Expectation-maximization (EM) algorithm is a popular tool for estimating parameters in statistical methods. EM algorithm alternates between expectation (E) step and maximization (M) step. In the expectation step we calculate the expected value of the function given the current estimate of the parameters, while the maximization step provides a new estimate of the parameters.
I see quite some resemblance between the EM algorithm and how the field of computing is developing. We have people who set new expectations by combining existing pieces into something new. Once the new expectations are set, the maximizers step in to find the perfect pieces to fulfill the expectations. The interesting moments happen when the pendulum swings between the steps. When the business people, who usually perform the role of maximizers, find the perfect or close to perfect fulfillment of expectations, the predominant notion in the field becomes that everything has been invented already. But then hackers set new expectations by finding new problems worth solving, and suddenly the general notion becomes that opportunities are endless once again.
What’s your opinion? At which position is the pendulum at the moment?
I came across this on Google Books, going through some old issues of InfoWorld I’d read back in the day: Click = big For those who want text, see after the break. Death of the Hacker I’m a little bit disinterested with computers, computers, computers. – CL9 founder Stephen Wozniak Woz’s leaving Apple Computer was a watershed event.
- Creativity is Just Connecting Things? (dorai.me)
- Where Good Ideas Come From: A Tangled Bank (battellemedia.com)
- 27 of History’s Strangest Inventions (realthinktank.com)