Hacker News Reader

Comment by telesilla | Hacker News Reader

I am of the age of these kinds of devices and I don't think, A.I. was really a concept in how it is these days. I remember robot after robot in popular sci fi but outside of the Asimov stories my Dad read, I don't remember thinking about robots as intelligent in the way we think of A.I. now, as being autonomously able to think without being specifically programmed. I think I thought they were kind of clever logic.

View on HN

You must have been just a little too young to have seen 2001 and been caught up in those 60's days of sci-fi. HAL-9000 was a big deal and set the tone for expectations of AI with fully conversational, human-like behavior.

AI is an unattainable goal, not because we'll never create it, but because "what is AI" changes over time as we figure out how to do things, and give them their own name. Decision trees, machine learning, deep learning - each of these was "AI" at some point, but now that we can do them the magic is gone and "AI" has become more distant than it was before.

(Heavily paraphrased from a longer post I read years and years ago, no idea where)

To answer "What is AI?" you first have to answer "What is I?"

Check out the Apple ][ at the Dolphin Research Center in 1982, and Ron Reisman’s thoughts about dolphin intelligence:

https://www.youtube.com/watch?v=CWHCTNztnwQ&t=392s

“It’s gotten to the point that I never say anything about intelligence in general. I don’t know what it means any more. I used to. But then I started trying to test it. And if you think about it for a while, you don’t know what it is.” -Ron Reisman

The best way I heard it phrased was "AI is whatever computers can't do yet".

I think among common people what AI is is clear - a computer that is as smart and clever as a human. That has never changed, and has never been achieved.

But in academia what you say applies since they first thought "well a computer that can hold a conversation must be intelligent", then they created ELIZA, then "well a computer that can win in chess must be intelligent", then we got Big Blue, "well a computer than can recognize objects in an image", etc etc

I get the point but also fully disagree. At the point where computer cognition is good enough for humans to truly fall in love, build friendships or come to hate as an enemy, I personally think not recognizing it as AI will become a fringe belief.

I don’t agree with that. Having my grown up in that era there is t any shortage of content in popular culture that have smart or sentient AI:

- Asimov novels (as you’ve referenced yourself)

- 2001 (Hal)

- Star Wars (R2D2 & C3PO for example)

- Buck Rogers in the 25th C (Twiki & Theopolis)

- Battlestar Galactica (Cylons)

- Alien (Ash)

- Star Trek TOS (I forget the episode names but there’s a few with smart AI)

- Star Trek: The Motion Picture (V’Ger)

…All from either the same era as 2XL or before.

The problem with “AI” as a term is that it can mean multiple things (like “HD”, “smart”, etc) so it will be used and abused by whoever feels it might sell a few copies of something.

An example of this is computer games. AI has never been used to refer sentience but rather the programmed logic of a computer opponent.

Then you have AI in the research sense; which are the technologies another poster described.

You have AI in stories, TV and movies which are often depicted as human-like intelligence, complex reasoning, or even outright sentient.

It was never a term that had one definitive technical definition.

Now you do see researchers retrofitting different classifications of “AI” in a hope to clear up the confusion around the use of the term “AI”. But that’s a more recent (in comparison) development and certainly isn’t something the general populous are aware of.

In the 70s/80s there were a lot of intelligent robot characters in popular fiction. Metal Mickey, Little Wonder, robot in Lost in Space, K9, Twiggy from Buck Rogers, Muffit from Battlestar Galactica, the ship in Flight of the Navigator, Transformers (sort of), T-bob (M.A.S.K), tons more I'm sure. It was very common for a squad to have an intelligent robot in, or a kid-hero to have an intelligent robot sidekick.

Star Wars!

The 1920 play that literally introduced the word "robot", Karel Čapek's "R.U.R.", featured fully autonomous AI robots.

There were plenty of "thinking robots" in Sci-Fi of that era. Asimov aside, one example that immediately comes to my mind is Mike the computer in "Moon is a Harsh Mistress", and the titular villains in Saberhagen's "Berserkers" series; both are from 1960s.

Jonny 5 is alive!

A decade later but you are right there was plenty of AI back then.

I’m sorry Dave…

Basically, by the "if it has so much as a single decision in the code, it's AI" definition a lot of people like to push, this is actually AI.

What do you suppose AI is if not “clever logic?”

Reply on news.ycombinator.com