The Internet Of Automation

[Arthur Hanson]

Many technologies are coming together at a critical juncture in time and they all involve semi technology to varying degrees including robotics, 3D printing and 3D scanning. This will be just another link in the Big Data/Communications/processing power/Mems chain. To the end of this can be added automation and robotics (of which many mems are just nano scale automation). Semi technology has dramatically brought down the cost of these technologies as they have done to communications, processing and storage. Along with the Internet of Things, we will have the Automation of Everything (Production, Prototyping, Knowledge Processing, even everyday life).

Not only have these technologies brought down costs very dramatically, but will allow the progress of many, if not all technologies to pick up speed by compressing the timeline and cost structure of experimentation and prototyping. Already there are companies that do this over the wire as a service and not just in regular physical items, but in actual experimentation. Wuxiapptec has even brought out a mobile app for doing this in the medical and pharmaceutical field. Soon everyone will have access to remote access to do research and prototyping, then production.

Robotics are a significant part of this and with 3D printing with much of the wiring and some of the semiconductors printed in will increase the use of robotics and automation to a degree we haven't even dreamed of. The age of robotics and automation from nano scale to automated ships is soon upon us.

This will create a wave of growth that will create whole new fields of opportunity and study. The wave of growth this creates in the semi industry will rival the rise of the processor and change the world as we know it just as much. Education and training are going to have to radically change to adapt to a speed of change that not only is never going back, but increase and diversify at an ever increasing rate. Our educational, social, regulatory and government structures are going to have to change the most for they aren't even keeping up with the rate of change we have now. All of this will reach into every facet of our lives and the world around us.

I feel SemiWiki will be a key player in this and maybe become the key player in helping this trend progress and be of help in spreading the benefit of this trend for all. SemiWiki is not only an ideal place to have forums on technology, but the on the wisdom to put it to good use. We are like children that are being given great power and we must give as much thought to using it as we give to creating it.

Comments, thoughts, ideas and additions solicited, wanted and welcome.

 

[benb]

Robotics and 3D printing benefit from Moore's Law, but the applications aren't clear yet, and need to clear social and legal hurdles. Do we want robotic cars? Drones? How many of us will say "Not In My Back Yard" and opt out? Should they have this right? I am no luddite but would be concerned about drones and robotic cars, because electronic malfunctions could cause injuries and death.

3D printing may remain a hobbyist technology for a long time; it depends on how fast the 3D printer companies can evolve the technology resolution and speed. Currently it has to evolve a lot to reach into the mainstream, outside of aviation. But disruptive technologies usually start out like that, so I think 3D printing is likely to be something, (I can't say what) eventually.

 

[Arthur Hanson]

Many, many companies are using 3D printing on serious projects, like Boeing. The cost of 3D printing is about to drop dramatically and improve in quality with the entrance of HP as soon as key patents expire this year and next. This is already reflected in the drop in many 3D company stocks. 3D printing is advancing at the same rate as semis in sophistication, dropping prices, market penetration and applications. I believe we will see house structures made using 3D printing large scale. It has already been done and will only get better. How much disruption will that cause?

 

[Li Yisuo]

Now You Can 3-D Print Chocolate - Personal Tech News - WSJ
What do you wanna print for today?

 

[Daniel Payne]

This is interesting, a 3D printer for PCB from a company called Nano Dimensions in Israel:

[video=youtube;s4m5skH0sGw]https://www.youtube.com/watch?v=s4m5skH0sGw[/video]

 

[Arthur Hanson]

Dan, Fully integrated 3D printing on everything from microscopic objects to houses will have the process integrated with different materials to come much closer to a fully completed product. This will include human organs that are already being printed for lab use. Being from the EDA industry, how do you feel about the automation of knowledge. I feel such fields as pharmacists shouldn't exist, just like hand powered drills. Automated intelligence in this area could do a much, much better job than a person in all respects, from customer interface in numerous languages, to checking interactions and other options, to actually preparing and packaging the prescription. Vinod Khosla feels even doctors in many cases should be replaced by AI or close to it. It's going to more a political game. That's why US medical costs 400% more than the average industrial country and delivers 35th in quality. Imagine if EDA wasn't allowed by government regulation? No one working in medical would tolerate paying 400 % more for a car from a company that ranked 35th in quality. Any comments, thought or ideas welcome.

 

[ippisl]

Arthur :

Sure, Software in medicine had for a very long time, lots of potential to improve treatment and reduce costs.List you say , it's just a regulatory problem.

And definitely, there will be a lot of growth in robots and in technology in general. And yes 3d printing have lots of potential(altough there's currently much hype that will probably never be achieved).

As for semiconductors ? Volume won't probably come from 3d printers, or even the IOT , because their computational needs are quite low, and mcu processes are becoming quite advanced(28nm/40nm/55nm).

My guess is that semi volume will come from vision, virtual reality, maybe large scale data analysis, possibly cognitive computing(although IBM has made watson quite efficient), and maybe some other AI stuff.

And a lot of this is best implemented in generic compute platforms(cpu/gpu/fpga) which to the best of my knowledge are quite mature, so most of the growth will probably happen in a few large companies.

 

[Arthur Hanson]

ippisl, The computational needs of advanced diagnostics using massive data bases generated from people mounted devices in and out of hospitals will need almost AI level processing and large memory to actually find new causes and treatments for medical purposes. Worn IOT devices will soon provide massive data bases over long periods of time with far more inputs than we have now combined with being applied over millions of people, increasing the information by literally factors of millions, if not billions, will definitely require very sophisticated automated handling and processing to make it useful. Just like we have automated physical tasks, we will automate mental tasks in everything. This is why the IOT will lead to the Internet Of Automation. Combining the automation of physical and mental tasks will lead to a level of automation few have even contemplated or even dreamed of. Sending robots to explore the universe or under our largely unexplored oceans is but one task where automation and robotics will change everything. The wave of automating everything, physical and mental, will create massive needs for all areas of the semiconductor industry.

 

[ippisl]

Arthur, so we're talking about large scale data analysis because of the IOT. We're doing this today. The best chips for this are usually GPU's and CPU's - which are pretty mature industries , so most likely a few old players will benefit from it.

Worn IOT devices usually won't provide that much useful data at least according to some physicians. But labs-on-chips are starting to appear and make blood testing cheap, almost painless(pin prick instead of needle) and comfortable(maybe even at the home). This will surely increase the rate of blood testing and hugely increase the types on things measured. Add that to DNA testing, and automated medical questionnaires filled before a medical visit(some start-up is succeeding with that). Lots and lots of very useful data.

But let dig deeper. An average american visits the doctor maybe 4 times per year. That's not a lot. And probably most often for simple treatments.

But what about the research part ? analysing all that data and building all those models ? I don't know how to calculate it. But here's an estimation: the NSA is building a new center[1]. it surely runs lots of data(just are tweets are much more data than any possible blood testing, and text analysis is more computationally heavy than numeric data analysis in general). it costs $2 Billion. Let's multiply it by 5. let's divide it over 3 years. it's $1.5 billion per year. it's a nice sum of money for intel/nvidia. But As a growth driver for the semi industry, it's something around 0.5%. and that's a reasonably optimistic guess.

On the other hand , let's take the case that virtual reality is as popular as smart phones. could happen. Let's say we're talking about $200 in semi's per person. Let's say only 200 million people from the western world bought it. That is $ 40 billion dollars. divide by 3 years - that's $13 billion. yes that's optimistic. But look at the growth of smartphones.it's possible. And even that will mostly be concentrated in intel/nvidia/tsmc/samsung.

[1]http://www.wired.com/2012/03/ff_nsadatacenter/all/

 

[Arthur Hanson]

I was just looking at a business show this morning that stated medical is now one of the world's biggest industries. Last time I checked it was at 7 trillion dollars and growing quickly due to increased wealth, population and number of older people. I feel reform is especially needed in the US where our medicare debt with prescription benefit is now way over 100 trillion (usdebtclock.org). In the US our medical is 300% higher in cost than the average industrialized country and we rank 35th in quality with still escalating costs Any other industry that delivered results like this would not be tolerated. If they had to truly compete on a world scale, they would go out of business. The only explanation for this is a deeply flawed system in almost every respect. If the medical industry adopted the ever higher performance at ever lower cost ethos of the tech industry it would dramatically increase the heath and well being of our overall economy and society in general. I hope the tech sector can change an the medical industry that is deeply mired in the past in many, many respects. What causes this performance/cost disparity with the rest of the world is whole separate area of discussion that will probably be riddled with recrimination and hurt feelings. I hope a productive dialogue can be opened in this area.

 

[ippisl]

There are many reasons why the medical industry is so expensiveower structures, regulation, real complexity and importance, wrong incentives, etc. Some are valid, some are just various interests abusing the system.

One interesting point though is that usually ,adding technology to the system usually makes it more expensive. It might offer better care, but at a higher cost. But usually in other markets, a typical cost reduction effort goes like this: some new technology that offer lesser quality product but at a much lower costs comes, gain some customers , and with time improves to become even better both in price and quality than the product it replaced.

But this cannot happen in medicine - usually when creating a new product/service, it must be better than current care , or it won't be approved, won't be recommended by doctors , and it'll have big resistance from people. So there's no space for low-cost products to improve. And most technologies that are added only add cost.

But it might be starting to change, there's some reports that healthcare costs haven't risen recently, unlike previous years. So that's good news.

And we also see lots of innovation starting to creep into the system. For example IBM watson diagnosing cancer. With IBM publishing the system widely, maybe cancer patients will inquire with their doctors about using the system, and will be very angry when doctors don't. This could be the entry point for smart technology to medicine.

Also i see some potential for political battles about using smart systems in medicine.

Anyway, complex and fascinating subject.

 

[Arthur Hanson]

The political power of the medical lobby from professional organizations, unions, trade organizations, standards bodies and even the medical education establishment to twist and distort the free market to their benefit should not be underestimated. At three trillion a year in US medical spending and our costs being 3 times other industrialized countries for 35th in quality, seem to indicate 2 trillion in corruption/distortions/waste and bad processes. Any comments on this would be welcome.

 

[ippisl]

Yes, definitely. Lots of money is wasted, and even worse, lives. But that's a political game ,and in general the politics in the u.s. are far worse towards the common people than in many other countries. This will even get much stronger once we'll start to see automation in mass, like is sure to happen. At least european countries might have some chance to achieve a sort of a safety net. But the u.s. ? it doens't seems so. It could end very bad.