Postulation: Is anyone else concerned with the proliferation of AI?

[Bill_Bright]

Don't just move on. It was a genuine question.

Yes, it was a totally genuine question. The question is not the problem. Some of the response are.

Some here are convinced they totally understand AI.

As Lex puts it, "moose muffins".

To make matters worse, they so convinced they fully understand AI today, they have convinced themselves they fully understand where AI will be next year, or five years from now.

Bull feathers!

Some are convinced AI cannot be independent because it takes a human to press the start button. Hogwash! I gave the example of my tech encountering an unexpected obstacle. I noted he can improvise, modify, and veer from standard procedures as the needs arises - all on his own. That is being "independent". You claimed AI is not capable of making those independent decisions.

Hogwash!

What is clear to me about AI is that it is still in its infancy, but rapidly growing by leaps and bounds. AI has gotten it wrong many times - but lessons have been learned by those mistakes too.

But hey! What do I know? Unlike you, I don't fully understand AI. But then neither do the folks at MIT, or Google. Perhaps you can explain to researchers How AI Knows Things No One Told It.

Or maybe you can educate these scientists: Scientists Have a Dirty Secret: Nobody Knows How AI Actually Works

Perhaps you can show the world how AI is "pretty harmless". For I sure can't.

 

[freeagent]

Have you ever seen moose muffins? They are huuuuge.

I wonder where a bull might have feathers..

I used to gargle with hogwash, tastes a bit gamey, but its not bad after awhile.

I wonder how many "levels" of AI there are?

There's the kind we play with, the kind runs stuff, but what about the really high level stuff?

Like powered by a dark pyramid type of thing..


And Google is like yeah, we found the multiverse, and everyone is like oh, ok..

 

[AusWolf]

Yes, it was a totally genuine question. The question is not the problem. Some of the response are.

Some here are convinced they totally understand AI.

As Lex puts it, "moose muffins".

To make matters worse, they so convinced they fully understand AI today, they have convinced themselves they fully understand where AI will be next year, or five years from now.

Bull feathers!

I never said I fully understand AI. What I said is, everything in a computer is a bunch of 0s and 1s, that's what I was taught at school, and unless someone shows me otherwise, that's what I believe today. I'm happy to learn, but by hearing "that's not what it is" or "that's not how it works" from left and right, I'm not gonna learn anything. Tell me what it is, tell me how it works, don't just throw vague ideas around.

Some are convinced AI cannot be independent because it takes a human to press the start button. Hogwash! I gave the example of my tech encountering an unexpected obstacle. I noted he can improvise, modify, and veer from standard procedures as the needs arises - all on his own. That is being "independent". You claimed AI is not capable of making those independent decisions.

That's also not what I claimed. What I claimed is that while AI is capable of doing the job independently, it is not capable of understanding context around the job - why it needs to be done, who's asking for it to be done, what are the implications of it not being done, what happens if it's done, etc. It also doesn't think outside of the job. Once it's done, it's not gonna start doing something else. Someone prove me if I'm wrong.

What is clear to me about AI is that it is still in its infancy, but rapidly growing by leaps and bounds. AI has gotten it wrong many times - but lessons have been learned by those mistakes too.

But hey! What do I know? Unlike you, I don't fully understand AI. But then neither do the folks at MIT, or Google. Perhaps you can explain to researchers How AI Knows Things No One Told It.

Or maybe you can educate these scientists: Scientists Have a Dirty Secret: Nobody Knows How AI Actually Works

Perhaps you can show the world how AI is "pretty harmless". For I sure can't.

Thanks for the links, I'll have a look. Just let's drop the attitude, and have a normal discussion, shall we?

 

[Shrek]

I don't think we're beyond the point of no return yet. But we need to recognize the dangers approaching us and take decisive actions now.

Will it help to take action as a bad player will not follow suit.

 

[freeagent]

I wonder if the real high level AI is people who have uploaded themselves just before they passed..

Freaky shit man.

 

[AusWolf]

Will it help to take action as a bad player will not follow suit.

That's hypothetical. We still need to do our best, regardless. One player cheating the system doesn't mean that the system is wrong.

 

[lexluthermiester]

Have you ever seen moose muffins? They are huuuuge.

Yes they are. It's my polite(r) way of saying "bullsh#!".

Some here are convinced they totally understand AI.

Which is not possible as no one person can know everything about AI, even those who work on AI.

 

[ShrimpBrime]

I wonder if the real high level AI is people who have uploaded themselves just before they passed..

Freaky shit man.

Now there is a concept for AI that I could get behind.

Grow brain organisms integrated into silicon based computing where we would call this species sentient with having the potential to have organic processing using things such as instincts, perhaps fear of death, hindsight and evolutionary future both biologically and mechanically.

_________

But why go through all this trouble? If you want to produce something near human intelligence, wouldn't it be easier to just procreate? The gene pool will produce more Einstein like people. It's just a matter of the odds....

 

[Shrek]

I believe the intent is to greatly exceed human intelligence.

 

[lexluthermiester]

I believe the intent is to greatly exceed human intelligence.

Not sure I would agree with that as a blanket statement on a world-wide basis. There's likely someone working on such an effort, but it seems implausible that it's widespread..

 

[dragontamer5788]

AI is never a real tool. AI is always marketing.

When a tool graduates from research/hype into... a real tool.... it gains a real name. Search Engines used to be "AI" from the 80s for example, but today we call them search engines. Automated Theorem Provers were called AI before, but today we call them Logic Programming, Automated Provers, 3SAT solvers or whatever (depending on the technique). Etc. etc.

AI will never be a problem because AI never exists. It has never existed, AI is always the dream.

When we come down to reality, LLMs, Automated Generative Painters, and other such stuff will likely become useful as tools. In fact, I think the "Generative AI" painters are at least being used as reference material for art, if not as quick cheaply made commercial art.

LLMs is the interesting one to me. There's so much money being thrown at LLMs but its successes seem... arbitrary. I think people think that LLMs can replace programmers and managers or something. But that's just not what LLMs do, they're closer to incredibly fancy autocorrect (and indeed, their algorithms are about "guess the next word" given the context). Ex: people are using LLMs as summary-bots (but clearly failing. LLMs as summaries consistently fail to find the thesis of documents).

So if we specifically talk about LLMs (probably the biggest hype "AI" today), its weird. There's something interesting about these LLMs but its not clear how to use them to high efficacy. I know my coworkers like to pretend that they can make good Emails or whatever, but conciseness is how I was taught and having LLMs pad out extra paragraphs of uselessness is against my writing style. And LLMs never beat me in a conciseness or straightforward discussion. (LLMs always verbosely meander about).

I believe the intent is to greatly exceed human intelligence.

Google can search documents faster than you can. Cars can run faster than you. Etc. etc. Tools are tools. Tools always have beaten humans at their job.

You are right: the intent of modern AI is to "greatly exceed human intelligence", or "AGIs" or some bullshit. But that's never been where computers are successful. Computers are fundamentally, a tool. Something that should improve someone's job. Excel makes people analyze patterns more easily. Word helps people write documents (spellcheck, grammar check, etc. etc.). Photoshop helps make illustrations. Etc. etc.

If the goal is... to deliver better advertisements. Well guess what? That was all computers before and remains all computers today (be it AI or not). A/B Testing of advertisements and seeing which humans prefer which ads always was a thing done by computers, and automatically at that. If backend algorithms are replaced by AI, a lot of the modern internet isn't going to fundamentally change. (Do you think a human was recommending you new items after each Amazon purchase? No, that was, and always was, a robot).

 

[lexluthermiester]

Tools are tools. Tools always have beaten humans at their job.

But until now they've never been able to "out-think" us. We are on the bleeding edge of that reality.

 

[ShrimpBrime]

But until now they've never been able to "out-think" us. We are on the bleeding edge of that reality.

If you compare an AI to a brain, the program is simply running a task.

Your brain runs the heart.
A LLM is running only that, off or CPU/GPU (brain). The LLM is just a function of that brain.

So we are trying to code a bunch of functions. This does not dictate intelligence however.

We ascribe definition of consciousness to self awareness levels.

So we just have to code as well as biological evolution to create our level of intelligence. Then figure out how to run that code on PC hardware. Or perhaps emulate it on our latest X86 processors ....

Perhaps describing a program as intelligent is difficult for me, I dunno.

 

[lexluthermiester]

Perhaps describing a program as intelligent is difficult for me, I dunno.

I think you're viewing AI very differently than most others. You have an interesting perspective though.

 

[dragontamer5788]

But until now they've never been able to "out-think" us. We are on the bleeding edge of that reality.

Four color theorem - Wikipedia

en.wikipedia.org

Initially, this proof was not accepted by all mathematicians because the computer-assisted proof was infeasible for a human to check by hand.[2] The proof has gained wide acceptance since then, although some doubts remain.[3]

Others took up his methods, including his computer-assisted approach. While other teams of mathematicians were racing to complete proofs, Kenneth Appel and Wolfgang Haken at the University of Illinois announced, on June 21, 1976,[17] that they had proved the theorem.

Automated Theorem Provers have been proving mathematics (that humans couldn't prove) since 1976. If you're in the realm of computers, its just a long list of things computers (with the right algorithm) suddenly become better than humans over... and over... and over again.

------------------

Maybe the only thing is that humans have run out of "things" that count as thinking. In the 90s it was chess. Then 00s was handwriting recognition, 2010s was the game of Go and Watson for the game of Jeopardy. Now we have LLMs and Generative AIs encroaching upon a new subject but its the same old same old for the past 50+ years.

Once the hype dies down and Watson is forgotten, all that remains are the useful tools and the artisans adapt to learn how to use the new tools.

 

[lexluthermiester]

Automated Theorem Provers have been proving mathematics (that humans couldn't prove) since 1976.

Really with cherry-picked example? I suppose next you're going to mention the Chess match which involved "Big Blue"?

I did not say "out-perform" us, I said "out-think" us. No computer anywhere has done that yet.

 

[Vayra86]

What I'm seeing is all the stuff that happens around AI.

We have a world and economy right now, where Big Tech is on a power trip and governments are fast losing that battle. Between tax evasion and Musk in an administration now, things are really escalating and we're getting pretty close back to the Middle Ages, where a select few owned all the things and the rest was kept in service with scraps and leftovers, fear, uncertainty, and doubt. The erosion of social securities and government institutions, the erosion of justice and trias politica.... they're all telltale signs of it. Not to mention the political influencing that tech CEOs do now. Zuckerbergs' recent moves to make Facebook another X are to be viewed in a similar trend.

There is no way AI is going to help us with anything here. Its a tool to improve the grasp on power. Information is power, information control is a much greater power. Right now, Big Tech is literally making tomorrow's news, and we're all clicking.

Welcome to the dystopia

 

[dragontamer5788]

Really with cherry-picked example? I suppose next you're going to mention the Chess match which involved "Big Blue"?

I did not say "out-perform" us, I said "out-think" us. No computer anywhere has done that yet.

You misunderstand the example. Chess was "thinking", until everyone decided it wasn't thinking.

Computers have always been better than us at various things. And new algorithms make computers better than us at new things. Always been true, always will be true.

Now there's a human need for ego and self-worth. Which is what drives the changing definitions. Maybe that's finally been beaten down this cycle.

What I'm seeing is all the stuff that happens around AI.

We have a world and economy right now, where Big Tech is on a power trip and governments are fast losing that battle. Between tax evasion and Musk in an administration now, things are really escalating and we're getting pretty close back to the Middle Ages, where a select few owned all the things and the rest was kept in service with scraps and leftovers, fear, uncertainty, and doubt. The erosion of social securities and government institutions, the erosion of justice and trias politica.... they're all telltale signs of it. Not to mention the political influencing that tech CEOs do now. Zuckerbergs' recent moves to make Facebook another X are to be viewed in a similar trend.

There is no way AI is going to help us with anything here. Its a tool to improve the grasp on power. Information is power, information control is a much greater power. Right now, Big Tech is literally making tomorrow's news, and we're all clicking.

Welcome to the dystopia

Well that's a bit more political, but its just the 1870s again with the Railroads and Telephone networks taking over the country.

We're at the later stages where the public is suddenly aware at what we've given the "robber barons", and the need to clamp down on the behavior. With Google being declared a monopoly last year, we're on the path towards fixing the larger ills of society. But it will be a slow slog to fix things up over the next decades. It wasn't until the 1910s that worker reforms were enacted, and not until the 1930s that society finally made something resembling the modern era.

We have the benefit of peering back into history and seeing the problems of monopolies of the 1880s though. So maybe we can have a faster go at things this time around.

Much like the late 1800s, money buys more money, which buys power which buys more money. Eventually Democracies grow tired of that cycle and changes the rules to be more fair.

 

[ShrimpBrime]

I think you're viewing AI very differently than most others. You have an interesting perspective though.

I believe that intelligent can't be described or ascribed to a computer program or algorithm in programming. It's an intelligent program. But it does not have intellectual thoughts or thought processes similar to a human to exceed our level of intelligence.

But a belief is no different than going to church. So I shouldn't use this word, but I don't think AI can believe something. It won't have this type of thought process for a very long time to exceed our intelligence.

Perhaps it's the what where when how and most importantly WHY do we want to consider a computer program intelligent that I don't understand all these simultaneously.

Is it even artificial?

 

[dragontamer5788]

I believe that intelligent can't be described or ascribed to a computer program or algorithm in programming. It's an intelligent program. But it does not have intellectual thoughts or thought processes similar to a human to exceed our level of intelligence.

But a belief is no different than going to church. So I shouldn't use this word, but I don't think AI can believe something. It won't have this type of thought process for a very long time to exceed our intelligence.

Perhaps it's the what where when how and most importantly WHY do we want to consider a computer program intelligent that I don't understand all these simultaneously.

Is it even artificial?

Do you even have a specific algorithm that you're talking about? As I stated earlier, AI is a false marketing from the start. At a minimum, you should get more specific.

Are LLMs intelligent? Have you tried asking them how many r's are in strawberry yet?

-----------------

Now with the right prodding and the right algorithms, you can get ChatGPT to get around the "Strawberry" problem.But there's all kinds of flaws being pointed out in these things today. All you have to do is use these LLMs for a bit and you'll notice them yourself.

I've talked about this earlier but summaries seem to be the biggest "false-use cases" of LLMs. And that's where I see the big dangers of LLMs (and other "AI" tools). LLMs are terrible for summaries, but there's a whole bunch of dumbasses out there... powerful dumbasses, who are shoving LLMs into summary-bot positions.

 

[Vayra86]

We're at the later stages where the public is suddenly aware at what we've given the "robber barons",

Are they? I'm not seeing much of that. There is a tiny percentage that seems to see it. There's a far bigger group that's reducing that group to a bunch of whiners that should just STFU.

So far its mostly Europe that has taken real steps with things like DSA and GDPR. The US? asleep at the wheel if you ask me, and its going to be really interesting where the power balances towards the next 4 years. Either way, you're right, I don't want to make this topic political, and I'll leave that part of the discussion here I'm mostly looking at the regulation part of the deal.

 

[ShrimpBrime]

Do you even have a specific algorithm that you're talking about? As I stated earlier, AI is a false marketing from the start. At a minimum, you should get more specific.

Are LLMs intelligent? Have you tried asking them how many r's are in strawberry yet?

View attachment 379765

-----------------

Now with the right prodding and the right algorithms, you can get ChatGPT to get around the "Strawberry" problem.But there's all kinds of flaws being pointed out in these things today. All you have to do is use these LLMs for a bit and you'll notice them yourself.

I've talked about this earlier but summaries seem to be the biggest "false-use cases" of LLMs. And that's where I see the big dangers of LLMs (and other "AI" tools). LLMs are terrible for summaries, but there's a whole bunch of dumbasses out there... powerful dumbasses, who are shoving LLMs into summary-bot positions.

Lol.
I totally follow. AI it's self, the naming alone is marketing. So we just call useful programs AI and it sells like crazy.

But I don't know what AI (programs) took jobs away from Lex (OP) stated 3 of them. So it's difficult to be specific I suppose.

Intelligence is a very strict term we use to define biological creatures. I see totally why the word does not apply here.

 

[dragontamer5788]

Are they? I'm not seeing much of that. There is a tiny percentage that seems to see it. There's a far bigger group that's reducing that group to a bunch of whiners that should just STFU.

So far its mostly Europe that has taken real steps with things like DSA and GDPR. The US? asleep at the wheel if you ask me, and its going to be really interesting where the power balances towards the next 4 years.

And yet: https://apnews.com/article/google-s...e-department-84e07fec51c5c59751d846118cb900a7

U.S. regulators want a federal judge to break up Google to prevent the company from continuing to squash competition through its dominant search engine after a court found it had maintained an abusive monopoly over the past decade.

The proposed breakup floated in a 23-page document filed late Wednesday by the U.S. Department of Justice calls for sweeping punishments that would include a sale of Google’s industry-leading Chrome web browser and impose restrictions to prevent Android from favoring its own search engine.

We are ahead of the year 1870. But we aren't in like 1910 where public awareness is fully aware of the problem. I'd say we're probably in like 1900s where novelists and other writers are beginning to write and/or publish stories like "The Jungle". There's an awareness to the problem but we haven't quite made the case to the general public yet.

 

[lexluthermiester]

You misunderstand the example. Chess was "thinking", until everyone decided it wasn't thinking.

Not everyone thought it was "thinking". It was just running through sequence of possible move on a moment by moment basis. No "thinking" was taking place.

Now there's a human need for ego and self-worth.

That has nothing to do with the concerns being expressed. For you to say that shows you haven't paid attention to the bulk of the ongoing conversation.

 

[dragontamer5788]

But I don't know what AI (programs) took jobs away from Lex (OP) stated 3 of them. So it's difficult to be specific I suppose.

"AI" didn't take anyone's jobs away. Dumbass bosses who were sold on non-functioning technology took away people's jobs.

Job loss is an economic problem. If companies cannot figure out how to use workers in an efficient manner to changing tools that's a whole different problem and ballgame.