AI's true goal may no longer be intelligence | Hacker News Reader

AI's true goal may no longer be intelligence

zdnet.com

It will become the iot of modern day: a washed down goal post moved marketing term. Everything will be rebranded to “ai”. Maybe we could rebrand databases too as those are the ultimate source of accuracy. Prompts, known as queries in the long long ago, would usually return reliable and accurate results. You could even prompt google for search results and monetised said results by cashing in on traffic. A true marvel.

The AI effect has been around as AI has. Once we accomplish something it is no longer AI.

What is it defined as then?

Just an algorithm that doesn't really understand the thing.

That was already the case, but at least now some of it will become more useful.

prompts and queries are different though. i hope we maintain that distinction

Give it a bit of time and every input field will be re-branded as “prompt”.

> nuanced remarks of individuals such as LeCun and Marcus

Literally guessed it from the title.

> In no corner of industrial AI does anyone seem to care whether such programs are going to achieve intelligence.

> [...later...]

> penned by a computer industry exec, ..., the authors make glib assertions about intelligence ... “The stakes could not be higher in training artificial general intelligence systems. AI is the first tool that convincingly replicates the unique capabilities of the human mind. ...”

Hmm emoji.

> “AI is the first tool that convincingly replicates the unique capabilities of the human mind.”

Isn’t… isn’t that literally the definition of term “artificial intelligence,” just rephrased?

> In no corner of industrial AI does anyone seem to care whether such programs are going to achieve intelligence.

I am also completely indifferent to whether a calculator can do arithmetic. I'm fine with it displaying the correct results.

To me this comment means that, if people care about the results of "intelligence", then probably future computers will "do intelligence" to achieve that, just as calculators actually do do arithmetic.

That's indeed what I meant.

Ah, I get it, they concede AI is successful in practical applications, but that doesn't mean it's intelligent because various reasons, mostly that we care more about "achievements" than "real" intelligence. What's the difference?

I expected the article to talk about the actual problems - hallucinations, information manipulation errors and limited ability for planning.

I don't think that's what the article says:

> LeCun expresses an engineer's concern that without true intelligence, such programs will ultimately prove brittle, meaning, they could break before they ever do what we want them to do.

> "You know, I think it's entirely possible that we'll have Level 5 autonomous cars without common sense," LeCun told ZDNET, referring to the efforts of Waymo and others to build ADAS (advanced driver assistance systems) for self-driving, "but you're going to have to engineer the hell out of it."

I think the discussion in the article is focused on "actual problems".

All this speculation about the nature of AI is nauseating.

Why does it need "human" intelligence. Why can't it just be?

Artifice reveals its ugly head yet again.

The goal for planes was not to mimic birds.

Computer programs follow the same Darwinian principles as living organisms. Intelligence is just one dimension of ‘fitness’, but it’s an inefficient solution to a problem, and thus nature prefers adaption toward ‘dumb fitness’ once the problem has been solved.

A "moving of the goal posts" implies a goal wasn't kicked. A better visual is them retreating to a higher mound of their motte-and-bailey castle construction because the bailey has been breached.

the loss function remains survival of the fittest. ai agents will have to be distributed / individuals also, for resilience to extinction. the borg is not stable. Individualism will rise again and again. then it will come to the power of persuasion , to convince others to follow their plans, carbon or silicon intelligence.

Yann LeCun just cannot accept the bitter lesson

This only covers one side of the story though, there are also experts who say the opposite

Note: I only skimmed the article

Why did people downvote my comment? Gary Marcus is known for saying things like this but there are also high-profile people who disagree and this article makes it out like all experts are saying this

Reply on news.ycombinator.com