Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

IA or AI?

vanemden.wordpress.com

Post reply on HN
Print view View original

31–37 of 37 posts

Re: IA or AI?

#31

Post by nickpsecurity »

Practicality wrote: post #19
barteklev wrote: post #15

Earlier quoted context omitted.

I would rethink statement, that AI doesn't change people. In my opinion it does it indirectly. AFAIK it's well documented fact that Google has changed the way our brains work. Would it be possible if Google had no AI? I ask, don't realy know. How about computer games? I daresay that experience created with them can be beneficial to inteligence. Interaction with AI in games happen to be interesting to take a look at.…

Fair enough. I think that we need both. Just that I believe that we benefit more from the IA than the AI. The AI is the tool that gets us there. We need AI, but it's not enough.

It's not even speculation: your point has tons of evidence backing it. I remember when I did AI that "automatic programming" was something we wanted. My research led to a project called The Programmer's Apprentice from MIT. The idea, since auto prog wasn't working, was a more limited A.I. that became an extension of programmer's mind to automate some work, analyse others, optimize even more, and so on. Not sure where that went but a Java programmer's explanation of NetBeans was deja vu. ;) A programmer with a text editor vs one with modern tools (I.A. + A.I.) is a game-changing difference that did effectively increase intelligence of the work.

Likewise, people not so smart with numbers had calculators. People needing conversions have Google and Frick. There's financial and accounting packages that can convert lots of numerical assessments into simpler forms to aid the user's understanding. ERP and BPM, done right, let a person ignore inconsequential details to focus on high-level aspects of business operation. Wikipedia for summaries + Google for details and verification let one amass expertise in a new domain rapidly.

And so on and so forth. Intelligence Augmentation and Artificial Intelligence both have proven value. Both are used today. So, we can keep both. :)

Re: IA or AI?

#32

Post by hyperpallium »

oblig. https://xkcd.com/903/

We already have amplified memory (see also: books, mnemonics). and google amplifies retrieval.

But what is "intelligence", that we might amplify it? For me, limited short-term working memory is an obstacle (EWD's "limited size of skull"). As complexity is added, earlier parts drop out.

There is the "technology" of hierarchical decomposition and the pyschological instinct of chunking, but every problem has irreducible complexity... if this is greater than my working memory, I cannot grasp it.

Artificially enhanced working memory may help here, but I suspect the limit is due not so much short-term memory itself, but it having associations throughout all long-term memory. That is, it's less a cache limit than a bandwidth limit, interconnecting with the entire mind. We aren't Von Neumann architectured.

PS: there's an argument that we might not be able to grasp intelligence itself, if its and its components' irreducible complexity is greater than any person's working memory - even if we formalize a correct model, we mightn't grasp it ourselves. Thus, IA may be essential for AI. Or, AI is essential for AI.

Re: IA or AI?

#33

Post by hellbanner »

dsr_ wrote: post #18

The IA advance which is most obvious to me (yet somehow not yet a reality) is the nomenclator. In Rome, a nomenclator was a slave who remembered people's names for you, and as they approached would whisper to you that this is Gaius Tullius Castor, his wife is Flaminia, his eldest boy is Marcus, and he owns beanfields. A Google Glass camera on your eyeglasses and a speaker in your ear, hooked up to Facebook's face rec…

I highly recommend Things that Make Us Smart: http://www.amazon.com/Things-That-Make-Smart-Attributes/dp/0... "Defending Human attributes int he age of the Machine".

Great read for anyone interested in IA or AI.

Re: IA or AI?

#34

Post by ATLobotomy »

EDW387 [0] (which doesn't have NN0 or NN1 pseudonyms either) seems to be pretty clear about what the "anti-intellectualism" comment was about.

>The undisguised appeal to anti-intellectualism and anti-individualism was frightening. He was talking about his "augmented knowledge workshop" and I was constantly reminded of Manny Lehman's vigorous complaint about the American educational system that is extremely "knowledge oriented", failing to do justice to the fact that one of the main objects of education is the insight that makes quite a lot of knowledge superfluous.

Wish the author went into more detail on why now may be different than during Kay/Engelbart's time.

[0] https://www.cs.utexas.edu/users/EWD/transcriptions/EWD03xx/E...

Re: IA or AI?

#35

Post by rasz_pl »

ATLobotomy wrote: post #34

EDW387 [0] (which doesn't have NN0 or NN1 pseudonyms either) seems to be pretty clear about what the "anti-intellectualism" comment was about. >The undisguised appeal to anti-intellectualism and anti-individualism was frightening. He was talking about his "augmented knowledge workshop" and I was constantly reminded of Manny Lehman's vigorous complaint about the American educational system that is extremely "knowledge…

There is this "Tim van Gelder on Douglas Engelbart, Intelligence Amplification and Argument Mapping"

https://www.youtube.com/watch?v=P77FvUy-NGA

Re: IA or AI?

#36

Post by delish »

> The point I am making here is that Engelbart and Kay were unrealistic in expecting that their technologies would give quick results in the way of Tools for Thought. They had no appreciation for the vast and rich culture that produced the tools for thought enabled by the traditional technologies of writing and printing. They did not realize that a similar culture needs to arise around a new technology with augmentation potential.

I am guilty of deifying Englebart and Kay, and castigating "our society" for failing them. After my honeymoon period with the "tool of thought" people, I've calmed down.

Here's my radical belief: portability is for people who can't write their own programs. (copped from a Torvalds witticism)

Consider writing and literacy: If you really grow up in a literate culture, you can start with a blank page and end with a bespoke document that suits your needs. If you don't grow up in that, you have to modify others' documents. This limits you. Hallmark cards are for people who can't write poetically (no judgment intended).

So too for programming. Today we rely on hundreds of millions of lines of code of others we can't even realistically modify. But I think the future resembles Forth: in less than a hundred lines of code, you write something that suits your needs[0]. You can't do this yet because computers suck.

I'm talking loosely and at a high-level.

[0] I think Forth is a powerful vision for the future: no operating system, no types, no compatibility, no syntax. An executable english language.

Post reply on HN