quote:

Originally posted by .38supersig:
Seems with the removal of entrance exams, maybe AI is the way to go.

They now don't even have to take the MCAT to get in to over 40 medical schools and will probably use AI to cheat on whatever test they do have to take.

Would the thing to do is not trust any doctor under the age of 25? The older more experienced doctors had to learn which end of the scalpel to hold.

Actually, if one goes to HS, then to 4 years of college, then to 4 years of Med school, then into an internship followed by a residency (minimum 3 years), with out any break, one would be 28 or 29 years of age before one starts practice as an independent physician.
As for AI replacing physicians, I can't speak for other specialties, but Neurology will be a difficult one to replace. One needs to learn the Neuro physical exam, which is very specialized and subtle, and also the fine art of taking a detailed Neuro history, which is also a difficult task. Even among today's Neurologists, I read way too many incompetent Neuro H&Ps and notes that missed the boat so to speak...
One of the reasons I chose that field, was that I was pretty certain I would NOT be replaced by a machine during my career (and I was right).

Between a global economy, and AI, and the demographic collapse, I suspect stuff will get "weird".

Plumbers, masons, etc will be difficult to replace - too analog, too many variables.

Nurses will be impossible to replace, in the near term.

An AI "screening" doc, especially if the FDA ever approves some of the rapid lab on a chip tests we were working on 20 years ago, could really lower the burden on docs.

I'm not sure they will be replaced, so much as very augmented. Same with lawyers. Legal Zoom, more or less, tried to replace lawyers, and their product is really, really bad.

(Good for basic stuff, like being an agent for process, but the work product I've encountered from them is awful.)

Again, it might help lower the workload, though.

quote:

Originally posted by Bytes:

quote:

Originally posted by Skins2881:
Is now too soon?

IBM plans to replace 7,800 jobs with AI over time, pauses hiring certain positions

Probably too soon. IBM is top heavy as hell. They've given up all their hardware money making divisions and are relying on software support. Don't get me wrong, AI could replace a bunch of positions everywhere. AI could probably replace IBM's CEO and the company wouldn't skip a beat. AI replacing jobs is not the issue.

I'm guessing SoftLayer makes up for their hardware divisions. Cloud is a huge component of their business now. Hardware is cheaper to make now and not as profitable.

They are replacing people with AI using attrition, not layoffs. So their moves won't be all that noticable, but multiply it by all Fortune 500 companies.

quote:

Originally posted by WaterburyBob:
We haven't needed too many buggy whips for quite a while.

When making buggy whips was a profession, there weren't eight billion people in the world. This AI crap is going to be the downfall of a lot of formerly successful people, who will have to turn to- what? Manual labor. Service industry. If most of the tech jobs disappear, well...

AI will benefit corporations and governments and it's going to fuck everyone else and it ain't gonna be pretty. This is the kind of stuff that could lead to worldwide riots in fairly short order.

I always appreciated sci-fi as a genre. Espically DUNE. They did not allow for AI to overtake them a second time. That author saw the future very clearly

This has all happened before…

quote:

How soon will AI start replacing human jobs?

Given the way that every single item on my take-out order this evening was wrong, AI can't come soon enough in the fast-food industry.

A lot of businesses would like to replace employees with AI, but the reality is that won’t happen for most jobs.

I have some experience with AI. Neural nets AKA “Machine learning” is very cool tech, and very effective at what it has been trained upon. But if you present it with a case outside it’s training data, what we used to call “exemptions”, it will choose the closest case, and it will be wrong. Think of it as an idiot savant. It can’t “think outside of the box”. It has no common sense. And to train it, you need a huge volume of data from a trusted source, because you can’t un-train it from bad data.

It can aid people to make decisions, and is blazingly fast. But it isn’t AI like in Terminator or Wargames, it is more like the original Westworld movie.

I've always said, if they could replace me with a toaster and save $0.40, I'd be gone by morning. This current round of AI reveals is impressive, and for sure, there will be many applications in the near future, but as for fully replacing staff, I think we're far enough out that much of any transition will be in a generational rotation. With the possible exception of authors and programmers. Those who make language. Since it's the language models that have quantum leaped.

Basically, if you consume digital, process and output digital, I think that will move quickly. My son is in school for cybersec. I'm hoping he'll land a job steering the AI. But, maybe I should have let him go to art school.

Oh I’m SURE airline executives would LOVE to replace me in the cockpit with AI. No more fatigue calls. No more haggling over contract negotiations. No more required crew rest…departures 24/7. Millions in savings on hotel expense.

Problem is, what happens when something really non-normal arises? Someone posted above that AI isn’t very good at thinking outside the box. Well…that’s EXACTLY what the “Sully moment” was on the Hudson.

Nah…I’ll start taking the train for travel the day that happens. That’ll be AI, too, but at least I’ll already be on the ground when the train crashes…

^^^ Could trim a flight officer or engineer out of each route though possibly?

quote:

Originally posted by Nuclear:
A lot of businesses would like to replace employees with AI, but the reality is that won’t happen for most jobs.

I have some experience with AI. Neural nets AKA “Machine learning” is very cool tech, and very effective at what it has been trained upon. But if you present it with a case outside it’s training data, what we used to call “exemptions”, it will choose the closest case, and it will be wrong. Think of it as an idiot savant. It can’t “think outside of the box”. It has no common sense. And to train it, you need a huge volume of data from a trusted source, because you can’t un-train it from bad data.

It can aid people to make decisions, and is blazingly fast. But it isn’t AI like in Terminator or Wargames, it is more like the original Westworld movie.

This. This is not skynet people. You train it in simple to complex logic. I think the danger is in who trains it & what bias they can train it with potentially.

All you haters! What about cyborg rights? Or rights for those who were born human but "identify as cyber?" What about human-cyborg marriages? Court-ordered software "upgrades" for "cyber offenders?"

"Natural rights for artificial beings," I expect to hear that mantra in my remaining lifetime.

"Who" will be the first bot to get a CCW? More to the point: when will the first "mass shooting" by an AI-enabled bot occur? Which will come first, open season on bots, or "bot rights?" Is peaceful co-existance even close to possible? Should we start a pool?

As far as Erica goes, the video would be much more convincing if they had "her" performing a pole dance.

Mostly kidding here, the prolific use of quotes in this post indicate that we don't even have the language yet to deal with this phenomenon, much less understand the eventual outcomes. The buggy whip example is tired and wrong, most of the jobs that replaced those of the buggy whip industry were not even imagined of at the time, and I would challenge those who'd claim that we are not better off than we were back then (perhaps with a few exceptions in certain aspects).

quote:

Originally posted by WaterburyBob:
I expect my job, application software design and development, to be done by AI in 3-5 years.
It won't affect me as I'm semi-retired, but it's going to put a lot of people that make decent money today out of work.

But a computer can't dream up a new app out of the blue? Musk - "What we need is a new way to pay for things on the internet . . . "

quote:

Originally posted by 1flynDO:
I always appreciated sci-fi as a genre. Espically DUNE. They did not allow for AI to overtake them a second time. That author saw the future very clearly

Thou shalt not make a machine in the likeness of a human mind.

Machines will do the work of people.
People will take care of the Machines.
Machines will be built to take care of the machines.
Those people will be reduced to those who take care of those machines.
...and on and on until we need no people.

AI will initially be a demonstration of our technical abilities.
Then some minor productivity.
Quickly it will be commandeered to use for an advantage - political, power/control, financial.
Some AI will be used to defend above.
Then AI will figure out how to defeat the defense.

No more need for people.

Also a good plot for future movies.

We are already seeing the built it bios to the liberal cause in its answers.

quote:

Originally posted by jhe888:
But a computer can't dream up a new app out of the blue. Musk - "What we need is a new way to pay for things on the internet . . . "

Not now, but we are getting very, very, close to doing that. The new apps will probably be crude as hell but gen 2, 3, 4, ...

quote:

Originally posted by erj_pilot:

at least I’ll already be on the ground when the train crashes

You'll probably be on the ground if the airplane crashes.

quote:

Originally posted by jhe888:

quote:

Originally posted by WaterburyBob:
I expect my job, application software design and development, to be done by AI in 3-5 years.
It won't affect me as I'm semi-retired, but it's going to put a lot of people that make decent money today out of work.

But a computer can't dream up a new app out of the blue? Musk - "What we need is a new way to pay for things on the internet . . . "

As soon as a user is able to just ask AI for what it wants, the code will be generated for the App.
I don't think that's too far in the future.

The code might not be the most efficient at first, but it will improve.

quote:

Originally posted by WaterburyBob:
As soon as a user is able to just ask AI for what it wants, the code will be generated for the App.
I don't think that's too far in the future.

The code might not be the most efficient at first, but it will improve.

Look at the history of coding. Even since I first learned how to code it has gone from machine/assembly languages to modern structured languages that are practically "tell the computer what you want, and it will generate the code." AI is more "we'll tell you what you want, there's no need for any 'code.'"

Put your faith in the hackers. The "blackhats" of today are the anti-authoritarian heroes of tomorrow! As long as they "win" anyway. Where is John Conner? Is he even among us?