Hu Resources Replaces Human Resources in Health Care for humans would have been a sufficient reason to switch from paper medical records to electronic records. But the potential for more effective use, for multiple re-use of medical records, adds to the reasons we benefit from electronic medical records interpretable to software through meta-data and other forms of knowledge representation. TIME LINE We cannot say that new intelligences will arrive February 29 of this or that year. The unfolding of these new intelligences is a process, not a single event. To understand the time line of the transforma- tion of hu intelligence, we need to consider both a scalar and an exponential time frame. The human transformations needed draw heavily on further evolving current wet hu functions and behavior. Even with extending current habits, it will take years for humans to change. It will take several annual planning, budgetary, and project cycles for large and currently successful enterprises to change direction and adapt to the changes needed to optimize intelligent wet, soft, and hard hu devel- opment. That is, the rate of many wet hu changes is scalar: we add a fixed amount of change within set frequencies of time. Soft and hard hus appear to develop at an ex- ponential rate, following Moore's Law. Gordon Moore, a co-founder of the silicon chip company, Intel, noticed within a decade of first making chips that the number of transistors or electronic switches Intel employees were fitting into their computer chips was doubling every eighteen months for essentially the same manufacturing cost. That is, the rate of productivity or evolution of silicon chips is exponential, where we do not add on a fixed amount of change, but rather we double the old amount to get to the new (Kurzweil, 2005). Imagine your bank savings account with ac- celerated compound interest where the principal doubled every year and a half. Other systems re- lated to these silicon computer chips also improve exponentially, although often over a longer time frame. Software developers' productivity doubles every five years. Information workers double their productivity every ten years. Bandwidth, or the capacity of telecommunication systems to move larger and larger electronic content, triples every two years (Kurzweil, 2005). Industrial competitors ever increasingly create new and improved models of health care devices 285 Self-Improving Have you worked with a voice recognition or transcription system that improves in interpreting what you are saying as it works with you and has you actively correct it? Perhaps you notice that your own word processor's spell checker learns new words from you. Search engines and online book search systems automatically learn from you and other users. Database management systems in your IT department's computers continually build more and more efficient indexes as its us- ers ask certain types of queries out of the very large number of possible query types. These are examples of systems which are self-improving. These are systems which improve their efficiency and their effectiveness with use. More esoteric are artificial intelligence sys- tems, such as those using genetic algorithms, genetic programming, and neural nets to rapidly generate a large number of solutions, keep the best, generate more solutions from these best, and repeat these steps. The early adoption of intelligent systems will be because they become consistent with routine tasks. But increasingly they will help us with managing complexity and compressing solutions into smaller time frames and using fewer resources. Then they will help us to manage situations so complex that we could not do without these allies. More aggressively, we use thermostats not only to monitor the temperature of buildings, but also, without humans, to cause changes in the function- ing of a building's heating and air conditioning systems. More energy is consumed solely on the orders of the thermostat, a machine.