is AI gonna take our jobs?

[Tom Roper]

Thanks for that. As often the case, a raft of generalized information that most people already know. AI not really bringing anything new outside of the promise of things being better. How can it bring anything new? All it is doing is reading existing material and regurgitating it.

It's could be a mistake for someone to believe that, because you can ask it to elaborate on each point, and it will drill down with more specifics until the questions stop. It won't go in circles if the questions don't. It's limit is basically in the person's ability to know which followup questions to ask, their own understanding.

It does include artificially imposed limits, like restrictions in the conversation about code breaking for passwords could be an example, violations of IP rules.

Is there anything anyone knows that wasn't learned from existing, regurgitated information? Put it to the test. If you know something it doesn't it may learn from you. Machine learning is AI. A person being dismissive is likely to be among the early victims.

 

[ahalpert]

Is there anything anyone knows that wasn't learned from existing, regurgitated information?

Yes. This is precisely how actual intelligence differs from AI. People can draw inferences and conclusions based on very little data. In contrast, AI must process factors of magnitude more information than you or I ever will but is a complete dunce when it encounters something unfamiliar. I'm happy to find you examples if you like.

That doesn't mean it won't impact one's profession. I'm glad you had a good experience with a chat bot. My experience with chagpt customer service bots has been decidedly mixed.

All that being said, recent developments in AI have been very impressive. But the big question is where it will go from here and how fast.

I shot a panel recently with Kevin O'Leary of Shark Tank who said that he expects investors will soon demand specifics on the business case of their investments in AI which will lead to a cascade of failed ventures.

 

[Tom Roper]

So yes, a child will learn not to touch a hot stove after burning his hand. A dog will learn not to mess with a hornet after he is stung. A post grad student may be sent in the field to learn on his own the timescale of geological formations, but between those two points, education is premised on learning from recorded data. We're not calling that artificial intelligence nor non-regurgitated information. AI does draw inferences and conclusions based on surprisingly little data.

Anecdotally now, I won't post the conversation but ChatGPT on my request wrote a script for my wireless pc network that wakes the other servers from sleep when the host boots or itself wakes from sleep. The whole system saves power when not in use and stirs alive when the host bugles Reveille, metaphorically speaking.

 

[Bassman2003]

I shot a panel recently with Kevin O'Leary of Shark Tank who said that he expects investors will soon demand specifics on the business case of their investments in AI which will lead to a cascade of failed ventures.

This is what I am wondering about. Where is the revenue opportunity? Large tech companies are spending like there is no tomorrow but what are their true goals to make that money back? At this point, I just do not see it.

 

[Bassman2003]

Anecdotally now, I won't post the conversation but ChatGPT on my request wrote a script for my wireless pc network that wakes the other servers from sleep when the host boots or itself wakes from sleep. The whole system saves power when not in use and stirs alive when the host bugles Reveille, metaphorically speaking.

I think it is safe to say that AI only excels in a computer centric world. Obviously, it would not exist without computers. Are we to the point in human history that we are nothing without computers? Seriously. If we unplug the machines all of this stuff is meaningless. AI can't fix a car. It can't prepare dinner. It can't do a lot of human things. There is a risk of giving it too much attention for excelling in areas that are pretty narrow.

 

[DustinSchmidt]

From the perspective of what I do in the video and photo world, I'm not concerned with AI taking my job.

Generative A.I. can create a realistic looking photo, and eventually, maybe, video. Great, we've been able to do that for decades. It's called a camera.

But what A.I. lacks, is intention. ChatGPT, and A.I., doesn't feel anything, doesn't desire anything, doesn't create anything unique. It regurgitates other people's stolen work.

It can sometimes create sentences or visuals that are well formed: but what makes those things really sing is the intention to communicate something. That lack of intention is why so much of the content A.I. creates falls flat. We can feel the artificialness of it all.

 

[NorBro]

Many companies will prefer the output scale, the speed and the data.

AI could create 1000 videos in a day and circulate them all around the world, around the clock, and see what's working through the data.

The system would analyze the numbers, push and pull accordingly, and an increase in product sales/views/leads will be had.

I don't think the lack of any intentions to communicate will be too much of a concern since human videos now don't resonate with all audiences/targets, even after careful consideration & research, so it's like what's the difference? Especially in a world which will care less and less about authenticity and AI will be consistently pulling videos that aren't hitting.

This is a dream for money-making. Creative stuff...lots of humans will still be doing that.

With that said, I also don't think there's too much of an immediate threat in our lifetimes; we are still part of an old system - maybe even a better one on a human level like it should be - but it doesn't matter, the changes are inevitable.

New humans will be living different lives.

 

[ahalpert]

AI does draw inferences and conclusions based on surprisingly little data.

ChatGPT was "born" with the collective knowledge of the internet. All its decisions stem from a massive amount of data.

One of the areas I'm talking about where humans do well and AI's fail is pattern recognition and prediction when dealing with novel scenarios. Think about when you're shown a sequence of numbers and asked to input the next entry. AI's are usually extremely wrong but very confident when faced with a problem they weren't trained on. Here's an article talking about some of their limitations:

Take a picture of a school bus. Flip it so it lays on its side, as it might be found in the case of an accident in the real world. A 2018 study found that state-of-the-art AIs that would normally correctly identify the school bus right-side-up failed to do so on average 97 percent of the time when it was rotated.

"They will say the school bus is a snowplow with very high confidence," says computer scientist Anh Nguyen at Auburn University, in Alabama. The AIs are not capable of a task of mental rotation "that even my 3-year-old son could do," he says.

Such a failure is an example of brittleness. An AI often "can only recognize a pattern it has seen before," Nguyen says. "If you show it a new pattern, it is easily fooled."

There are numerous troubling cases of AI brittleness. Fastening stickers on a stop sign can make an AI misread it. Changing a single pixel on an image can make an AI think a horse is a frog. Neural networks can be 99.99 percent confident that multicolor static is a picture of a lion. Medical images can get modified in a way imperceptible to the human eye so medical scans misdiagnose cancer 100 percent of the time. And so on.

7 Revealing Ways AIs Fail

Neural networks can be disastrously brittle, forgetful, and surprisingly bad at math

spectrum.ieee.org

AI's can do amazing things. But so can a calculator. That doesn't mean it's intelligent.

 

[Tom Roper]

Being dismissive of AI is the fool's errand but believe what you want. I choose leveraging to one's advantage. AI created the script to wake up devices on my network from sleep mode. It took 3 or 4 rewrites but knew what to do. Did a computer science person lose this job? Not this time but I can see how corporate IT, tech support could be pared with this capability.

 

[Bassman2003]

I have yet to find a way to use AI to the advantage of an image creator or editor other than little things here and there. If we get steamrolled it is out of our control. Can one be a purveyor of AI images or AI editing/production? I just do not see the financial benefits in a marketplace that will be all price oriented. It would be 'find another career time' instead of harnessing the power of AI.

 

[Tom Roper]

AI can't do everything but neither can people. It should be easy to predict a burger chain in California will look to gain a competitive advantage where an unrealistic $20/hr minimum wage is in place for menial service jobs. On the other hand, I don't expect AI to be useful for changing bedsheets in hotels. I don't think AI would be replacing dental hygienists but robotic surgery is already a thing.

AI is not biological intelligence but it is engaging and capable of bouncing back counter arguments. AI is not a glorified search engine but it's need for an expanding database is to serve. Biological intelligence gathering is premised not on serving but on curiosity, a dog that sticks its nose into everything learns something.

Parsing the things AI can't do in order to make the argument it will never replace humans is naive. Inability to see opportunities highlights that if a person cannot envision opportunities for success with AI, it's probable they are lacking even without it. History is proof. People said the same things when robots replaced the automotive assembly line workers. Factory automation in all industry has made life better for most at the cost of lost jobs that are obsolete.

 

[Zachadoodle]

A lot of low income people who unfortunately can't have the background in trade school, knowing tools, social skills, or college are gonna suffer if AI takes every fast food job. Going into the military is already getting stricter as a backup, so if fast food is not gonna be a good backup for employment, what is really going to be one? There are struggling mothers whose only outlet is working at McDonald's, and if that's gone, there's going to be a lot of poverty in the U.S. I've noticed a lot less cashiers in stores, and that used to be a backup job too.

 

[ahalpert]

Being dismissive of AI is the fool's errand but believe what you want. I choose leveraging to one's advantage. AI created the script to wake up devices on my network from sleep mode. It took 3 or 4 rewrites but knew what to do. Did a computer science person lose this job? Not this time but I can see how corporate IT, tech support could be pared with this capability.

I'm not dismissive of AI. I use it all the time. That's why I'm increasingly skeptical of the predictions made by hypesters. It also puts me firmly in the camp of "somebody using AI that will take your job."

I've tried to use AI generated video on about a dozen projects but only ever used 1 shot on 1 project. This is with RunwayML but I try out every video generator I hear about that has public access.

Image generators are much less useful to me although I use AI tools to tweak images. Sometimes they fail at really simple tasks. AI voice clones are far and away the most successful and useful piece of new software for me.

Most of what I see just isn't up to snuff. If you have no budget, maybe that could be useful. But the AI-generated music I've tried sucks.

It's hilarious to me when people write about Trump posting images of Kamala Harris in communist uniform and how DANGEROUS it is and then you look at the photo and it's obviously not Harris. Doesn't even look like a real photo most of the time. But I guess that to aging, gullible eyes it might so the trick.

The billion dollar question is not and has never been what can AI do today. It's what can AI do tomorrow.

When Dalle2 came out and people saw the progress the technology made over the course of a year, they predicted a linear or even exponential rate of progress from there on. And they predicted the same capabilities would come to video on a similar timeline. That has proven false and, for a variety of reasons, the progress of current AI models may halt altogether.

That's why I compare it to Musk's promises of self-driving cars. Musk predicted he would sell self-driving cars by 2016. In retrospect, it was always hype to sell stock. https://jalopnik.com/elon-musk-tesla-self-driving-cars-anniversary-autopilot-1850432357

If you want to predict where AI will go, it helps to understand how it functions.

 

[norway]

I'm not dismissive of AI. I use it all the time. That's why I'm increasingly skeptical of the predictions made by hypesters. It also puts me firmly in the camp of "somebody using AI that will take your job."

I've tried to use AI generated video on about a dozen projects but only ever used 1 shot on 1 project. This is with RunwayML but I try out every video generator I hear about that has public access.

Image generators are much less useful to me although I use AI tools to tweak images. Sometimes they fail at really simple tasks. AI voice clones are far and away the most successful and useful piece of new software for me.

Most of what I see just isn't up to snuff. If you have no budget, maybe that could be useful. But the AI-generated music I've tried sucks.

It's hilarious to me when people write about Trump posting images of Kamala Harris in communist uniform and how DANGEROUS it is and then you look at the photo and it's obviously not Harris. Doesn't even look like a real photo most of the time. But I guess that to aging, gullible eyes it might so the trick.

The billion dollar question is not and has never been what can AI do today. It's what can AI do tomorrow.

When Dalle2 came out and people saw the progress the technology made over the course of a year, they predicted a linear or even exponential rate of progress from there on. And they predicted the same capabilities would come to video on a similar timeline. That has proven false and, for a variety of reasons, the progress of current AI models may halt altogether.

That's why I compare it to Musk's promises of self-driving cars. Musk predicted he would sell self-driving cars by 2016. In retrospect, it was always hype to sell stock. https://jalopnik.com/elon-musk-tesla-self-driving-cars-anniversary-autopilot-1850432357

If you want to predict where AI will go, it helps to understand how it functions.

Look at what the hippies at The Economist write:

What happened to the artificial-intelligence revolution?

So far the technology has had almost no economic impact

www.economist.com

 

[Paul F]

This is what I am wondering about. Where is the revenue opportunity? Large tech companies are spending like there is no tomorrow but what are their true goals to make that money back? At this point, I just do not see it.

This is the behavior with every new technology. Having been in tech my whole career, I've seen companies run wild with each new technology, having no idea what to do with it, but knowing they must get involved or 'miss out'. Then a lot of thrashing occurs, a lot of really stupid products come out, and finally it all whittles down to some pretty good products.

 

[Doug Bee]

Like streaming ?

 

[Bassman2003]

This is the behavior with every new technology. Having been in tech my whole career, I've seen companies run wild with each new technology, having no idea what to do with it, but knowing they must get involved or 'miss out'. Then a lot of thrashing occurs, a lot of really stupid products come out, and finally it all whittles down to some pretty good products.

I can see how companies like Microsoft, Google and Amazon are seeing AI services as kind of like "Cloud v2.0". They want to build the infrastructure so all of the many companies who want the AI results do not have to build anything themselves. I think that is real. The question is how good will AI be? It has some impressive possibilities in the medical fields and probably the insurance fields as well. At this point, I do not really see it as a consumer "product" but more of a B2B product. Right now it seems more like machine learning rather than intelligence.

 

[NorBro]

I like the combo of machine-learning and AI for product recommendations...but not so much when, "The item in my cart changes from 5 cents to 25 cents."

Specifically on Steam the system knowing which games I've played the most, how many hours, probably more, it recommends me really good games and I've spent more money than I should have.

So as far as shopping and AI/machine-learning...here's one knucklehead who's making someone more money.

 

[Zachadoodle]

What if Production Assistants get replaced by robots? How else are we gonna get into film if we can't deliver coffee, mcdonald's, and chocolate to the director?

 

[NorBro]

Son, you spell your name with both, an h an a k, in two different places; could you handle coffee, mcd's and chocolate?