn some new skills and that will be all that's required. So it would be sensible to think about the future employment situation and start moving in the direction of decoupling income from wage labour.

I think robotics automation has a lot of potential to improve quality of life, but that the way that the economy is arranged tends to work against that. For example, disabled people could get a lot of benefit out of improved home automation or robotic mobility systems, but if you look at what's happening to that category of people in the UK today it's an extremely bleak picture in which being able to afford robotics systems seems like an absurd and almost derisory fantasy.

On a practical note for the entrepreneurs my advice is that robotics R&D is not like pure software, so it's not like investing in the development of some app or internet service where you can rapidly throw some code together and deploy it. Doing any actually innovative robotics involves non-trivial capital investment, so it's not just about the cost of labour and some laptops. It also involves having electrical testing equipment, ability to manufacture prototype parts, facilities suitable for physical assembly and testing, complying with safety regulations and so on. In other words it's more similar to traditional industrial manufacturing which is a quite different mindset from Silicon Valley "unicorns". Bugs in robotics R&D can exist in software, in the electronics or in the mechanical design and so you need to be equipped to adjust all of those things. Investment in shared co-working or hackspaces would make sense in order to reduce these fixed capital costs and give entrepreneurs access to the tools they may need to get started.

Also robotics can be disconcerting for people coming from the pure software world because of the non-determinism involved. Fundamentally it's about systems which are capable of taking action in the presence of significant uncertainty, where the uncertainty is both about the state of the robot itself and the world it inhabits.

The best known business model for making progress on robotics in a relatively short time is the Willow Garage arrangement. You set a budget and then try to hire the best robotics people you can find. Their only remit is to try to push the envelope of what robotics systems can do and to publish their results as open source so that it can be reused, together with demo videos. That kind of model resulted in the biggest improvement in robotics capabilities that I've seen within my lifetime. Avoiding re-inventions of the wheel is very compelling.

I was amused to read in the article that the UK might lose it's competitive status in robotics, since as far as I know the UK doesn't have any competitive status. Historically most robotics R&D has been done elsewhere - typically in the US or Japan. That could change though if there's enough investment in the UK and it's appropriately focused.

So despite my general cynicism I welcome the fact that more money is going to robotics research. Long may it continue, and bring on the future of "fully automated luxury communism".