Advertisment

"People underestimate the human mind"

author-image
CIOL Bureau
Updated On
New Update

Lampman has made the long journey with the company having seen dramatic shifts in the company, industry and technology. Initially, he worked with HP’s product development team and then moved to research in 1981. In 1999, he took on the mantle of leading the Labs.  

Advertisment

He has worked on so many technologies that he felt “he changed jobs every five years!” He has been involved in many projects such as the Itanium project with Intel, digital photography and the PA RISC architecture.

Lampman intends to retire this year after choosing a successor in his role at HP Labs. Priya Padmanabhan of CyberMedia News, spoke to him, on topics spanning the changing face of corporate research, new research models and technologies to look out for.

Excerpts:

Advertisment

You have been associated with HP Labs for over two decades. How has the Labs kept in step with industry changes?

When people think of a corporate research lab, they think of the Bell Labs model. A lot of things have changed since the time when Bell Labs was a model for industry research. For companies that invest in research, they have to keep reinventing not just in new technologies but also in the way they invest in research.

There used to be a time when most investment that companies spent was on microprocessors. The industry is now consolidated around a number of small microprocessor companies. Similarly, companies used to make huge investments into operating systems that are also now well consolidated. We find ourselves at research looking at the intersection of technology and opportunity and not just technology. We can work on a lot of technologies but picking the right ones to work on plays a huge part in creating value. If you just follow your curiosity, you may miss out on some of the big opportunities.

Advertisment

How has HP Labs evolved in terms of its research model over the past decades?

Twelve years ago, if you looked at the HP Labs model, it involved universities, HP Labs and HP Business. We used to do our own work, tap into what universities are doing, try and draw the best out of that and deliver new technology for the company. Both customers and HP would benefit from that. Now that chart has become a lot more complicated because we have VCs and small companies that are very good at certain things and not so good in other areas. This is a new model where you have partner companies.

I think that IT industry consists of networks of companies, not just one company that does everything. We work with national labs; we have research people working directly with HP customers not only in helping the customer but also collaborating on pilots and helping the customer get early access to technology before it is for sale. The benefit we get is in learning what the important problems are.

Advertisment

Our decision to offer data center automation actually came out of the dotcom bubble. Most companies at the time were throwing a lot of money at IT. CIOs kept telling us that they were under unrelenting pressures to cut cost and deliver more. But they did not see the IT industry solving those problems, but instead looking at microprocessors and operating systems, which was just a small part of their costs.

So, we actually restructured our programs to say how to attack the total cost and benefit that companies get out of their IT investment. That was ultimately was adopted by HP as a core enterprise strategy and a lot of products and services were developed around that.

Today, this is our biggest program at HP Labs.

Advertisment

We used to do everything from top to bottom. To me, the shift started somewhere around 1990. That was even before the start-ups. Emergence of Unix, the PC operating system and emergence of Intel architecture were the fault lines in the industry. Customers also wanted more choice. These happenings brought balance to the industry.

What are the various milestones that HP Labs has achieved over the last few decades?

Looking at HP Labs as an innovation machine, it has been fundamental in terms of bringing about change in the company. This would range from printing technology-Inkjet and laser-printing technology came from the Labs; PA-RISC architecture came from the Labs and more recently our move into commercial printing. The next generation data centers is something that HP Labs worked on nine years ago-this was well before people talked about it.

Advertisment

Another area is the Itanium-enterprise scale processor project along with Intel. We did a lot of the early work in web services software by contributing our technology for open standards. That was not a product but an important development. Not done but still important is nanotechnology. That is starting to pay off. But it will take a few years more before to see its results. That is going along quite strongly now.

How do you make a call on what technologies would work and contribute to the company’s business and what won’t?

We are constantly looking for technology breakthroughs. If you look at computer systems, it requires a detailed understanding of industry and business problems. For instance, we looked at the underpinning technology for the Semantic web which is the next level of extraction on the web after HTML and XML. We started on this, and developed a toolkit that was open sourced and is well used. The first application of this solves the frequent customer problem of integrating different kinds of databases. This is when the technology intersected with a fundamental business problem.

Advertisment

The Labs’ role is to look at new sources of value for HP in future. It is really exciting since it gives researchers a chance to think broadly on what they do. The other thing we did is to take a high level view of the opportunities in HP to set a framework for researchers.

After internal debate, we choose a set of topics that would be long-term drivers. One of the things that characterizes research is many independent ideas coming up all over. An oft used phrase used is, ‘Let a thousand flowers bloom!’ Many ideas popping out is great, but all ideas are not equally valuable.

If you look at India, one of the things one looks at is the price of technology and rising incomes in India. But is that enough to make IT more usable? What we look at are the non-economic issues that can make IT more usable. The emphasis is on language, mechanisms, telecom infrastructure and address issues that can make HP more accessible.

In some ways, we are creating a new type of research that is not done in the labs, and get business inputs in their thinking. There is more emphasis on industry-specific solutions and services.

In your years at HP Labs, which are the technologies that did not live up to expectations?

There were certain areas where excitement exceeded reality. Artificial intelligence (AI) was one area that people were very optimistic about. AI has become very important in machine learning. But in the early days, people underestimated the sophistication of human mind. That led to enormous optimism. Machines are good at certain things, but then there is the complexity of the human mind.

We have a limited understanding even to this day of what’s pre-programmed and what’s not, and the bandwidth of things that people absorb. The idea of replacing human intelligence is too farfetched.

Another area was virtual reality. This turned out to be more of a niche technology. A thing that anyone in research has to accept is that you have to accept whether some technology can be important or not. Data center automation is important. Heating and cooling of data centers is something we worked on before people ran giant data centers. Sometimes the technologies become important and hit inflection points. And sometimes they don’t. The Labs’ role is to take more risk and also accept when they fail.

And those that could take off in future?

Biotech and nanotechnology will deliver in ways never imagined before. We are just scratching the surface of that now and it will take a long time. If you look at something like the integrated circuit and see what it has done today-right from the first transistor to where ICs have become the prime force of the economy, it has taken four to five decades.

How has the global spread of Labs aided research?

In the `90s, we did a lot of work on Linux and Open Source, a lot of people considered it academic curiosity and not something important from the business point of view. We thought that maybe Linux would turn out important and it did. Work on that started in Europe. We were fortunate that we had a lab in Europe. One of advantages of having a lab in other locations is that though the technology world is flat, it actually has a lot of bumps and valleys. For instance, for mobile phones, the US is the worst place to understand what’s going on. But Japan and Korea are the places to watch.

We shifted our model from the initial big lab model. The issue with that is when there are many technology centers in the world, we needed to be in those places to tap into those technologies, meet market requirements and be close to customers. So it made sense to have more labs for good coverage.

The tech industry has seen the prevalence of the NIH (Not Invented Here) syndrome, in the sense that a lot of big companies like Cisco would rather buy out companies rather than build it on their own. What are your thoughts on this trend?

Cisco is aggressive on acquisitions. What people forget is after that acquisition, suddenly, the smaller firm’s team becomes part of Cisco’s work force. So it increases Cisco’s internal R&D capability. I think within the IT industry, what you see ultimately is a kind of a mix. The benefit of acquisitions is that it makes you move more quickly. It is a way to fill in a missing piece. Two things are important in shaping a research program: one is to have a clear vision and second, to build a map saying here are the things we are missing. It is important to look at which ones can be filled and which ones to do ourselves.

Doesn’t this approach add redundancies to the existing research program?

This used to be a problem. If you look at HP, in the mid-90s, there was a lot of concern about whether start-ups were replacements to doing our own work. We decided to look at start-ups the way we approach universities.

So we felt we should regroup and focus on something else. We have numerous examples like that. People did not realize that start-ups were an efficient way of doing certain things and not very good at others. VCs look for high returns in a relatively short (period of ) time. If that’s what you are going to do, you can’t do large enterprise scale projects as a start-up.

What happened after the bubble was that people started focusing on building products that people really needed. Our own attitude changed. Our job was not to compete with start-ups. They were just a new player in the model. We decided to take advantage of this. In one case, we were asked to help a startup called Indigo that was into commercial printing. We worked with them and helped refine their technology and HP ultimately ended up buying the company. So that’s part of a new model.

If we see ten start-ups in a certain area, we won’t probably invest a lot of effort in that area. But start-ups are not a cure for all problems.

What kind of research can be outsourced?

I think some parts of research can be outsourced. There are examples like open innovation where we let people try and solve problems. It comes from understanding how you can create value. For example, in the PC industry, everyone uses OEMs, and in standard systems, that is more efficient. Variations tend to come through quality and product selection and distribution. All the systems are built of the same components. But you can differentiate in the software.

Take last year’s laptop battery crisis. HP did not have this problem because that was one of the areas where we decided we do that on our own. Though we use suppliers for a lot of components, we decided to do batteries on our own since we considered it a sensitive area. We were not just lucky that it did not happen to us but the fact that we had engineers working on it. So though some areas may be standardized, you can decide which area you want to specialize in.

How do you measure innovation? Is it in terms of IP and patents or something else?

Patents are not a good indicator of innovation. Economists seem to like them because they are things that are countable. We look back at the biggest thing we did in a program. Sometimes we give an idea, or come up with a prototype that was turned into a product. One way of doing this is backward looking metric. We tend to keep a kind of an honor roll by looking at key businesses or changes in businesses that would not exist if it weren’t for our work. In general, people are in agreement about what these things are. Then there are other soft contributions that involve work with standard bodies.

The higher the risk, higher the return. Most people conclude that if you use very rigid metrics like the kind you use in business, you run the risk of driving behavior you don’t want. You actually want Labs to take risks and not be successful all the time. (HP Founder) Bill Hewitt once said, many years ago, “If the Labs is always very successful, it must not be trying very hard!”

The real value of Labs is that we have helped define a lot of HP’s evolution as a company. Then there is a forward-looking metric that is softer. There is an engagement model at every level from the strategic to front-line engineering teams. They will see programs where the paths are well aligned.

© CyberMedia News