The main idea behind the Duke-NUS IT infrastructure is to place tools on the desktop of my computer that facilitate access to information, thinking, problem-solving and learning. My strategy is based on the assumption that the information available through the Internet, which I call my Internet-memory, will enhance many facets of my biological memory, and that web tools (e.g. web-services) will enhance my execution of tasks. (The real reason I am including the Internet-memory in my list of recall resources is that it helps me avoid the forgetting curve - i.e. learning things that I'll use infrequently and thus forget.) It is probably obvious why Internet-memory and web services will enhance problem solving. What is probably not so obvious is the behavioral change required to develop what I call the Internet-reflex, a reflex that immediately sends us to Google as a preface to problem solving (has someone else solved this problem or are solution hints readily available).

Each day brings problems to my table. For each problem, I collect data, look for common threads (analysis) and eventually synthesize a draft solution. The draft solution is often incomplete, thus triggering another cycle of data collection, analysis and synthesis. Analysis and synthesis are thinking while data collection is both doing (acquiring new data/insights) and remembering , (recalling facts and insights). If I forget something or remember it incorrectly, then my analysis and synthesis will be off target. If I find that I don't know something then I either have to go learn it or ask someone else - and both take considerable time. All these activities are work for my brain - and along the way, I must decide how much effort to allocate to learning, recalling, doing and thinking. (Sometimes, my brain is just running in neutral - sort of fermentation of ideas and impressions. This happens while I'm running - and I don't know how to label fermentation.)

Thinking is not possible without a substrate for building new ideas and synthesizing solutions. Actually, the substrate is more like a scaffoding, where I can hang new pieces of information. I am unable to remember unrelated facts without great effort. Give me a frame of reference, i.e. some scaffolding, then remembering is a piece of cake. There are two kinds of substrates:

• facts or observations and
• concepts that link facts and observations.

My early years started the life-long process of building my substrate for thinking. This time was spent learning core concepts and absorbing experiences - a mix determined jointly by me and my parents. Graduation from my youth placed me face to face with life, which was complex and required thought before deciding which path to follow. It's my observation that each day, I have a finite amount of energy I am able to devote to neural activities before I burn out. I must decide how best to partition this energy between learning, remembering, doing and thinking. Each day would be more fun if I could shift some of the time I spent remembering (sometimes the wrong stuff) to something more interesting like thinking. Similarly, if I could limit my learning to that necessary to solve my problem, then I could retarget the time usually spent mastering things I do not need or will immediately use. I suppose its the child in me that delights in building something new as opposed to blindly remembering something old. Similarly, its the rebel in me that delights in demonstrating I don't need to master all the discipline in order to solve a serious problem.

The reallocation of doing effort from me to software tools has been painless - simply shifting things I did not enjoy doing to a program that had no feelings. The more difficult reallocation of effort has been that of using the internet to enhance segments of my memory and to facilitate learning. Commodity computing + global connectivity + search engines are available to extend my memory if I learn a new skill - that of building a high yield Google search before trying to remember an obscure fact or asking someone else. The good news is that the internet memory is available 24 hours/day, 7 days a week reduces my dependence on my error-prone memory and enables high speed self-learning. The bad news is that I have to unlearn my reflex of trying to recall obscure facts or asking someone else for the information and unbelieve the paradigm that content-based learning is superior to problem-based learning. Because the internet memory is global, integrating it into my learning and recalling enhances my problem solving by bringing the world's data, speculations and ideas to my desk. This new resource has become the basis for exploring new ways to retarget my mental energy, shifting away from error-prone and possibly low-yield biological remembering and toward high-yield internet-remembering. The left over time is available for new thinking or learning or jogging.

Content mastery is the 20th century learning paradigm and is based on mastering concepts and memorizing facts. Problem-based learning is becoming the 21st century learning paradigm - limiting those concepts to be mastered to those relevant to solving today's problem. With the internet-memory, problem-based learning brings us an efficient and competitive way of working.

Some internet accessible information is high quality - while other information is trash. Improving my skill with critical and analytical thinking will be essential for efficiently sifting through search engine results to locate good stuff . With an internet memory, remembering is effortless (requires a click) and, in contrast with biological memory where unlearning obsolete concepts and facts requires considerable energy, internet unlearning is trivial - simply erase the element from the internet memory. I can look at my biological memory as a sort of cache memory - keeping only themost frequently used concepts and facts. The internet memory then becomes my main memory. My use of the internet memory has another advantage: it shields me from memory loss as I age. The end result is that each day reveals a more internet-centric Frank and a less Frank-centric Frank. To use these new tools, my team is building an internet-centric work environment by creating an IT infrastructure based on