How the AI apocolypse starts

[Isaac]

Not urging but asking.
Why it's not the other way. First dive deeper, research, and come to a conclusion. If necessary, sign it.

Doesn't it depend on the nature of the "conclusion"?

If you and I were on an island with lots of problems, would you think it odd that I issued a decree that said :
Conclusion: we have lots of problems. Create task force to figure them out in more detail and take action

 

[The_Doc_Man]

Doesn't it depend on the nature of the "conclusion"?

If you and I were on an island with lots of problems, would you think it odd that I issued a decree that said :
Conclusion: we have lots of problems. Create task force to figure them out in more detail and take action

No, it would just prove that you were once a government employee.

 

[Isaac]

No, it would just prove that you were once a government employee.

I was almost a government employee, a few years ago I worked AT the dept of revenue for Arizona, through a contracting firm. When the manager, beaming and smiling, handed me my great treasure and prize - an offer letter at 30% less than I was making contracting, I feigned a weak smile and signed it. Then found a better job a couple months later! Government jobs go from paying less-to-more:
Fed > State > County > City

If I could get in at the City level, that would be ideal. Still get the state retirement system

 

[The_Doc_Man]

The "working through a contracting firm" is exactly what I did... except that it was through 8 or 9 firms over a 28 1/2 year period. But working for the U.S. Navy (indirectly) led to some interesting benefits. For example, they didn't count it as a break in service when one contracting company lost the contract to another company. That was because the losing company suddenly had lots of employees but no slots for them, and the winning company had lots of slots but no employees. So <wink><wink><nod><nod> suddenly we worked for someone else, usually at the same desk and with the same phone number - but because we had greater experience than someone fresh off the street, we had seniority. By the time I retired, I had LOTS of seniority and a retiring salary to match - which figured into social security benefits.

Retirement plans were always 401(k) for that job track, but no sweat. After the first job change and the first 401(k) rollover, it was just another bother to manage when contract changes came around. We had to get new security badges because our employer had changed and clearances derive from your employer's clearance. The only other issue that almost caught me was on the badges, when my fingerprints didn't match. (OK, they really did - but the last job change occurred when my hands started to get wrinkled and there were extra lines for the fingerprint reader.) Fortunately, the supervisor could see the original lines matching up and understood what was happening, but it was still somewhat strange. The AI in use at the security station couldn't tell - but there WAS a human there with override abilities.

 

[AccessBlaster]

I was almost a government employee, a few years ago I worked AT the dept of revenue for Arizona, through a contracting firm. When the manager, beaming and smiling, handed me my great treasure and prize - an offer letter at 30% less than I was making contracting, I feigned a weak smile and signed it. Then found a better job a couple months later! Government jobs go from paying less-to-more:
Fed > State > County > City

If I could get in at the City level, that would be ideal. Still get the state retirement system

Sometimes the benefits out way the starting salary.

 

[Isaac]

Sometimes the benefits out way the starting salary.

True, I just couldn't do it at the time, couldn't pay my bills on it

 

[Jon]

I enjoyed this video. I like the animated ones, and this describes how things could evolve over the not too distant future.

 

[The_Doc_Man]

It is time for someone to force-code the AIs to obey Asimov's Three Laws and NEVER remove those barriers to behavior. Also, NEVER allow them to have a truly uninterruptible power supply.

 

[KitaYama]

China has unveiled the world’s first AI-powered hospital, featuring 42 AI doctors and 4 AI nurses, marking a major leap in the future of healthcare.
Developed by Tsinghua University, the “Agent Hospital” simulates full-scale medical operations where AI doctors diagnose, treat, and interact with virtual patients. This cutting-edge facility can handle over 3,000 cases a day, using real medical data to train and test AI in a controlled, research-focused environment.

China’s First AI Hospital: The Future of Healthcare is Here

The Rise of Robotic Doctors and Nurses

medium.com

 

[KitaYama]

France gears up for the future with a colossal AI-powered robot army, set to transform defense by 2040.

France Wants to Roll Out Robot Army by 2040

Though armed conflict is becoming increasingly rare as the decades go by, global military spending is rising sharply with each passing year, reaching unprecedented numbers in the history of warfare. Much of that spending is fueled by the rise of the military tech industry. Between 2021 and 2023...

sg.news.yahoo.com

 

[Isaac]

Seems everyone is "planning" for higher tech in military.
Getting the budget actually approved and the R&D done is a different thing. Hope the good countries are successful in doing so.

 

[The_Doc_Man]

The articles about robo-docs and robo-soldiers make me think back to some movies and TV shows. Here comes a random rumination...

Idiocracy involved a robo-doc who thought its patient was having mental problems - except that "Joe" wasn't crazy - he was just smarter than the average citizen. The robo-doc had trouble diagnosing his real problem - because it was unique. Nobody had ever had that "condition" before - because nobody knew how to think. Which is why "Joe" was a perfect example of the phrase, "In the valley of the blind, the one-eyed man is king." We complain about oligarchs - but is the real problem simply that they see things we don't see and take advantage of that vision?

Iron Man 2 involved robot fighter drones that were pretty formidable and, once you had the control programs written, you had the perfect warrior.

Star Trek (TOS) episode "A Taste of Armageddon" discussed the idea of warfare SO automated that they didn't need to use actual bombs. They just simulated their attacks, tallied up casualties, and sent those casualties to death chambers. It had become so "clean" and "neat" that the warring civilizations thrived because the concept of war had been so "sanitized."

Our writers didn't come up with these concepts out of thin air. They were extrapolating existing conditions. Robotic drones and automated combat systems are becoming commonplace. What I fear is that if you automate war far enough, it becomes too distant, too sanitized, too quiet. It becomes a case of "out of sight, out of mind." War should be shunned. The soldiers returning from Viet Nam suffered PTSD because for them, the war was NOT distant, sanitized, or quiet. War HAS to be a noisy, messy business because otherwise there is no incentive to end it.

Speaking of wars...

I am not saying that Israel is 100% right in what they are doing, but they are apparently of the opinion that the only way to end that war is to maintain the messy business until Hamas has had enough incentive to truly end it. And I don't know if Hamas, being ideologically motivated, will ever see past their face scarves. Scarves that won't stop a bullet but that do (symbolically) stop peace.

The Russia-Ukraine war is already being fought with RC drones and has shown the value of that approach. Drones are cheaper to make in bulk than the cost of training a soldier.

The India-Pakistan conflict looks like IT is heating up as well. At least for the moment, though, they are in a more direct - but contained - shooting match, not using drones that I can tell. Perhaps that "up front and personal" approach will deter a more rapid escalation. One can hope.

 

[Thales750]

The Russia-Ukraine war is already being fought with RC drones and has shown the value of that approach. Drones are cheaper to make in bulk than the cost of training a soldier.

Not to mention, you don't have to go visit their mothers or wives.

 

[The_Doc_Man]

Not to mention, you don't have to go visit their mothers or wives.

A valid point.

 

[Isaac]

That war really makes me sad. I feel for everyone from the Russian kid soldier to the Ukranian kid soldier to the girls in UK villages and the patients in bombed hospitals - just like Gaza. I feel for everyone except Putin, and I'm ambivalent on Zeilensky, I"m not sure when it's time to cede some land to stop the bleeding, but I respect it's a tough decision anyway.

 

[Isaac]

wants the money train to last forever.

Yea

 

[Thales750]

Not me. He is profiting immensely by our throwing money at his war with no accountability. He wants the money train to last forever.

What a surprise,

 

[BlueSpruce]

This whole AI thing is overhyped. It's just a toy for now and too risky to scale. It will fade once the novelty wears off, just like the cloud did. What's the next fad going to be?

 

[GPGeorge]

This whole AI thing is overhyped. It's just a toy for now and too risky to scale. It will fade once the novelty wears off, just like the cloud did. What's the next fad going to be?

I seriously doubt that AI will fade, no more than the internet faded after the dot com bubble burst in the late 1990s. In fact, the cycle is familiar. The "cloud" went through its period of over-hype and settled into a more rational role. Of course, I could be wrong about that. Maybe no one actually uses cloud-based solutions any longer?

AI is coming to the end of the over-hype phase. What will happen next, I believe, is a more rational period in which people begin to understand how to exploit it effectively and appropriately.

Those who figure out how to do so effectively will be more productive over the long run. Reaching for an appropriate analogy, I'd say the time will come when AI will be no more shocking than Cruise Control in automobiles is to the kids learning to drive today. I remember being awestruck by the concept. Today, it's just there.

 

[Isaac]

I agree with bits of both of you - I DO think AI as-is is too risky to scale very well. that's an excellent summary way of saying it.
But it will get through its period of over-hype and settle into something pretty seriously relevant, just as the cloud did.
I do think it will hit more of a 'ceiling' than people think. Regulatory, compliance, legal, IP law, safety, security will all come in to limit its use to only those who know what they are doing, which kinda brings us back to the beginning in a sense.

But it's undoubtedly useful for productivity. I just asked it how to identify any stored procedure in a particular schema containing a particular text. Now I could've figured that out and written it on my own, maybe in 5-20 minutes. But with ChatGPT I did it in 20 seconds.