The best news of the last month was something that most people entirely missed. Amidst all the distractions and noise that comprises modern media, a quiet press release discloses that a supercomputer has suddenly become more effective than human doctors in diagnosing certain types of ailments.
This is exceptionally important. As previously detailed in Chapter 3 of The ATOM, not only was a machine more competent than an entire group of physicians, but the machine continues to improve as more patients use it, which in turn makes it more attractive to use, which enables the accrual of even more data upon which to improve further.
But most importantly, a supercomputer like Watson can treat patients in hundreds of locations in the same day via a network connection, and without appointments that have to be made weeks in advance. Hence, such a machine replaces not one, but hundreds of doctors. Furthermore, it takes very little time to produce more Watsons, but it takes 30+ years to produce a doctor from birth, among the small fraction of humans with the intellectual ability to even become a physician. The economies of scale relative to the present doctor-patient model are simply astonishing, and there is no reason that 60-80% of diagnostic work done by physicians cannot soon be replaced by artificial intelligence. This does not mean that physicians will start facing mass unemployment, but rather than the best among them will be able to focus on more challenging problems. The most business-minded of physicians can incorporate AI into their practice to see a greater volume of patients on more complicated ailments.
This is yet another manifestation of various ATOM principles, from technologies endlessly crushing the cost of anything overpriced, to self-reinforcing improvement of deep learning.
Related ATOM Chapters :