re: what jobs are “AI proof” and why?

construction

quote:

---

quote:

---

Self-driving vehicles are on the verge of eliminating millions of jobs.

---

Not really.

---

Care to explain? In the very recent past, it might have been a silly thought, but already self-driving vehicles have logged many millions of miles. Computer vision is improving rapidly. NPU (neural processing unit) tech is advancing rapidly.

Computers are better drivers than people. The only question is how *much* better they have to be for society to accept them. We can accept a fatality per 100 million miles with humans driving. I don't know if we will accept that with computers driving.

quote:

---

People would have predicted that white collar jobs and creative jobs like art and music creation were safe from automation, and art and white collar work were put squarely in the crosshairs a short time later.

---

Right, we've been eliminating and automating "blue collar" work for thousands of years. AI does white collar work.

But of course we still have plenty of blue collar jobs after all these centuries, and I suspect the same will be true of white collar jobs in the future. We just don't know what those jobs will shift to. And the transition will be much, much faster this time. I hope we can keep up.

quote:

---

Care to explain? In the very recent past, it might have been a silly thought, but already self-driving vehicles have logged many millions of miles. Computer vision is improving rapidly. NPU (neural processing unit) tech is advancing rapidly.

Computers are better drivers than people. The only question is how *much* better they have to be for society to accept them. We can accept a fatality per 100 million miles with humans driving. I don't know if we will accept that with computers driving.

---

It's not about the capabilities of the computers to drive the vehicles, it's about the way that the trucking and delivery companies operate.
Commercial vehicles break down a lot, and you always need somebody there with a broke down vehicle, or things can and will go south in a hurry.
Drivers don't really make a lot of money, they just put in a lot of hours.
To take that guy out of the seat, is going to mean changing the way an entire industry operates, and it will be cost prohibitive IMHO.

quote:

---

Trial lawyers. Sure AI could "litigate" a case up until the jury trial, but AI is not going to replace the trial lawyer or a jury. The right to a trial by jury is embedded in our Constitution.

---

It could replace many of the lawyers assisting in cases for defense and prosecution. It can do research, write briefs, file things with the court, etc…

quote:

---

My job, internal audit. We operate a lot in grey areas. AI will prolly get to a place where it can do a lot of the testing, but I don't see how it will completely take over those jobs.

---

AI could definitely take the majority of your job. Instead of needing an entire internal audit team, the job could be done by a single person. The AI would be able to make all the black and white fixes and then flag the grey areas for review by the one internal auditor.

Need to differentiate between Artificial Narrow Intelligence(ANI) and Artificial General Intelligence(AGI).

ANI for 200 years has replaced or redefined the human role in the workforce.

ANI will continue to change the work force. However, AGI is more complicated.

I think most jobs currently will be impacted by more advanced ANI but AGI is a long way off from replacing humans in the workforce.

quote:

---

quote:
Care to explain? In the very recent past, it might have been a silly thought, but already self-driving vehicles have logged many millions of miles. Computer vision is improving rapidly. NPU (neural processing unit) tech is advancing rapidly.

Computers are better drivers than people. The only question is how *much* better they have to be for society to accept them. We can accept a fatality per 100 million miles with humans driving. I don't know if we will accept that with computers driving.

It's not about the capabilities of the computers to drive the vehicles, it's about the way that the trucking and delivery companies operate.
Commercial vehicles break down a lot, and you always need somebody there with a broke down vehicle, or things can and will go south in a hurry.
Drivers don't really make a lot of money, they just put in a lot of hours.
To take that guy out of the seat, is going to mean changing the way an entire industry operates, and it will be cost prohibitive IMHO.

---

AI may drive/operate the vehicle, but a human driver will still be in the cab. Kind of pilots are still in the cockpit, we've proven the plans can take off, fly and land themselves without a human in the plan, but a human have to the there in case something happens.

If truck drivers go away, will lot lizards start giving top to the AI?

quote:

---

"Automating many of the jobs that skilled hands can do"? There has been robotic surgery performed in the medical field for a good while now.

---

Yeah I specifically mentioned computer controlled precision equipment.

quote:

---

What job requires more skilled hands than surgery?

---

I said "many", not all. Robots perform lasik, something human hands could never do. They also cut metal and other materials with extreme precision. "Skilled hands" doesn't just refer to precision, it's more about adaptability. And not just in range of physical abilities but also the brain controlling them that adapts to the situation. As you mentioned, human surgeons are still involved and probably will be for the foreseeable future. Too many things can go wrong and every body is different. Decisions have to be made. Ideas must be had. I'm not sure AI will ever have a truly novel idea.

It's not always obvious what a machine can or can't do until we get there. Using the example others posted, it's probably harder to design/build a plumbing robot than a surgical one.

quote:

---

I said "many", not all. Robots perform lasik, something human hands could never do. They also cut metal and other materials with extreme precision. "Skilled hands" doesn't just refer to precision, it's more about adaptability. And not just in range of physical abilities but also the brain controlling them that adapts to the situation. As you mentioned, human surgeons are still involved and probably will be for the foreseeable future. Too many things can go wrong and every body is different. Decisions have to be made. Ideas must be had. I'm not sure AI will ever have a truly novel idea.

It's not always obvious what a machine can or can't do until we get there. Using the example others posted, it's probably harder to design/build a plumbing robot than a surgical one.

---

very well said, specifically the philosophical note on if an AI as we currently understand them can ever generate a truly novel idea. I personally draw a hardline stance on this one as "no, they cannot".

I'm not sure if a thread on this subject would gain any traction at all, but it's at least tangentially related and you may have seen this story or found it interesting. After AlphaGo beat Lee Sedol, I think most people thought the game of Go had joined chess as a game where AI superiority over humans was solidified forever. The ELO rankings certainty suggested no human would ever top the best Go bots like KataGo.

However, with the help up another AI system, humans identified a flaw in KataGo that with a fairly trivial strategy, the bot which is regarded as the best Go player in the world, fails to combat. What is interesting about this is the bot acts as though it does not understand the fundamental objective of the game when this strategy is employed, and a research team got an amateur Go player to go 14-1 against KataGo using this strategy.

LINK

I think this ought to bring up the philosophical point on if these AI systems are capable of true conceptual "understanding" of the subject matter or if they are simply being trained on a superhuman mimicry of sorts.