NFR AI - How It Will Affect Jobs In The Next 5 years

[SurfnFish]

First read about AI in the 80's when reading about a symposium in which William Gibson, the godfather of near future sci-fi fiction that became tagged as cyberpunk and is at the top of my fave writer list, predicted that it was inevitable that the advancement in computers would first lead to autonomous computers that would replace man in many low to mid level jobs, and would eventually evolve into AI machines capable of passing the Turing test, with the next and final step in machine evolution achieving AGI, Artificial General intelligence = sentience.

The Turing test was proposed by math genius Alan Turing in 1950 as a test to assess if a machine can mimic human conversation and behavior so convincingly that it appears indistinguishable from a person. We're already past that.
In April 2025, a study using a specific version of GPT-4.5 found that the AI was mistaken for a human 73% of the time, a rate higher than the actual humans in the same test.

As to achieving human equivalent sentience, that would require a machine capable of developing a soul, that mysterious core at our center that elevates us from animals to humans. And it seems unfathomable to consider that happening to a machine.

 

[Matt B]

One of these:

 

[Porter2]

HAL 9000 was my first memory of AI

 

[509]

Try coaching modern day Boys and Girls Club athletics, it’s real sad how many kids have never even thrown a baseball with their “parents.”

Where do you live Scott??

The reason that "many kids have never even thrown a baseball with their "parents" is they are playing soccer with their kids. In Wenatchee, over 60% of the school kids are immigrants or born here to immigrant parents.

Between the immigrants and their kids that are born here that accounts for ONE-THIRD of the American population today.

I am a immigrant kid, and while my parents stressed education they could never give me a education in American culture and society. The education system assumes that everybody is a middle-class American with those values and cultural orientation. That is a minority of the school population these days. We need teachers and educators that have a background outside the "education" establishment.

There are so many things wrong with "government schools" that I am unsure that we can save them from themselves.

But back to more important things.

One of the best things about becoming an American is I no longer HAD to play soccer. As soon as I discovered baseball I never kicked a soccer ball again. And yes, my father NEVER threw a baseball with me.

 

[Zak]

CAPTCHA= Completely Automated Public Turing test to tell Computers and Humans Apart

Here is the article I mentioned above about about AI translating whale codas (should be a free shared article): https://www.nytimes.com/2025/11/23/...e_code=1.5k8.MnGw.VFzDcgCtYSHU&smid=url-share

 

[SKYKO]

Blade Runner -Replicants, 1982

I forgot about Blade Runner for a sec, what a great movie. I think it was based off a Philip K Dick sci fi novel but can't remember for sure.

 

[RCF]

View attachment 173754

I remember that movie so well! After seeing it, I went to my car and just sat there for at least a 1/2 hour ---> mind blown!

 

[wanderingrichard]

I just read about a cool use case for AI: deciphering sperm whale language/codas. No idea if that's a pipe dream, but wouldn't it be cool if we could understand whale song?

They have been talking and singing for a lot longer than we have!

It's also been used to identify and decode elephant language (yes there is such a thing). Surprise , elephants have names for each other.

Elephants have names for each other like people do, new study shows - Warner College of Natural Resources

Wild African elephants address each other with name-like calls, a rare ability among nonhuman animals, according to a new study. The Colorado State University research suggests elephants do not imitate the receiver's call to address one another but instead use arbitrary vocal labels like humans.

warnercnr.source.colostate.edu

 

[wanderingrichard]

Putting Ozempic in the water supply would fix this.

Might also kill about 1/3 of the population. That stuff is nasty

 

[wanderingrichard]

I forgot about Blade Runner for a sec, what a great movie. I think it was based off a Philip K Dick sci fi novel but can't remember for sure.

It's part of the book "Do Androids Dream of Electric Sheep?" Which brings up ethical questions about PKD's ancestry and personal life i dont even wanna think about.

 

[Scottybs]

Where do you live Scott??

The reason that "many kids have never even thrown a baseball with their "parents" is they are playing soccer with their kids. In Wenatchee, over 60% of the school kids are immigrants or born here to immigrant parents.

Between the immigrants and their kids that are born here that accounts for ONE-THIRD of the American population today.

I am a immigrant kid, and while my parents stressed education they could never give me a education in American culture and society. The education system assumes that everybody is a middle-class American with those values and cultural orientation. That is a minority of the school population these days. We need teachers and educators that have a background outside the "education" establishment.

There are so many things wrong with "government schools" that I am unsure that we can save them from themselves.

But back to more important things.

One of the best things about becoming an American is I no longer HAD to play soccer. As soon as I discovered baseball I never kicked a soccer ball again. And yes, my father NEVER threw a baseball with me.

Dude… “baseball” is symbolic… it’s not just a baseball. It’s why I coach and started a scholarship. It’s easy to bitch and be a keyboard social justice warrior. I hope that through my actions, I’m able to reach some kids and make their lives better in the short term and help them succeed in the long term.

 

[Fourbtgait]

One of these:

View attachment 173754

Lost in space…….
So many youngsters here

 

[offensivename]

Let's get back to the doom and gloom of AI.

Reserve the off tangent rants for other threads.

Representations of "AI" in media are older than anyone here:

I've been around what we are now calling AI for a while, tangentially people I knew were doing stuff with early NL stuff in the early 2000s when I was in undergrad at Georgia Tech. As I've worked in tech fields over the past couple decades the ML side of things has started to become more and more prominent, I'd say the product people in my field first started asking it out and pretending like it would do things it couldn't around 2012-2015. "We can make AI search for that problem" were pretty common declarations (spoiler, none of those great ideas ever came to fruition). The common thing I've seen is that people think AI as we're seeing it today is Skynet, or HAL or MU-TH-UR, but its really nothing like that. Whats really getting used today with ChatGPT, Claude and the like are much more like really smart versions of Google in the earlier parts of this century (which is a weird thing to say but accurate). The impacts are, in my opinion, quite similar as well, but probably magnified. 20 years ago people complained that kids were able to do their homework by googling the answer, finding a relevant article or site and copying that.
I'd say it was pretty easy back in the day because there was a lot less on the internet. As it became harder to sort the wheat from the chaff you had to learn how to google. This is a thing I've had to try and explain to junior developers many times over the years. They would come to me with a problem and, eventually, I'd come back with an answer. I didn't know the answer off the top of my head, I knew how to take their problem, turn it into a google search, know which results were relevant, turn those into more google searches and so on until I solved the problem. Which is really similar to what a lot of "AI" is doing today.
I use AI pretty much every day anymore. Its actively promoted at the company I work for. What it lets me do a lot of the time is take my initial question, which is in natural language, and turn it into an answer by automating the recursive searching and processing that I used to do manually. The work now is in knowing how to ask the agent a question that gives the result you want (there is a lot of things to be learned there) and then reading what the AI spits out and learning how to properly interpret the results. Just like 20 years ago if you google something and use the wikipedia article word for word, or just copy paste the Stack Overflow result into your code blindly you're going to have a bad time.

 

[509]

Dude… “baseball” is symbolic… it’s not just a baseball. It’s why I coach and started a scholarship. It’s easy to bitch and be a keyboard social justice warrior. I hope that through my actions, I’m able to reach some kids and make their lives better in the short term and help them succeed in the long term.

My point wasn't about baseball.

It was about culture. Throwing a baseball or kicking a soccer ball with your kid is not done in all cultures. The family relationships between adults and children are very different depending on the culture.

 

[Zak]

I listened to this audio article in the car the other day and I thought it was interesting. The guy created an AI agent based on himself and had it make calls and interact with friends and strangers for him (starting with spam callers).

My Other Self - This American Life

What happens when people create alternate versions of themselves and release them into the wild?

www.thisamericanlife.org

 

[Stonedfish]

Lost in space…….
So many youngsters here

Indeed Will Robinson. Warning, warning!
SF

 

[509]

I have started using AI in a variety of technical subjects. Here goes with some observations.

As a forester, I have used AI in writing forestry subjects for the layman and used AI. That was interesting. So the typical AI searches the internet. It took my professional opinion and used articles written by journalists at the Seattle Times to tell me I was wrong!! If the AI is trained on mass media just forget about getting any useful information.

I then wrote an article , and loaded it into several different AI engines. But this time I specified that AI could only use Google Scholar for its information base. Google Scholar is limited to published science articles in peer reviewed journals and published legal opinions. That was worthwhile. For example, the article was on mega-fires and biomass as the primarily cause of them. Using the "internet" option yielded me lots of comments about global warming and it being the cause of mega-fires. Using Google Scholar and the professional published papers, climate change was hardly mentioned as a cause of mega-fires.

I did take one of my articles and asked AI to find a list of scientific references that supported my assertions in the article. At first, I was pretty excited since about 70% of the citations were articles that I had read in my professional readings. It was the other 30% that was a problem. It appears that AI found a reference and then appended to it some other references making a complete mess of the citation. There were a couple of citations that I could not find, period. One actually, referenced a living, breathing research forester and credited him with a citation that I could not find.

I did a couple of searches on history questions using AI. I asked AI to give me information on Woodrow Wilson and his historical legacy. It was pretty worthless. I then split the question into two queries: 1). Why was Woodrow Wilson the best President of the 20th century and 2) Why was Woodrow Wilson the worst President of the 20th century. That gave me lots of useful information. That was an interesting learning experience on how to use AI.

I am also into amateur astronomy and the equipment market is today is totally Chinese manufacturers. Who have NEVER been able to write a readable or functional technical manual. I asked some questions on integrating the astronomical camera, with the telescope mount, telescope, and the software package that aligns the scope, finds the objects to image, and finally runs the exposures and processes the resultant images. Pretty complex stuff.

I asked AI to give me step by step instructions for specific tasks. It did that, plus also identified problem areas that were noted by other users, complete with their solutions. That was impressive. I am basically using AI, question by question to write a book on how to use the Chinese astronomy software and hardware.

I gave AI a topic in forestry and asked it to write an article on a specific issue. I then went ahead and wrote my own article on the same issue. Both addressed the same topic, but they were significantly different in tone and writing style. I then sent both articles to friends that were journalists, foresters, and even a couple of civil and mechanical engineers and asked them which article they preferred without telling them that one of them was AI generated. Hands down, they preferred my human written article. The AI article was found to be boring, and said to be written by soulless bureaucrat.

When I was working as a "soulless bureaucrat" writing Environmental Impact Statements I wish I would have had access to AI at that time!

My guess is that AI will definitely change things particularly in the professions. In fact, to protect themselves I suspect the professions will create their own learning databases and then limit public access to the professional AI models.

The public will have the garbage version of AI that is dependent on "internet" learning.

 

[RCF]

The old SISO dilemma.

There was a post earlier about AI using AI generated information to enhance the LLM. In fact, it degrades the AI output.

I applaud you for taking the time and effort and 'bouncing' the output among your friends and peers. The results are very interesting. Not totally unexpected considering how young AI is and also demonstrating that so much improvement is still necessary.

We all know information is key to making money. Google and Meta make a lot of money off of it. So, did you have to pay for access to Google Scholar? The next money maker. How do I invest in it? When I say invest in it, not Google but the stakeholders of scientific, peer-reviewed, authenticated datasets and respositories..

 

[509]

...............We all know information is key to making money. Google and Meta make a lot of money off of it. So, did you have to pay for access to Google Scholar? The next money maker. How do I invest in it? When I say invest in it, not Google but the stakeholders of scientific, peer-reviewed, authenticated datasets and respositories..

Google Scholar is part of the Google search engine.

I did not follow it closely, but the scientific journals tried to limit access to published science articles that were funded by the Federal government. Anything you do as a Federal employee or material produced by Federal funding is in the public domain in most cases.

If the Federal government, followed Canada's lead, and charged businesses and the public for the use of GPS, internet, USGS maps, weather data, etc. etc. and a host of other inventions and products paid by the taxpayers we would probably would not have a deficit.

The journals lost the lawsuit and now have to provide access to all scientific articles funded by the Federal government.

I do not think anybody is even thinking beyond Meta and Google as to providing AI services based on the sciences. I suspect some of the government agencies, right now, are building and funding their own AI systems.

The mass media outlets are suing Meta and Google for using them as sources for training AI without payment. Nobody cares about newspaper articles!!!

The hard science and professions is how AI will change the world. Nobody is talking, out loud, on how to manage that aspect of AI and to whom will it be available.

 

[cody-]

Google Scholar is part of the Google search engine.

I did not follow it closely, but the scientific journals tried to limit access to published science articles that were funded by the Federal government. Anything you do as a Federal employee or material produced by Federal funding is in the public domain in most cases.

If the Federal government, followed Canada's lead, and charged businesses and the public for the use of GPS, internet, USGS maps, weather data, etc. etc. and a host of other inventions and products paid by the taxpayers we would probably would not have a deficit.

The journals lost the lawsuit and now have to provide access to all scientific articles funded by the Federal government.

I do not think anybody is even thinking beyond Meta and Google as to providing AI services based on the sciences. I suspect some of the government agencies, right now, are building and funding their own AI systems.

The mass media outlets are suing Meta and Google for using them as sources for training AI without payment. Nobody cares about newspaper articles!!!

The hard science and professions is how AI will change the world. Nobody is talking, out loud, on how to manage that aspect of AI and to whom will it be available.

Anthropic (the devs of "Claude") seem to be the ones positioning themselves to be the big player in the hard sciences niche. The press release also mentions licensing and deployment of everyone else's models (Google, OpenAI, grok). https://data-science.llnl.gov/llnl-introduces-anthropic-ai-tool-claude-facility-wide-operations

Tip for @RCF if trying Google Scholar without being a part of an institution that gives access to most publishers: always try the "All # Versions" link under an article which often leads to repositories like OSTI, sites like ResearchGate where people post their own preprints, or university websites.