September 26, 2003

A sense of wonder at MIT Emerging Technologies Conference

- by Lee Schlesinger -
I hear frequent advertisements on my local radio station for brain supplements. I got the same effect without a pill by attending Technology Review's Emerging Technologies Conference at the Massachusetts Institute of Technology this week. Speakers in various disciplines provided fascinating glimpses of future technology.

Nathan Myhrvold, former Microsoft CTO and co-founder of Intellectual Ventures, a for-profit company whose mission is to develop inventions, delivered the first day's keynote address. He noted the best inventions are those that grow exponentially, like semiconductors (think Moore's Law), software, and communications bandwidth. With these items, the more you have, the more you want. In order to have exponential growth, Myhrvold said, you need something that works, intellectual depth, and market demand. The next areas likely to meet those criteria, he thinks, include genomics and nanotechnology (though the latter currently fails the "works" test).

In his new venture, Myhrvold said, "I'm looking for people that are narrowly insane" -- that is, people who believe in an idea and have the drive to see it to fruition. This was a theme for many speakers -- that successful innovators are optimists.

Later in the morning attendees listened to four speakers on the topic Where Technology is Heading. MIT professor Edward Roberts noted that pure research in large corporations results in improved growth rate and profitability when it focuses on three factors: linkage of its technology to its business strategy; leveraging its use of resources, notably human resources, outside of its home base, and leveraging its opportunities through alliances and acquisitions; and leadership among its key competitors in growing its technological base. He noted a common weakness in corporate research efforts -- a relatively low degree of participation by an organization's marketing department and CFO.

GM's Lawrence Burns gave a vision of technology in the auto industry. He envisions fuel cell propulsion, hydrogen fuel, and electronic and software systems replacing the internal combustion engine, petroleum fuel, and mechanical controls. (He called hybrid automobiles a transitional technology.) Challenges to that vision include hydrogen storage, high cost, and building a fueling infrastructure. He said that at $50 per kilowatt-hour fuel cells will be practical. Today we're at 10 times that figure, but Burns hopes advances in technology will let us reach the target by the end of this decade. If that happens, hydrogen appliances, electrolyzers, and reformers will replace petroleum refineries and pipelines, he said.

In addition to the propulsion system, ubiquitous GPS and WiFi networking will allow automobiles to be connected at all times, opening up a host of new applications.

Pfizer's Peter Corr followed Burns with a discussion of trends in pharmaceuticals and health care.

But the most interesting discussion came from Intel's David Tennenhouse. He noted that we've passed the one person, one computer milestone. The next wave will include networked embedded computers that give real-time information, enabling proactive computing that can anticipate our needs. Enabling technologies for that trend include physical control of biological and chemical elements; ubiquitous computing in the form of smart dust, RFID, and software radio; planetary-scale distributed systems (PlanetLab is one example); and new probabilistic methods of machine learning.

Tennenhouse sees information technology moving out of the home and office and into the environment. Smart dust will allow us to monitor our environment for safety and health. It will give us a vast increase in the spatial and temporal fidelity of information. Systems will directly query these tiny distributed processors instead of going to a central database to which the remotes have uploaded their information.

Tennenhouse said stochastic modeling and Bayesian networks will help computers infer cause and effect relationships from raw data. It will then be up to humans to decide among the computers' choices for the most promising avenues to pursue.

Self-organizing wireless networks

A panel session on wireless networks focused not on today's 3G or WiFi networking but on a future of self-organizing webs or meshes of tiny independent nodes, each of which routes messages to the others without needing a central authority. Four entrepreneurs presented their visions of this technology. They see embedded wireless devices as invisible technology, running 8-bit CPUs with 4K of RAM and 64K of program storage. They can be used to sense and control larger devices. For instance, they can turn off lights or air conditioners in hotel rooms when they're not needed.

This is commercial, not consumer, technology. All four companies represented sell their products through OEMs and systems integrators.

Among the most common such devices today are Berkeley motes, which integrate computation, communication, and sensing into a single small device. Their peer-to-peer networks can scale to thousands of devices, from which can emerge advanced behavior. The motes run on an operating system called TinyOS, which fosters an event-based execution model and supports flexible component modules.

A key issue with such nodes is power savings, since devices in the environment have to last for years. That requires optimized hardware and communications protocols. Today's devices are quite small -- as big as the tip of a ball-point pen -- and cost only about 25 cents to manufacture in quantity, but require support chips that bulk up the package. The goal is to put all the functions on a single chip.

Another major issue is the need for very low latency routing. Because the devices are often mobile, routing must be reactive, requiring new algorithms, which get complex as the number of nodes increases.

The industry today suffers from a lack of protocol standardization. Various manufacturers' motes use different protocols. Standards bodies are working on the problem, but until these protocols emerge, the industry's growth will be hampered. Nor did any of the participants have ready answers about issues of privacy and security. One panelist proposed using the technology today in places where there is no need for security.

Other obstacles include laws of physics (in the size of components) and regulatory laws. Power use, data rate, and range are all correlative variables, and you can't maximize them all at the same time. However, silicon is getting negligibly expensive, according to Ember Corp.'s Robert Poor, and "at that point it's about the data."

Dell words of wisdom

Michael Dell gave the first afternoon's keynote interview. He covered some of the secrets of his company's success. At Dell, suppliers deliver materials every 90 minutes based on what customers order. The company turns over its inventory about 100 times a year. CEO Dell said Dell Corp. doesn't want to reinvent technology it can get from partners. He said the Internet is not a sales channel, but rather the Web is a front end to a business process.

To determine what businesses his company wants to be in, Dell looks for large markets where there are inefficiencies or high mark-ups. He looks for standards, because markets don't usually become high-volume until standards exist. Dell said standards benefit users, while proprietary hardware benefits only the company selling it.

Dell cited consumer electronics as an area that fits his market criteria now -- a strong hint of what the company announced the next day. Today consumer electronics are "proprietary" in that they don't communicate well with each other. He sees IP-based devices communicating with a central PC, transmitting digital video and audio. Growth areas in the computer industry, he said, include enterprise servers, storage, peripherals, networking, and globalization.

Dell was the only speaker to mention Linux, when he noted that his company works with Red Flag Software to put Linux on computers for the Chinese market.

Dell advised potential entrepreneurs not to be afraid of making mistakes, to get help by building a team around themselves, to be curious and learn, and to do something tremendously better than what others are doing.

Young entrepreneurs

A central feature of the conference was the presentation of the Technology Review 100 -- 100 innovators under age 35. The final panel of the first day of the conference included four TR100 honorees. They gained recognition from widely varied endeavors, from giving robots "humanized intelligence" to making silicon chips using printing techniques to developing a collaborative spam filter.

Vipul Ved Prakash, the developer of the spam filter, began writing his software when spam began clogging his 2400bps modem line in New Delhi. It became an open source project, and gained a lot of users when it was highlighted on Slashdot. Eventually Prakash co-founded Cloudmark and commercialized the software by developing a plug-in for Microsoft Outlook. The back end and protocols remain open source.

All the panelists agreed that they got into their particular fields because they felt they were "cool." They also agreed that they share an optimistic outlook. "You look at the world and you don't see problem, you see potential solution," said Colin Bulthaup of Kovio.

Imagination at work

Thursday's keynote speaker was Jeffrey Immelt, chairman of General Electric. Immelt said he sees times of slower growth and higher risk ahead. China and India will be strong international competitors, while the Internet levels the playing field for price. Immelt said companies will succeed based on their ability to innovate and differentiate.

Immelt talked about leadership and innovation. Good leaders, he said, prepare the organization to innovate, pick the right places to innovate, make innovation pay investors, and know how to use their organization's size as an advantage. He said leadership is about controlling the head, the heart, and the wallet of followers, and aligning them toward the organization's goals.

Part of leadership is personal. Immelt said people in the organization need to feel you're a part of their life, that they're a phone call away from the top of the company. They need to be able to trust the organization's command chain to pass information in an unfiltered way.

Later in the morning, a panel of scientists presented visions of future developments in biotechnology and nanotechnology. Stan Williams of HP Labs said it's physically possible to build a handheld device that can perform the same amoung of work as all the Earth's present computers -- though it may take 50 years.

Williams researches molecular electronics at HP Labs. He looks for people for his team with deep knowledge of computer science, physics, and chemistry, who in talking with each other come up with "eureka moments." His team has discovered that affordable nano-scale manufacturing requires defect tolerance, implemented by significant redundancy and by finding, cataloging, and avoiding defects. He expects an evolution toward hybrid chips with silicon CMOS underlying molecular electronics, with the latter eventually taking on most of the processing load, beginning within the next five years with sensors.

The future of e-commerce

The panel on the Future of E-commerce began with moderator Evan Schwartz introducing some hot areas. Grid computing is replacing enterprise software. Paid context is becoming sustainable. Contextual marketing -- for example, ads on Google searches -- is gaining ground over non-contextual content, such as e-mail lists. Patents are more important than business plans.

The panelists, however, largely failed to follow through on the panel's promise. They came from a diverse group of companies: Red Swoosh provides a data grid for distributing digital media; Peppercoin aims to provide a cost-effective way to process micropayments (payments of less than one dollar); Sprint provides telecommunication services, and appeared to be included only because the company was one of the event's sponsors; and similarly, a representative of OpenWave, the company that developed WAP, appeared to be content-free panel filler.

Travis Kalanick, founder and CEO of Red Swoosh, sees the future of digital media commerce including access by any device, anywhere, to a full library of content, with a commerce backend to pay for it. He said piracy is going to become passé because it doesn't provide the quality of services the digital grid will provide.

Perry Solomon's company, Peppercoin, aims to lower third-party transaction costs that make micropayment sales unprofitable. Peppercoin takes 100 of these transactions (typically for digital downloads, including music, games, and mobile content) and presents only one to the bank via cryptographically secure selection. If that sounds vague, it's all the detail either Solomon or the company's Web site offered.

Solomon predicts multiple business models will co-exist. Today's free, ad-supported content and subscriptions will live side-by-side with pay-per-use. Solomon envisions micropayments to network service providers attached to every email message as a means of defeating spam. His service will launch next month.

Thomas Reardon of OpenWave maintained that phones are becoming richer devices, and are much more likely to grow up than computers and PDAs will be to grow down.

Time well spent

Despite some disappointing sessions and panelists, the Emerging Technologies Conference provided valuable information for anyone interested in innovation. A scan of participants' nametags showed a large proportion of titles like CEO, president, and general partner. If these business movers learn lessons that make them more effective, we all will benefit.


  • Migration
Click Here!