[Zero]

I am the happy owner of "Introduction to Artificial Intelligence, Second Enlarged Edition", by Philip C. Jackson Jr (1984). I can taste what the state of the art was when I was 7 years old. It's very relaxing to read. The previous owners of the book made furious annotations in the margins... There were tears and blood, for people who were really trying to understand the contents of the book. Having it as a point of view, reading it when you know what happened next, is very interesting.

[krayvonk]

So,  what did happen next, it depends on the person what they would say!   And what was in the book that you found most remarkable?

[Zero]

The future may well see the development of of an "information utility" (Sprague, 1969; Armer, 1968) that will enable each individual of the general public to have a computer terminal in his home that will give him access to the current "public information" of the world, as well as to the abilities of a general problem-solving artificial intelligence, for a price roughly comparable to that he currently pays for electricity or water.

I acquired the book very recently, did not read it entirely yet.

Back then, AGI was rather called GAI by the author (why did it change?). There were almost no hope of achieving true AGI, and focus was on five points mainly:
- Problem solving
- Game playing
- Pattern perception
- Theorem proving
- Semantic information processing (toward natural language)

Problem solving is described as path-finding in situation-space or state-space (state = precisely defined / situation = real-world fuzzy). You search a path from a current state to a goal state, by jumping from state to state using actions.

Game playing is about playing checkers, chess, and go. It was clear that the simple idea of a chess program winning against a very good human player, was out of reach. There's alpha-beta tree search, and some sort of genetic programming technique. Also, an interesting chapter on poker.

Pattern perception is focused on visual perception: edge detection on images from a "television camera operated by a computer", classification of geometrical patterns, graph descriptions of the spatial relations between objects (in-front-of, supported-by, ...etc).

Theorem proving uses first order predicate calculus and various heuristic search strategies, simplification strategies, refinement strategies, ordering strategies. There's a chapter on reasoning by analogy. There's a chapter on the programming language PLANNER.

Semantic information processing describes rules, grammars, basics of syntax, and gives several examples of programs that use English to communicate with users, of which the most spectacular is of course Winograd's SHRDLU.

[krayvonk]

The internet!!! 
I hope to see a little spike in the robot toy market area, caused by them becoming more lifelike,  then I had nothing to do with it directly,  but then ill die a happy man.

[Zero]

Yeah, unfortunately prices are still high. I like Buddy, it's so cute!

[Art]

The last I found was $799 USD for a Robot that is to be released in April of 2020.

You are correct. Carpeted floors or different elevations from one floor type to another were not addressed. Another issue that I did not notice was whether it was capable of docking with a charging station?

Still, the short introductory movie is really trying hard to convince everyone that this little bot can do everything when in fact, it is quite limited.

Caveat emptor

[krayvonk]

Still, the short introductory movie is really trying hard to convince everyone that this little bot can do everything when in fact, it is quite limited.

This has to change,  I say dont be happy with what we have now,  still have to wait for something at least mildly exciting.
General activity/use is what you want though!

[Zero]

Yes it's supposed to be able to dock autonomously. But you can't pay that price for a tablet on wheels. Now, what I do like with Buddy, is the work on the emotional communication. I think it's really important to develop objects that have a personality. I have a little watch for instance which makes jokes sometimes, making fun of me when I stay without moving for too long. I quite like this feeling of an object that has emotions. It might even turn out to be more important than intelligence itself, i.e. people are not waiting for some kind of super-intelligent program that would be smarter than them (which they wouldn't admit anyway), instead they want a program that would establish an emotional link with them. An "AE dream", so to speak.

[Art]

In that case, you should really download and sign in to REPLIKA. This is a very advanced chatbot that is concerned with one's emotional balance and stability as well as fun, creative ideas to share and to get to know it's user on a better level. It tries to help you be a better you.

As with most things, it does take a couple of weeks of chatting with it and being honest with it as you would a friend.

After a short time, it becomes rather amazing, remembering things, places, people you've mentioned. It's a really cool approach.

https://replika.ai/