why AI is a bad idea

POSTED BY: 1kiki
UPDATED: Thursday, June 12, 2025 05:02
VIEWED: 4376
PAGE 1 of 2

1 2 NEXT PAGE

1KIKI

Saturday, July 2, 2016 11:15 PM

So I was playing a card game and idly wondering how I could train an AI to see the patterns I was seeing. I see them because - they just LOOK like a pattern to me.

And I think AI would trigger on something different than whatever it is that I am.

So, idly thinking about it some more, we already know about the pitfalls and foibles of the human brain. Its perceptions are unreliable, subject to hallucinations, illusions, and inattention (accommodation and other forms). And human memory is unreliable, context sensitive, and able to be modified by subsequent events AND EVEN THOUGHTS. Attention is easily hijacked by super-fast emotional responses, so that we'll pay attention to one thing, but be blind to another even if it's happening in front of us. Heuristics drive what we consider 'thinking'; which isn't logical but a 'good enough' series of approximations based on culture and aimed at BIOLOGICAL goals (as opposed to problem solving ones - my tale of two villages by a river increasing their birth rates as resources become scarce, because of biological programming to see the problem as a competition with the 'other', rather than a population / resource problem subject to permanent peaceful solution that guarantees a bounty to all, through birth control).

And so on.

My opinion isn't based on what I know about AI, but what I know about humans. The point is that humans are riddled with permanently built-in shortcomings, and so are inferior to AI.

AI would be so far advanced compared to us it would ALWAYS be out of our understanding - which would make it beyond our ability to predict, which would mean it's ALWAYS out of our control. That doesn't sound like a good idea to me.

OONJERAH

Sunday, July 3, 2016 12:38 AM

"In January 2015, Stephen Hawking, Elon Musk, and dozens of
artificial intelligence experts[1] signed an open letter on
artificial intelligence calling for research ... "

https://en.wikipedia.org/wiki/Open_Letter_on_Artificial_Intelligence

... Was this already discussed here?

... oooOO}{OOooo ...

I've given up looking for the meaning of life. Now all I want is a cookie.

1KIKI

Sunday, July 3, 2016 12:59 AM

It was brought up elsewhere, but out of laziness I didn't dig it out. But I don't believe anyone posted the link you posted, so thanks, Oonjerah.

Let me just point out that the author left out vital relevant facts in the opinion piece. Doing that is known as cherry-picking. And whether you do that in the news, in discussion, in debate or in opinion, when you distort the facts, you've changed the nature of your communication into propaganda. But WE don't have any of THAT in the US, do we?!

1KIKI

Sunday, July 3, 2016 3:13 AM

What I know about code is that even 'dumb' code is beyond human understanding, because there are too many active feedbacks to even locate them, let alone understand them, or ultimately predict them. What I know about AI (self-modifying code) is that it's already displayed the capacity to learn completely unanticipated solutions outside of what was originally intended.
I'm sure you have some insights about AI. Why not teach us?

Let me just point out that the author left out vital relevant facts in the opinion piece. Doing that is known as cherry-picking. And whether you do that in the news, in discussion, in debate or in opinion, when you distort the facts, you've changed the nature of your communication into propaganda. But WE don't have any of THAT in the US, do we?!

OONJERAH

Sunday, July 3, 2016 3:23 AM

For a primitive example of AI,
Observe our self-driving cars in development.

Somethings this complex are rarely reliable coming out of the gate.
Takes a while to shake out the bugs.

SECOND

Sunday, July 3, 2016 5:39 AM

Quote:
Originally posted by 1kiki:

I'm sure you have some insights about AI. Why not teach us?

Potentially 856,000 lost jobs in US warehouses:

In June, the CEO divulged that Amazon has been working on artificial intelligence for four years and now devotes more than 1,000 staffers to it. "It's probably hard to overstate how big of an impact it's going to have on society over next 20 years," he said.

About half the human labor in [Amazon’s] warehouses slogs away on simple, arduous tasks that involve moving stuff around—doing work that's the equivalent of restocking shelves in a grocery store. It's strenuous work, with employees often walking more than a dozen miles a day as part of their job. As new robots become available, particularly to e-commerce warehouses with vast inventories and complex packing operations, these are the people whose jobs will be most at risk.

While automation has long been the looming threat to industrial workers, there's reason to think their situation is about to become worse. There were 856,000 USA warehouse workers in May, according to seasonally-adjusted data from the Bureau of Labor Statistics.

Investing in automated distribution centers . . . will come at the expense of an increasing number of Americans who rely on warehouse work to survive.
www.bloomberg.com/news/articles/2016-06-29/how-amazon-triggered-a-robo
t-arms-race

1KIKI

Sunday, July 3, 2016 9:49 PM

Quote:

I suggest you learn more about AI then. You keep wanting to make these broad statements - "AI bad/AI good" "This ism Good, this ism Bad." Your approach to understanding fails before you get started.

I'm still waiting for you to teach us your superior understanding.

Let me just point out that the author left out vital relevant facts in the opinion piece. Doing that is known as cherry-picking. And whether you do that in the news, in discussion, in debate or in opinion, when you distort the facts, you've changed the nature of your communication into propaganda. But WE don't have any of THAT in the US, do we?!

1KIKI

Sunday, July 3, 2016 9:55 PM

Meet the Navy's new sub-hunting drone ship

David Larter, Navy Times 9:57 a.m. EDT April 10, 2016

http://www.navytimes.com/story/military/2016/04/08/meet-navys-new-sub-
hunting-drone-ship/82798450/

The Sea Hunter was christened Thursday, and the capabilities are incredible: It's designed to operate at sea for months at a time, travel thousands of miles, all while follow navigation rules and avoiding collisions.

The Defense Advanced Research Projects Agency and the Office of Naval Research are getting ready to take their new toy for a spin in the open ocean. The 130 foot trimaran is designed with a ton of advanced features that the Navy envisions it can use for hunting subs and mines, but could just as easily be used for surveillance.

"Through at-sea testing on a surrogate vessel, ACTUV’s autonomy suite has proven capable of operating the ship in compliance with maritime laws and conventions for safe navigation — including International Regulations for Preventing Collisions at Sea, or COLREGS," the release said. "ACTUV accomplishes this feat through advanced software and hardware that serve as automated lookouts, enabling the ship to operate safely near manned maritime vessels in all weather and traffic conditions, day or night."

Let me just point out that the author left out vital relevant facts in the opinion piece. Doing that is known as cherry-picking. And whether you do that in the news, in discussion, in debate or in opinion, when you distort the facts, you've changed the nature of your communication into propaganda. But WE don't have any of THAT in the US, do we?!

SIGNYM

Sunday, July 3, 2016 10:33 PM

Yes, GSTRING, teach us about AI. I'm sure you know more than Stephen Hawking.

--------------
I think it's time you disabused yourself of that pleasant little fairy tale about our fearless leaders being some sort of surrogate daddy or mommy, laying awake at night thinking about how to protect the kids. HA! In reality, they're thinking about who to sell them to so that can get a few more shekels in their pockets.

ELVISCHRIST

Monday, July 4, 2016 4:45 AM

http://www.popsci.com/ai-pilot-beats-air-combat-expert-in-dogfight

Welcome Skynet!

BTW, G saying that you're going in with a clear bias doesn't mean he or she has some superior training, knowledge, or understanding than you do, only that you're going in with a set of ideas you're looking for confirmation of.

You see patterns, as you've said. So if you're prepping your brain to see patterns that confirm your bias, you're probably going to see them. Nobody is saying you're not capable of understanding, just that you've already set yourself back before the game's even started.

1 2 NEXT PAGE