Is the ultimate smart skill no longer smart?

Towards a fair, thriving Web in the age of AI

In this issue:the notion of intelligence, again. The ultimate smart profession, no longer smart? The benefits of a Standard Plan for requirements.

Discussing the "Intelligence" in "Artificial Intelligence"

It has become fashionable to state that AI is not really  “intelligent” because  it supposedly “does not understand”. I discussed this matter in an earlier newsletter (the New Year edition, January 1st) which I later adapted for a Communications of the ACM blog article. The blog article will be republished as a regular “viewpoint” article in the magazine; it had to be shortened for that purpose and (as is often the case) the shorter version is, I think, better. You can read a preview.

I also wrote a development of the ideas in my own blog, with the title “Yes, AI is intelligent. Now prove me wrong”. That title is provocative and I predictably got some flak, but the challenge stands: since in so many respects AI tools show behavior that (if it were human behavior) would indisputably be viewed as intelligent, sometimes very intelligent, the burden of proof is on those who claim that it is not intelligent. (Arguments of the form “it does not do X” or “it sometimes messes up” are by themselves worthless, since they also apply to people considered intelligent.) I am still waiting for such a proof, refuting the assertion of the blog article's title.

Quote of the Week

Everyone thought that software programming was the ultimate smart profession. Look what is the first thing that AI is solving. Software programming. And so it turns out that the definition of smart is very different than what most people think.”

Jensen Huang (CEO of Nvidia), in the A Bit Personal podcast with Jodi Shelton.

What do you think?

Requirements plan

One of the contributions of my requirements book is a “Standard Plan” for requirements documents, divided into four “books”. The plan, refined over many years, has by now been widely applied, in both industry projects and university courses. It uses the book's classification of requirements into four categories, the “four PEGS” (Project, Environment, Goals, System).

The plan is reproduced below and explained in the corresponding chapter of the book, which is freely accessible (from the book page).

Complementary material on using the Standard Plan can be found on the Web page of the “Companion Book” by Jean-Michel Bruel, Sophie Ebersold and  Mariya Naumcheva: see here.

One of the benefits of applying the Standard Plan is that it provides a checklist of questions to be asked about the project (and its environment, goals and system). Forgetting any of these questions is a risk.

Try the Standard Plan on your next project!

 

Cover photo: Taj Mahal, Agra, Uttar Pradesh, India.