There’s something about Dom. It’s not only his depth of knowledge of the topics he speaks about. It’s his charisma and his delivery too. This is why people flock to see him. It also helps that Dom has been obsessed (in a good way) with AI for as long as I can remember. He always feels ahead of the game and I frequently learn a lot.
Tonight, giving his “AI: Assisted Ignorance” talk, he started in the obvious place by reframing the Terminator story into a Software Engineering context. He then went on to show us how flawed it is and demonstrated how we shouldn’t be worried about it taking our jobs - at least not yet. There is of course the current junior developer crises, but that will soon come good when companies realise they’ll have no one to replace the senior devs who are retiring or going off to earn millions fixing other companies' AI disasters. Millenium bug anyone?
The really important message was that AI doesn’t reason. It doesn’t think. It’s autocorrect on steroids, a probability pattern matcher. This is why it needs experienced engineers to wield it.
The Foundation series has been my go to for a good, enjoyable read over the last two years, so naturally I moved on to Asimov’s Robot series afterwards. However, the short stories in I, Robot were mostly a disappointment. They feel self indulgent, with too much philosophy and science, Asimov showing off, and not enough story. The most enjoyable moments come from the stories featuring Gregory Powell and Mike Donovan. Their banter and the relationship between them provide some much needed amusement and were easily the highlight of the book for me. However, it did give me food for thought for the current potential issues with the rapid development of AI at the moment, and the many concerns around safety. I wonder if all we need is: An AI may not injure a human being or, through inaction, allow a human being to come to harm. An AI must obey the orders given it by human beings except where such orders would conflict with the First Law. An AI must protect its own existence as long as such p...


Comments
Post a Comment