1
2	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Transfer what? Facilitating Invention & Diffusion The Moon Shot Principle—Use-inspired basic research The Priests versus the Shamans Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
3	"Grew up in Melbourne," Grew up in Melbourne, Australia, 1951-1975 Left because I met UCB Prof. Don Pederson in Melb. and helped him with SPICE—Open Source Ph.D. Student, UC Berkeley, 1975-78 Continued work on SPICE at UCB Observed problems arising from student consulting Professor, UC Berkeley, 1979-present Worked with Sematech, MCC, Bell Labs, Xerox PARC, Interval Research, many high-tech co’s. Chair, Department of Electrical Engineering and Computer Sciences, 1999-2000 Dean of Engineering, 2001-present
4
5	"Venture Partner," Venture Partner, The Mayfield Fund, 1988-2002 Helped found 20+ high technology companies Attended partner’s meetings Venture Partner, Tallwood Venture Capital, 2002-present
6	"Founding Director," Founding Director, The Gigascale Systems Research Center (www.gigascale.org), 1998-2002 Established in 1998 to address the challenges of the growing chip design productivity gap Cosponsored by government (DARPA) and industry (SIA) One of six MARCO national research centers Annual research budget of around $8M In 2000 had grown to 24 faculty at 11 universities, 11 postdocs, 56 PhD students, many industrial fellows
7
8	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Transfer what? Facilitating Invention & Diffusion The Moon Shot Principle—Use-inspired basic research The Priests versus the Shamans Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
9	Berkeley Engineering: A Tradition of Impact in Research Pre-stressed Concrete Ground Fault Interrupter Berkeley Unix Relational Database Technology (following IBM) Electronic Design Automation: SPICE to Synopsys RISC (with Stanford), RAID CyberCut online manufacturing systems NOW (Networks of Workstations) Salmon with antifreeze (grapes next?) IEEE Floating Point Infopad (now called WebPad, TabletPC,…) Semiconductor Devices & Modeling MEMS, Smart Dust, … Berkeley faculty are fundamentally motivated by high-potential-impact, long-range research
10
11	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Transfer what? Facilitating Invention & Diffusion The Moon Shot Principle—Use-inspired basic research The Priests versus the Shamans Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
12
13	Principles We must all take a truly global view Berkeley/Bay Area has a unique and differentiated role to play Major research universities are becoming the “DMZ” of research It’s all about leverage Tackle projects of a scale otherwise not possible by a single company Build informal connections to other companies—established and start-up Opportunity to really be in touch with latest developments, at the University but also in the start-up environment—globally Most important aspect is extracting maximum, bilateral value through synergy Requires focused, sustained investment on both sides Tech transfer in IT happens via people True collaboration requires ability to influence research agenda through dialog
14	Why Universities? Major corporate research laboratories of significance have mostly faded away (a few exceptions!), so who is going to do the work? Consortia take too long to assemble and often have very high infrastructure investment costs Universities can be seen as the “choice of last resort” Universities are attractive to very smart people from around the world—at all ages Universities usually have a broad range of human capital representing many different disciplines If they do it right, universities can form the neutral, research commons—the “DMZ of Research” Ideally, a global network of universities, inspired by a common set of grand-challenge “uses” (e.g. energy, education, communicable diseases), could “hold” the global community needed to address such Great Works of the 21^st Century—but that’s another talk!
15	"(1) People" (1) People (2) Markets (3) Products
16	"(1) People" (1) People (2) Innovation (Science & Engineering) (3) Relevant Application Context
17	Observations Almost all projects involve multiple sponsors (federal, state, and corporate) and overlapping research agendas In all cases, the specific research artifacts developed during the projects could have been circumvented relatively easily In all cases, many millions of dollars have flowed to the campus during and since the projects in support of faculty in their research—a tradition of support Conversely, in almost all software & embedded systems cases where technology was protected and licensed, the financial return has not been high and the full potential of the research has been (arguably) compromised
18	Summary Berkeley Engineering faculty are primarily motivated by the IMPACT of their work In ICT, impact is usually maximized by making the work widely available—ACE In other disciplines (e.g. Biological sciences), impact may be maximized by exclusive licensing for royalties—ADC Corporate sponsors that collaborate with Berkeley are primarily concerned about ACCESS to research results In certain areas, access is actually maximized by making work open and available Access clearly includes consideration of background rights and shared rights
19	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Transferring What? Facilitating Invention & Diffusion The Moon Shot Principle—Use-inspired basic research The Priests versus the Shamans Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
20	Impact of DARPA-Sponsored Design Technology Research at Berkeley
21	Transfer What? There is an important difference between Artifacts of Collaborative Evolution (ACE) and Artifacts of Discrete Contribution (ADC) Each may be made by an individual or a group Each may be comprised of, or influenced by, one or more of the other Both kinds of artifacts are extremely important, but must be managed entirely differently to maximize impact (Ed, was that “SUN” or “BUN”?)
22	Value Creation in Product Development
23	Value Creation in Product Development
24	Value Creation in Product Development
25	Value Creation in Product Development
26	The Productivity Gap
27	The Need for Long-Term Research We cannot “solve” today’s design productivity crisis We must change the problem to one we can solve, and where we can demonstrate efficient solutions! We have done this before, but it requires a comprehensive approach to the problem and a long-term investment It is a methodology change, not just a technology change
28	The Gigascale Silicon Research Center http://www-cad.eecs.berkeley.edu/GSRC
29	Artifacts of Discrete Contribution All you need to know is in a paper or a patent or a piece of code The key ideas can usually be exploited quickly via investment The key ideas can be protected and defended efficiently and effectively Usually what people have in mind when they speak of “technology” in the sense of “technology transfer” Nowhere near as common as ACE in terms of contribution and large-scale impact What University Technology Licensing offices usually imagine life to be like
30	Artifacts of Collaborative Evolution Impact is maximized through a community of smart people adding increments of value (technology, understanding, explanation) The “right answer” usually evolves over a long period of time Usually a lot of disagreement (people “not getting it”) in the early phases, and a “seems really obvious” attitude after success Usually a low protection barrier before the community is built To “get it” early you need to be immersed in the community Often preserved and taken to scale as a “standard”—formal or informal Often involves a number of protectable/defensible Artifacts of Discrete Contribution Examples: The Internet & Web, Berkeley UNIX, Window-based UI, TiVO, TinyOS,
31	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Facilitating Invention & Diffusion Transfer what? Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
32	Semiconductor Research Gap
33	Solution to Research Gap
34
35
36
37
38
39
40
41
42
43
44
45
46	Focus Center Metrics First Two Focus Centers: Quality of their Long Range Vision Number of ‘Out-of-the-box’ Thinkers on the Team Authority & Independence of Center Director Nurturing Environment for New Concepts and Radical Alternatives Focus Center Member Companies Buy-in to the Long Range Vision Patient Expectations for Research Results Exploitation of Anticipated Results
47	Elements of Evaluation Which payoff path will focus centers achieve?
48	3 Jobs for 3 Different Groups Technology Management--Create the Shared Vision and Manage the Project Technology Diffusion--Member Companies Buy Into the Shared Vision and Exploit the New Concepts Process Assessment--Evaluate the Focus Center Concept & Execution
49
50	"Protein kinase C activator" Protein kinase C activator Isolated from the stems of the small Samoan tree Homalanthus nutans
51	CITRIS Intellectual Policy Principles “Patents are expected to be rare, but…” If and where IP protection is indicated and desired: Ownership follows inventorship Except when obligations are imposed by State tax-free bond funding of facilities or by 3rd party agreements, existing and new. Licensing: CITRIS corporate participants will all gain non-exclusive, world-wide, royalty-free, access to all IP generated within CITRIS, as will any other company Exclusive, royalty/fee-bearing license an option In effect, we are defining an “Openness Agreement” rather than an “Ownership Agreement”
52	Outline My Background—A Context for my Comments “Berkeley’s” Credentials PeopleàImpactàReputationàBrand ValueàPeople Why is this Topic Increasingly Important Today? The Evolving (Critical?) Role of the Modern Research University Diffusing Knowledge & Understanding An Example: Electronic Design Automation (EDA) Facilitating Invention & Diffusion Transfer what? Impeding Collaboration & Diffusion Example: Faculty consulting in the UC system today
53	Introduction and Principles Advantages: Engineering is a discipline of practice Technology diffusion in ICT is via people Essential for Visibility/Resources/Impact for College and for the University Challenges: Lost time of key faculty for instruction, research and administration Possible conflicts of interest: Bias in student research Bias in student placement
54	UC Consulting/Intellectual Property Guidance January 15, 2003 Employees must disclose all inventions “The University must be able to meet its obligations assumed under legally binding contractual obligations with regard to intellectual property rights generated by faculty and other researchers as a result of sponsored research agreements, material transfer agreements, and other research support agreements entered into on behalf of those faculty and research staff. Therefore, all inventions made by a University employee must be disclosed to the University, including inventions made on weekends, on leave, at home “in the garage,” or during paid or unpaid consulting work. Disclosure is a legal obligation of employment at the University.”
55
56	Other Topics… Issue of Trust Butterfly in China Incubators in the IT world Corporate schizophrenia & IP Sematech vs MCC vs SRC vs MARCO