Intel had been good to me, and I felt a responsibility to help the company even after I left. I offered help to Paul Otellini when he became CEO. Paul and I had been friends both at work and socially. I told Paul I would be willing to help in strategy issues and no one had to know, and I did not have to be paid. He politely thanked me but only after a few years did he has for help. That is when Intel started up the Ultra Mobility Group to develop Atom project which was a microprocessor designed for mobile applications. There was a new group set up under Anand Chandrasekher. Anand eventually left for Qualcomm. I provided Paul with a verbal report which was pretty negative on Intel’s effort. I felt it was too little too late and the Atom would not be successful. I also thought that the team was weak overall. I suggested to Paul that Intel should acquire a few early-stage companies not only for some new technology but to get some additional talent. I also suggested some things that Intel could do to make its offering stronger that had to do with new technologies to link devices and servers.
Paul shocked me. He was not interested in what I had to say, and I guess I was also offended by my suggestion that Intel needed new talent. He said “We have the best people. I would not trade any of our people for any of the people that left”. I was not suggesting bringing back anyone and took this as a swipe at me. It was unusual behavior for Paul who was a pretty nice person in general.
After Paul was pushed out, Intel installed a duo of Brian Krzanich and Renee James. I did not know Brian even though we had started at Intel around the same time. Brian came up through the manufacturing organization. I knew Renee very well, and we were and are still friends.7
Once again, I wanted to help. I reached out Andy Bryant, whom I also knew well. Andy was and is currently the Chair of the Board. I told him that I wanted to help and that I would do it without acknowledgment or compensation. I said that I would write up my ideas and send them to Brian and Renee as long as they agreed that they would read what I was about to write. I think I also contacted Renee about this. I then send the following letter to both of them. I never got a response. Please note that this letter was written more than five years ago.
Intel and the Next Wave of Computing
Feb. 10, 2014
Dear Brian and Renee,
Brian, we have never met during the time I was at Intel (1984-1999). As you may know, I co-founded Intel Capital and conceived and led Intel’s broadband initiative staring in 1992 which lead to both the cable modem DSL infrastructure and residential broadband as we know it. While this effort had a very positive impact on Intel in both growing the consumer market for PCs and significant profits via our investments, it had no long term impact on Intel’s strategy. While I tried to present the opportunities, I saw for the company, and I had little or no impact. I regret this. Intel was good to me. I am hoping that my recommendations to follow might inform some of your thinking.
Renee, we have known each other for a long time now, although we have had little contact since I left the company. I am hoping that you understand my sincerity and still have respect for my abilities.
I am going to try to keep this short and concise because I am hopeful that both of you will read it and give it some consideration. My focus will be on the consumer market. I believe much of what I have to say is also relevant to enterprise and SMB. I would be happy to expand on any of this.
Intel missed the Current Wave of computing
The computer industry goes through waves of change. We went from Mainframes to Mini Computers, the networked PC, and now Cloud Computing and the post PC era. Almost no technology company is successful in going from one wave to the other.
The reason for this is pretty simple. New waves start up around the peak growth of the previous wave. It is a time when those riding the current wave have little interest in the new wave or even perceive it as significant. By the time the new wave is in ascension, it is too late for the incumbent to take a leadership role. This is what happened to both Intel and Microsoft.
Intel ended up having a strategy for the future which was pretty much “make today last.” That was certainly the situation in 1999 when I left the company and is now apparent. However, Intel has many significant assets. It is the leader in semiconductor technology, has talented people, a strong balance sheet and an excellent albeit, fading brand. This document is all about how Intel can once more become a leader in shaping the computer industry.
Time to skip a wave
Apple offers an interesting example. Apple never actually was a crucial player in the connected PC. They entered the PC market with the Mac 30 years ago but played a rather insignificant role in how the computer industry evolved. IBM was the leading computer company but on its way down. Digital Equipment Company was the second largest computer company and still on its way up. This was when Microsoft and Intel were able to set the standards and take the leadership of the PC era. But both companies got stuck when the action moved to the Internet. Interestingly, Intel, in particular, played a significant role in creating the transition for the consumer market by driving residential broadband and investing in crucial internet technologies.
I can take a lot of the credit for that, but also have to own the fact that I was not able to influence Intel’s strategic direction even though I pretty much understood what was about to unfold.
During the 90s when we would say correctly that “The PC was it” functionality flowed into the PC. I would call the PC a Black Hole which sucked in all that came near it. But by the end of that decade, the functionality was flowing out to the internet. Improvements in technology also allowed for powerful mobile devices. Mobility was the key because there was no way to put a PC in your pocket. Apple was the beneficiary of this, as you well know. I won’t go into that story. But it is an important example of how an established company can play a role in defining the next wave.
I think Intel has such an opportunity now. The key to this is understanding consumer needs and then matching them to innovative technology.
Cloud Computing – Kind of
Right now we have a hybrid of client and cloud computing. Some activities/applications run on the client and some run in the cloud. Some data is on our devices and may or may not be replicated remotely while some of our data is only in the cloud. Google is dominant in defining cloud computing. But Amazon and Apple have critical and different roles to play.
Google has taken over the leadership role that Microsoft once had. Like Microsoft, I believe they are an inch deep and a mile wide. Apple has significant long term problems because they need to introduce and dominate new categories for their business model to work. But there may not be such new categories available. TV, as Intel has learned, may not lead to the Promised Land. Apple does have opportunities in physical commerce. The combination of their large market share in smartphones, iBeacon, and iTunes (all those credit cards) can make Apple the leader in physical commerce. Amazon has fantastic potential with their extensive knowledge of consumers and AWS. Developing an Amazon strategy should be high on Intel’s list of objectives. Amazon may turn out to be the leader in the next phase of computing as strange as that might sound to you.
Why Cloud Computing is Broken
The strength of the PC era was due to the dominance of Microsoft and its ability to establish APIs. While the Internet has standards, there are many competing ways to do things. Technical and business issues limit interpretability between applications/services. Complexity is increasing at an alarming rate.
Part of the movement to the Cloud was the result of the complexity of synchronizing data between different devices. But now the mess has just moved. Take the example of photos. If I want to share my photos with my Facebook Friends, I have to upload them to Facebook. If I want to share them with my friends on Google+, I have to upload them to Google+, etc.
If I have my media on the cloud in let’s say, Google Drive and then want to share something on Facebook, the data will have to be copied and transferred from the cloud to my computer/device and then back to the cloud. The speed of this will be determined by the up and down of the consumers’ network rather than the rate at which data could be transferred between Google and Facebook. This is true if I have my data on Dropbox as well. And now there are two copies.
I am just giving one of the so many examples. Of course, each problem could be addressed by groups that would work together to create a standard but that is a slow process, and it is not clear that the leaders would be interested.
Key Consumer Needs
I am going to describe a set of consumer needs that are not being met.
This is a concept that goes beyond data security. Consumers want to be able to control who can access their information, including such things as photos. This means that while I might give my friend access to a photo, my friend can not share that without my permission. Consumers will want to be able to change permissions. Or set up other attributes. For instance, a photo can be visible to a particular person for only a certain amount of time such as SnapChat.
Consumers want to have only one copy of a photo to manage. They also want to be sure that their photos will not only be secure and under their control but that they will be preserved.
Currently, most of us have a one to one relationship with our devices/computers. It is my information that is on my Mac, iPhone, and iPad for instance. Once computing moves to the Net and the method is just a way of communicating to the Net, there is no longer a need to have such a linkage. Any desktop could be my desktop. Any phone could be my phone. All that needs to happen is that the device can identify who I am.
The way we communicate with the intelligence that is the Net will change a great deal. Frankly, I am not sure how all this will play out, but I am pretty sure it will. I think voice, gestures, location, and knowledge will inform the way we interact.
We are about to go through rapid change in the intelligence of the Net. This is being driven by increased computer performance and a better understanding of data mining (big data). But to be effective, the more information the net has about a consumer, the higher the IQ. This, of course, will lead to concerns about data privacy. But the power of this will out the way the concerns. If Waze can get me to work 10 minutes faster, I will be happy to let Waze keep information about my driving habits.
The OS in the Sky
I am going to describe a world in which all applications run in the cloud. I know that this will assume a high bandwidth, low latency and very reliable connection between the consumer and the network. Not all consumers will have this, and some may never have it, but I believe that the bulk of consumers will end up with such networks especially on a global basis. The wave I am talking about will probably take at least ten years to establish.
Devices will become the I/0, control, and security for the computing experience primarily. Allowing anybody to use any device. There will be many kinds of devices, and their functions will drive their form factors. Let’s consider four personal devices:
Desktop: This will consist of a flat panel of various sizes and resolution, speakers and mice, other peripherals such as keyboards and pointing devices. The client can be a small device attached to the display, but it can be a tablet or smartphone equivalent
Tablet: This will be a thinner version of today’s tablets and will come in various sizes.
Phone: This will be something that can fit into a pocket/purse
Wearable: This can be a watch or other wearable device.
These devices have only a few things to do. One is to display, get input from the user in various forms including video and voice, and most importantly to provide for identification of the user. For the sake of discussion, I am going to assume we are using a fingerprint.
Once the ID of the consumer is established, the device will display the appropriate information. Using today’s environment as an example, I could see a desktop on my big display similar to what I see on my mac or I could see the home page on my tablet. Of course, the look and feel of all this will change a great deal over time.
There will be other devices in the home that could be shared at the family unit level such as the equivalent of Apple TV or Chromecast. Eventually, these capabilities will be integrated into TVs.
No application runs in the device. Instead, the display, audio, etc. is streamed over the net. Using powerful GPUs this should not be a problem provided that network latency is low enough to support interaction. It is the latency issue that is most important. There are reasonable technical solutions to create networks with high enough QoS to be adequate for the task.
Now for the more difficult conceptual issues
Devices all communicate with a server that manages the consumer environment. There is a very high degree of security between the devices and the server. The server system needs to be highly reliable, but that can be done. An easy way to think of this is having the desktop virtualized and in the Net and then communicating with various applications and services in the cloud.
All consumer data would stay on the server. It would, of course, be backed up and archived in much of the way companies like Dropbox operate.
Now, let’s go back to the example of a photo. If I wanted to post a photo on Facebook, the server would provide a link to the photo. By legal agreement, Facebook would not be allowed to copy the photo. When they wanted to display it, they would have to get it from the server. If the consumer decided to remove it from Facebook, Facebook could no longer be able to show it. The consumer could, however, allow others to have virtual copies of the photo. Once again this would be a link. It is not much different from Dropbox. I can give you a link to a photo on my dropbox, and I can remove your access. The difference here is that you would not copy the photo.
What should Intel do
Intel could use its technology to develop, manufacture and sell all the devices and the servers. It could create an open set of API’s so that other companies can build products including servers that inter-operate in this new environment. Intel could develop the legal framework to make all of this work. It would probably have to bring in many companies, and this may prove very difficult, but I think it can be done.
Intel Capital should focus on investment activity aimed at companies that can build in this new environment.
Intel will need to push various network standards that will allow for the low latency and highly secure networks required by this architecture.
Intel is uniquely positioned to do all this. It has all the essential pieces. Intel can use its semiconductor capabilities to be able to offer the best devices and the best servers. It can use its knowledge of computer security to create a trusted environment that consumers will want. It can use its brand reputation to position this new computing platform.
Please let me know if you would like me to clarify or expand on any of this.