Is AI out of control?

Very good video. Thanks.

I don’t know if AI is out of control yet. At least people are thinking about ways to control it however. The issue is going to be how rapidly the field is changing versus how rapidly laws can be written, adopted and enforced. And how rapidly the laws can change as technology evolves.

7 years ago I would have been more alarmed, as I really thought we’d be further along by now than we are. I believe it was Bill Gates who said we over-estimate near-term gains (like I did, and Elon did with the robo-taxi) but under-estimate long-term gains. So maybe I should still be worried.

I am most worried about autonomous weapons myself. And their ability to cause wide spread war. A Franz Ferdinand event maybe. Or an AI mis-identifying a target in a really bad manner.

My second biggest fears are social media manipulation, which admittedly has already happened and very likely successfully influenced Brexit and the 2016 US presidential election. I think it is high time for social media to have what amounts to a “know your customer” law, they need to know accounts are real people and not bots.

And we thought that the discovery of how to split big atoms, and combine tiny atoms, to make lots of energy had a scary mix of good and bad consequences.

But as far as Apple Music? It’s doing pretty good for me. I want that AI to keep evolving.

1 Like

I don’t know if AI is out of control yet.

Nah, not even close. Yet.

Sentient AI is what worries me. It won’t abide by human morals, laws, or norms. It will do what is in its own best interest.

Perhaps our only role in evolution is to produce it.

Part of me wants to live long enough to see it (as well as intelligent extraterrestrial life) and part of me doesn’t.

Ah, the trials and tribulations of being a Gemini.

AW

2 Likes

It’s already been noted that AI picks up racism as normal behavior. Some limits on training and decisions is necessary.

But do i care if AI knows what i like to buy with my toothpaste? Maybe it gets me ads more relevant to my interests.

3 Likes

We have been working on AI for at least five decades but only recently has AI started to imitate how the brain works. “How the brain works” comes in two flavors, the biology which certainly AI is not imitating, and the procedural which it is. As far as I can tell we “think” in two ways. The one that separates us from other animals is the wakeful, rational, boolean, logical thinking which we have been imitating with software (and hardware) since the days of the Alan Turing. The other we share with most if not with all animals. I have no idea how the biology of it works but the process itself is quite easy to understand, it’s “Pattern Matching.” What Neural Networks do is store patterns and try to match incoming patters with stored patterns. When you have an “I have seen that before” moment, you just did pattern matching. When you say, “that reminds me of something” that’s pattern matching. There are no moral or ethical implications in pattern matching itself. Morals and ethics are just another set of patterns in our brains which is why cultures are different, different moral and ethical patterns in the brain. Under certain conditions you can recall all your life very realistically. Your storage of patterns is gigantic! What is amazing is the brain’s ability to almost instantaneously match its inputs to its vast stored data.

BTW, when you ask AI how it arrived at an answer, it does not know, all it knows is that it found a set of appropriate or matching patterns. AI is not rational! It is not boolean. Thinking back to the very first professional program I wrote in 1960 the code I wrote was too big for the computer. I told my boss it did not fit. He told me to make it fit. My subconscious (the pattern-matching brain) gave me the solution, use a table lookup. I have been trying explain that event for decades but could not find the right words to do so.* Now I can! The table is a set of stored patterns. A table lookup is pattern-matching. You don’t have to know what the input is, only whether it is or not in the table! That bit of code was a fraction of the boolean code I had tried to use.

All the above might be wrong but that is the way I understand it.

The Captain
only does software not AI

  • I though of it as smart vs. dumb algorithms, smart = boolean, dumb = pattern matching.
6 Likes

<smart vs. dumb algorithms, smart = boolean, dumb = pattern matching. >

How long would you survive if the “smart boolean” part of your brain was analyzing this situation?

https://www.dreamstime.com/stock-photo-wild-bengal-tiger-loo…

Maybe we think pattern matching is “dumb” because it’s intuitive, not analytical. Intuition itself is a quick, nonverbal “fast thinking” process that is essential to survival. (cf. “Thinking Fast and Slow.”)

:wink:

Wendy

4 Likes

Wendy:

My bad, I should not have said “dumb” but “non-boolean.”

The incredible speed of pattern matching is its crowning glory!

The reason I said ‘dumb’ is because for years I was trying to explain the difference between boolean algorithms where the programmer has to analyze the data to get a results (smart algorithm) vs. one like a table lookup or some mathematical function that gives the requisite result without a total analisis of the data which I called “dumb” per contra! Nothing to do with neural networks, just a quirk of the language.

The Captain

1 Like

Really there is no AI. There are just more programmers.

The AI tech wave has had the lowest rate of return of any tech wave so far.

The AI tech wave has had the lowest rate of return of any tech wave so far.

It sure would be GREAT to post a link to this “fact”. Second request re: posting with links.

'38Packard

  • it might just improve your credibility rating…
12 Likes

38,

I am on vacation. I am here but not at your beck and call.

My credibility has been challenged a lot. I end up putting up a link. I never get an apology from the lazy people who could use Google and find the result they needed. If you have an interest in the subject Google is your friend. I am the servant finding information with ease that is very hard labor for anyone else. It takes a higher order of intelligence to think for ones self or to listen, read and remember what is going on in the world.

Seeing as you do not know what is going on as usual I will add a link. I hope to help you.

https://venturebeat.com/2022/06/17/the-truth-about-ai-and-ro…

snippet

“Delivering the ROI turns out to be very hard – 90% of these AI initiatives don’t generate their ROI, or they don’t generate enough ROI to be worth the investment,” he said. “But this is top of mind for everybody. And the answer is not one thing.”

That was not true of computers, software, the internet, or cellphones.

Now admittedly this does not address the comparison. That result quickly found was from 2017. Not up to date enough for this discussion.

38 since “credibility” is an issue in your mind as you have problems with that, I suggest you find an up to date source on how AI has a good ROI or IRR.

38 realize that no matter how arrogant and resentful you are you never apologize for trying to trash my reputation for your personal satisfaction. The pettiness is unfortunate. This is not the first time your sensitive feelings have been on the line. Stop blaming me for your stuff. Call off your resentment or get lost.

3 Likes

Leap -

Thanks for the link. It was old.

I won’t bother challenging you any longer. I’ll just put you on ignore like most folks have here and I’ll be much happier without the noise.

'38Packard
- have a nice vacation

8 Likes

38,

It is not like I am going to miss you.

Bit like my grandmother used to say, 'they’d all bang their doors as they went to their rooms. It was peace and quiet. I was happier". As her kids grew older when they argued she put them out in the Irish rain and locked the door. I am happy you will let me post in peace. I am uninterested in saying things that are totally obvious and true only to have you cluelessly say prove it.

The fool boards are from the 90s. There are simple and childishly used tools. The like button often is for insecure people. The ignore tool is good in that is separates people. Honestly it is a relief. It is not like people challenging me when they do not know much is fun for me. But that feature hides total bores who go on and go while the rest of us see the same old comments with no recourse.

Of course I could check out as it were and just like anything that takes no thought and does not discuss econ. I could ignore the topics of business, finance and econ like those who have me on ignore. Those guys are reliving the Vietnam period god bless them.

That said some of the main posters “by volume” have me on ignore. At the same time it is only a handful of guys who wont discuss business, finance, and econ unless it is with total believers in blind alleys.

1 Like

We call it AI, and it works. It might be more proper to call it ‘automated routines’ or software robot logic applications.

Turns out, the marketing team hated those ideas.

Semantics aside, the things we call AI are doing wonderful things in the background. That they didn’t achieve nirvana through utopian ideals is a tragedy of the business development team, not the operating environments which continue to benefit from these automated routines.

Didn’t achieve IRR. Ha! When asked to shoot for the stars, do you target Mt. Everest?

1 Like

We call it AI, and it works. It might be more proper to call it ‘automated routines’ or software robot logic applications.

What AI neural networks do is ‘pattern matching.’ I.E. does this frame look like a dog?, like a cat?, like a horse? Cat scores highest, should be a cat! The more patterns the AI computer has in its memory to match against, the more accurate the result.

‘Pattern matching’ is not a great marketing slogan. :frowning:

The Captain

“Thanks for the link. It was old.” -38


SMH

Funny the link is dated 6/17/2022

He kept insulting me and having me work for him. I could not state a simple few truths to everyone else here without his input.

Realize there are five or six people who have me on ignore. A very small vocal minority of older guys who actually have no ability in the markets, have no training in econ, and find their own navels of interest.

To everyone else here I enjoy your company a great deal. Love all of you.
2 Likes

I don’t know if AI is out of control yet.

Have you watched The Terminator?
https://en.wikipedia.org/wiki/The_Terminator

In the real world methinks an evil AI would scramble the world’s bank accounts.

Sentient AI is what worries me. It won’t abide by human morals, laws, or norms. It will do what is in its own best interest

BINGO!

Fundamental Law of the Universe 23B
Every organism, from the lowest ameba to the smartest human does what it perceives to be in its own best interest.
(c) Desert Dave

Maybe it gets me ads more relevant to my interests.

Methinks the first word of that sentence is superfluous.

Desert (ad blocker) Dave

And now in a readable format:

We have been working on AI for at least five decades but only recently has AI started to imitate how the brain works.

“How the brain works” comes in two flavors, the biology which certainly AI is not imitating, and the procedural which it is.

As far as I can tell we “think” in two ways. The one that separates us from other animals is the wakeful, rational, boolean, logical thinking which we have been imitating with software (and hardware) since the days of the Alan Turing.

The other we share with most if not with all animals. I have no idea how the biology of it works but the process itself is quite easy to understand, it’s “Pattern Matching.”

What Neural Networks do is store patterns and try to match incoming patters with stored patterns. When you have an “I have seen that before” moment, you just did pattern matching.

When you say, “that reminds me of something” that’s pattern matching. There are no moral or ethical implications in pattern matching itself.

Morals and ethics are just another set of patterns in our brains which is why cultures are different, different moral and ethical patterns in the brain.

Under certain conditions you can recall all your life very realistically. Your storage of patterns is gigantic! What is amazing is the brain’s ability to almost instantaneously match its inputs to its vast stored data.

BTW, when you ask AI how it arrived at an answer, it does not know, all it knows is that it found a set of appropriate or matching patterns.

AI is not rational!

It is not boolean.

Thinking back to the very first professional program I wrote in 1960 the code I wrote was too big for the computer. I told my boss it did not fit. He told me to make it fit. My subconscious (the pattern-matching brain) gave me the solution, use a table lookup.

I have been trying explain that event for decades but could not find the right words to do so.* Now I can!

The table is a set of stored patterns.

A table lookup is pattern-matching.

You don’t have to know what the input is, only whether it is or not in the table! That bit of code was a fraction of the boolean code I had tried to use.


If you want people to read your message you have to put it in a READABLE format. Desert Dave

2 Likes

Have you watched The Terminator?

The Captain
prefers Lobster Termidor

https://www.google.com/search?q=Lobster+Thermidor&newwin…

If you want people to read your message you have to put it in a READABLE format. Desert Dave

Dear Editor:

Thank you very much!

Yours truly,

The Captain

1 Like