Hacker News Reader

Comment by Pils | Hacker News Reader

Same with the artificial intelligence graph. "Deep" seems to be the most popular category of AI, even more popular than "artificial intelligence," up until 2014. Considering deep learning is a subset of machine learning (which only starts appearing in the graph in 2012), I'm confused what "deep" is supposed to represent. Is there another meaning to "deep" in an AI context?

View on HN

I think it originally just meant "deep neural nets" ie. neural nets with more than one hidden layer. Now it's code for "I'm making magic AI, please fund me".

> Now it's code for "I'm making magic AI, please fund me".

Isn't that just AI? The term has never meant anything but "the future".

Well, no, AI has always meant "making computers do something that, right now, only humans can do." Of course, the term only applies to work-in-progress, because once it's working, it's not AI by definition.

Got a citation for that? Or is that just your definition? As soon as we understand a problem there are better terms that are actually meaningful.

It's a variation of a famous John McCarthy quote: "“As soon as it works, no one calls it AI any more.”

http://cacm.acm.org/blogs/blog-cacm/138907-john-mccarthy/ful...

It's not meant to be a serious definition but IMO the phenomena is very real.

The version I'm familiar with: "today's AI is tomorrow's computer science"

That resonates with me; thank you.

I still think it's a bullshit marketing term.

Reply on news.ycombinator.com