Log In

Your email is not your username

Register

If you were a member of the old Bodybuilding.com forums and would like to reuse your previous username, you can request it below. We use your email only for registration and do not store it. For more information, please see our Privacy Policy.

Confirm your email

A registration code was sent to your email. Enter it here.

Welcome

You have successfully setup your account.

Sign in

Quick Navigation Bottom Misc
Forum » Practicing medicine now, I see just how immune doctors are from AI/automation (srs)
  1. Results 361 to 390 of 470
  2. Page 13 of 16
  3. First
  5. 11
  6. 12
  7. 13
  8. 14
  9. 15
  11. Last

Thread: Practicing medicine now, I see just how immune doctors are from AI/automation (srs)

post 1596712981 01-30-2020, 12:35 PM
-
#361
  1. propreffered7
  2. Registered User
  1. propreffered7
  2. Registered User
  3. Join Date: Mar 2010
  4. Posts: 36,839
  5. Rep Power: 86393
Originally Posted By Scarrdbutsmartr⏩
His thought pattern is factual and aligns with how technology grows and adapts. There is actually a law that demonstrates his his position. The law is specific to hardware but its a generally accepted pattern within technology. Moores law...
Agreed. But its a very easy stance to take. So essentially, he says that. Then people point out SPECIFIC problems and he just reverts back to that wide angled stance. As i said, ive posted hundreds of problems with OUR OWN brain that we dont know about.

So to suggest that we could then replicate that is outlandish.

The intent of "artificial intelligence" begins with the idea of the human brain as a model seeing as that "intelligence" is all we know, correct?

So first and foremost before we begin to implement superior changes, we at least need to replicate that before we can worry about making different changes correct?
post 1596713151 01-30-2020, 12:36 PM
-
#362
  1. BloodFireDeath
  2. Registered User
  1. BloodFireDeath
  2. Registered User
  3. Join Date: Aug 2011
  4. Posts: 5,096
  5. Rep Power: 20639
Originally Posted By AltarOfPlagues⏩
this isnt always the case. sometimes a lot of people die before people start washing their hands after handling dead bodies.

its because they dont know how the tool works and they dont know who the patient is gonna sue in case the tool is hypothetically bad.
Using examples from 2 centuries ago. Lol who are you mad at? Just the doctor community in general? Is that why you think its dogma at work here that is stopping AI from taking over?
Originally Posted By MBisonSon⏩
Talk to someone with a legit smart home.

Is thecomputer in your car also a useless gimmick?

Applications will be refined, new ones will be discovered, etc.

Kinda like how first video games sucked and now we're on the verge of VR.

I bet 30 years ago there were plenty of people saying computers would never catch on.
Pretty much. If the world ended tomorrow and everything got wiped out and then respawned, nurses/doctors would be needed from day one. Same as farmers, builders, soldiers. Nobody gaf about self driving cars or even computers aka non-essential, they're meant to support/make easier the rest of the jobs to sustain humanity. We're so used to them now that we think we really NEED them but they aren't even remotely essential for life tbh.
post 1596713331 01-30-2020, 12:38 PM
-
#363
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By propreffered7⏩
The intent of "artificial intelligence" begins with the idea of the human brain as a model seeing as that "intelligence" is all we know, correct?
incorrect. increasingly people in the deep learning community wrestle with nomenclature. ive recently seen intelligence defined as "organized information" or some such.

at best, there is a spectrum of intelligence. what youre talking would be described as human-level AGI.
Originally Posted By BloodFireDeath⏩
Pretty much. If the world ended tomorrow and everything got wiped out and then respawned, nurses/doctors would be needed from day one. Same as farmers, builders, soldiers. Nobody gaf about self driving cars or even computers aka non-essential, they're meant to support/make easier the rest of the jobs to sustain humanity. We're so used to them now that we think we really NEED them but they aren't even remotely essential for life tbh.
you think computers are a gimmick so you are practically a flat earther bro.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596713491 01-30-2020, 12:41 PM
-
#364
  1. propreffered7
  2. Registered User
  1. propreffered7
  2. Registered User
  3. Join Date: Mar 2010
  4. Posts: 36,839
  5. Rep Power: 86393
Originally Posted By AltarOfPlagues⏩
incorrect. increasingly people in the deep learning community wrestle with nomenclature. ive recently seen intelligence defined as "organized information" or some such.

at best, there is a spectrum of intelligence. what youre talking would be described as human-level AGI.
Nah, ive posted so many problems in here. Im not going through them again

How do you know they arent shapeshifting their idea of "intelligence" around their limitations?

Given the problems ive proposed its very clear that could be the case
post 1596713541 01-30-2020, 12:41 PM
-
#365
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By propreffered7⏩
How do you know they arent shapeshifting their idea of intelligence around their limitations.
thats exactly what they are doing.

the entire point of tt is you dont need a human brain to beat a human brain at specific tasks. those tasks are increasing in complexity with no end it sight.

edit self driving cars is a watershed moment because for the first time, machines will outperform humans in calculation and decision making in an extremely complex euclidean space with many degrees of freedom.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596714171 01-30-2020, 12:49 PM
-
#366
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By propreffered7⏩
Agreed. But its a very easy stance to take. So essentially, he says that. Then people point out SPECIFIC problems and he just reverts back to that wide angled stance. As i said, ive posted hundreds of problems with OUR OWN brain that we dont know about.

So to suggest that we could then replicate that is outlandish.

The intent of "artificial intelligence" begins with the idea of the human brain as a model seeing as that "intelligence" is all we know, correct?

So first and foremost before we begin to implement superior changes, we at least need to replicate that before we can worry about making different changes correct?
I think you may have an incorrect vision of artificial intelligence...Artificial intelligence is not meant to replicate a human brain. It is meant, at this time, to replicate a single function of the human brain which is decision making. When you look at the brains ability to make decisions it is not only limited but its highly compromised.

How does a human brain makes decisions - It takes in input, it checks its memory store for past experiences and learned knowledge and then makes a decision. The decision is heavily influenced by another sector of the brain which stores emotion. This influences decisions and absolutely causes bad decisions.
How does AI work - It gathers hundreds of billions/trillions of data sources across billions of platforms. It identifies trends, isolates and learns patters and then suggest patterns with automation...It can be completely automated if you let it off of the chain (this is what worries people)

Why is this important...It would take a human mind millions of years to do a task that AI can. Why - because a human mind only has so much potential to experience and gather/store data. AI will always be superior because it can always review an unlimited data store and a human cant. It can make better and faster decisions because its literally looking at data analysis from the brain power generated by millions of people.

This is the first phase of AI. We can talk about future phases but this phase ^^ is what puts people out of work. The second phases are science fiction based and I wont go into them because we have to get through the first phase...and make no mistake we are doing that today....and the first phase will put doctors/lawyers out of work

I can give you a real world example of AI and its applicability that is working today. There is no voodoo to it. Its ones and zeros and can be explained how anyone can understand it. A real world example takes it from the ethereal to something you can wrap your head around. A real world example can show you how a human mind can never compete..Let me know if you want the example.
post 1596714251 01-30-2020, 12:51 PM
-
#367
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
that any of this is remotely provocative is fukkin nuts. ive never had a lower opinion of doctors, and ive had an obese sports medicine doctor advise me to never deadlift again.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596714361 01-30-2020, 12:52 PM
-
#368
  1. FAPhaggot
  2. Hawaiian shirt hoarder
  1. FAPhaggot
  2. Hawaiian shirt hoarder
  3. Join Date: Nov 2015
  4. Posts: 24,226
  5. Rep Power: 185354
I've spent the last 2 years doing research in electrical engineering, and the last year doing biomed and working on cognitive neuroscience, so I know a few things about this thread. A couple different thoughts I have:

1.Youreallydon't want to build a computer modeled after the brain. It's a fukking crazy place.

The biggest problem is how tightly coupled and interconnected everything is. For example, the brain area I'm exploring now has 3 different connections giving feedback into it, and we don't really know how 2 of them work. And because we don't understand how the feedback works, we don't know how that cortical area really works either - how much is it doing on its own, and how much is coming from other brain areas through those pathways? Picture a transfer system with an input, output, and 4 interconnected boxes - we basically have question marks in 3 of those boxes right now.

So when it comes to how the brain actually works on a low level, we don't know chit. We know how a lot of the high-level input/output works, if you will, but the nitty-gritty in between that controls it is by and large a mystery. IMO, if you asked me how long it will take before wereallyunderstand the brain, I would guess somewhere between 50 years and never.

If you want to take the neural-net idea that's like the vaguest high-level idea of how a brain functions, then sure. If you try and replicate it neuron by neuron, you are 1000% retarded.


2.The chit that AI is doing right now is actually kind of scary. BRB "just take people-facing jobs, robots can't do people chit." I can't say too much here b/c NDAs (someone actuallybroke NDA to show me this), but I'll just say that if I was an actor right now, I would be very worried about my future career. I've seen PhD engineers working in R&D whose jobs are probably going to start getting replaced in 10 years. I don't think there's any job that is completely safe long term, no matter how much education or how much "personal touch" you think you need. It's all on the table.


3.Surgical robots aren't going to be here for a while, because picking up and moving things with your hands is actually a REALLY hard job to do with a machine. You know how you can pick up a wine glass with 3 or 4 fingers, then put it down and start curling a dumbbell without even thinking? Yeah. There's not a robot arm I'm aware of on the planet that can do that. Some startups especially like to have carefully staged videos that shows cool chit going on, but when you put them out "in the wild", they really struggle (see below).

That being said, there is a T O N of money going into robot-arm chit. Amazon holds a competition every year to see if a robot arm can pick a bunch of random goods out of a box and put them on its conveyor belt warehouse packing line. IIRC, the highest rate any arm has succeeded at yet is 50%... but as soon as that picking can be automated, they're going to fire every warehouse employee they have, buy thousands and thousands of robot arms, and make someone VERY, VERY rich. And that's if Amazon's internal R&D doesn't get there first.

It's going to take a while, but IMO, it's just a matter of when it happens, not if. And once we have an automated arm with that kind of tactile control, the clock is going to start running for surgeons immediately. It's a job that's safe for 10 years, but pry not 20.
FA Crew
Always Pick 1 Crew

"Experience is something you get right after you need it."
post 1596714521 01-30-2020, 12:54 PM
-
#369
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By FA*******⏩
I2. The chit that AI is doing right now is actually kind of scary. BRB "just take people-facing jobs, robots can't do people chit." I can't say too much here b/c NDAs (someone actuallybroke NDA to show me this), but I'll just say that if I was an actor right now, I would be very worried about my future career. I've seen PhD engineers working in R&D whose jobs are probably going to start getting replaced in 10 years. I don't think there's any job that is completely safe long term, no matter how much education or how much "personal touch" you think you need. It's all on the table.
.
one of the craziest notions is that the AI also cant really tell you what it's doing. i read it described as like asking a pro baseball player how hes able to hit. he doesnt fully understand it himself, the know how is emergent and not fully articulated
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596715301 01-30-2020, 01:06 PM
-
#370
  1. MyGhettoFantasy
  2. Registered User
  1. MyGhettoFantasy
  2. Registered User
  3. Join Date: Mar 2016
  4. Age: 42
  5. Posts: 12,838
  6. Rep Power: 162504
Didn't read all pages but I have the impression that a lot of people itt need a better doctor.

Diagnosis is a small (albeit critical) part of effective medicine. The ability to positively affect patient compliance is, in my opinion, the difference between a good doctor and a greater doctor assuming all else is equal.

Most people are going to die from some sort of chronic condition that will gradually kill them or induce some acute episode that kills them. Often times these conditions can be prevented or managed by a competent , personable doctor that is able to convince a patient to implement a treatment plan, encourage their morale, and foster an honest relationship wherein a patient offers truthful answers and vital information.

Just lol at putting medicine down to computational power. I don't know how strong of an argument can be made that AI imitating human affect will be able to convince an incredulous patient to work with the prescription.
The closer we approach the uncertainty of life's ending the more we wish to trade all of the things we have acquired in exchange for all of the things we have lost: wealth for youth, knowledge for fresh curiosity, resignation for hope. We'd trade our wisdom for new experiences, but it is wisdom that will teach us that at the end of the road the only new experience is death.
post 1596715761 01-30-2020, 01:11 PM
-
#371
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By MyGhettoFantasy⏩
Just lol at putting medicine down to computational power. I don't know how strong of an argument can be made that AI imitating human affect will be able to convince an incredulous patient to work with the prescription.
so first, i dont know how strong of an argument can be made that an actual human can convince people to stop eating little debbies.

secondly, no one has hinted that a medical algorithm should be in the "convincing incredulous patients" line of work.

there is a strong "android" concept of machine learning itt.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596716041 01-30-2020, 01:14 PM
-
#372
  1. BulkingIsHard
  2. Registered Bigot
  1. BulkingIsHard
  2. Registered Bigot
  3. Join Date: Aug 2016
  4. Posts: 12,670
  5. Rep Power: 70015
Originally Posted By FA*******⏩
Wall if text.
AI is actually pretty slow at learning things. Humans are far more adaptable and quicker learning. Plus we utilize previous knowledge and experiences that make us better at other unrelated tasks.

For example, a human can learn linguistics and build off that into literacy. Then go to the gym and learn how to do a clean and jerk from another person or a book. An AI can only be good at one of those, and it needs to train on thousands on samples/scenarios for thousands of times before it can perfect it.

Of course once it reaches the state of convergence, it will consistently perform better than a human every time. If the data is already available, this makes training quite easy... only a few days in realtime or less and the AI is an expert. But through reinforcement learning without simulation training, expect it to struggle for a long time.

AI will continue to progress but the next leap forwards for AI regards general intelligence.
https://en.wikipedia.org/wiki/American_decline
https://en.wikipedia.org/wiki/Societal_collapse#By_absorption
post 1596716771 01-30-2020, 01:23 PM
-
#373
  1. MyGhettoFantasy
  2. Registered User
  1. MyGhettoFantasy
  2. Registered User
  3. Join Date: Mar 2016
  4. Age: 42
  5. Posts: 12,838
  6. Rep Power: 162504
Originally Posted By AltarOfPlagues⏩
so first, i dont know how strong of an argument can be made that an actual human can convince people to stop eating little debbies.

secondly, no one has hinted that a medical algorithm should be in the "convincing incredulous patients" line of work.

there is a strong "android" concept of machine learning itt.
A sincere, concerned doctor can influence patient compliance. Of course this is variable.

People are talking about doctors being replaced. The ability to convince an incredulous patient to fully and truthfully participate in a treatment plan is implied when you're talking about replacing a doctor (and everything that they do/ are capable of doing).

Perhaps the implications are lost on some.
The closer we approach the uncertainty of life's ending the more we wish to trade all of the things we have acquired in exchange for all of the things we have lost: wealth for youth, knowledge for fresh curiosity, resignation for hope. We'd trade our wisdom for new experiences, but it is wisdom that will teach us that at the end of the road the only new experience is death.
post 1596716811 01-30-2020, 01:24 PM
-
#374
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By BulkingIsHard⏩
AI is actually pretty slow at learning things. Humans are far more adaptable and quicker learning. Plus we utilize previous knowledge and experiences that make us better at other unrelated tasks.

For example, a human can learn linguistics and build off that into literacy. Then go to the gym and learn how to do a clean and jerk from another person or a book. An AI can only be good at one of those, and it needs to train on thousands on samples/scenarios for thousands of times before it can perfect it.

Of course once it reaches the state of convergence, it will consistently perform better than a human every time. If the data is already available, this makes training quite easy... only a few days in realtime or less and the AI is an expert. But through reinforcement learning without simulation training, expect it to struggle for a long time.

AI will continue to progress but the next leap forwards for AI regards general intelligence.
You make a strong point but the point you are making isnt really fair....It is true that AI can require a few weeks to a few months to learn and make decisions. The fallibility in your statement is that it took your brain years to be able to adapt and respond quickly in the instances you proposed. Your brain, as an infant, didnt instantly start make good decisions or adapting without challenges - but that is what you are expecting of AI.

When you want to compare a fully learned AI instance to a human brain, AI almost always wins if the data store is large enough. When you include the capabilities of Quantum computing its game over..
post 1596716921 01-30-2020, 01:25 PM
-
#375
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By MyGhettoFantasy⏩
A sincere, concerned doctor can influence patient compliance. Of course this is variable.

People are talking about doctors being replaced. The ability to convince an incredulous patient to fully and truthfully participate in a treatment plan is implied when you're talking about replacing a doctor (and everything that they do/ are capable of doing).

Perhaps the implications are lost on some.
we are a long way away from human level AGI androids, thats correct. thanks for your posts.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596717541 01-30-2020, 01:31 PM
-
#376
  1. MyGhettoFantasy
  2. Registered User
  1. MyGhettoFantasy
  2. Registered User
  3. Join Date: Mar 2016
  4. Age: 42
  5. Posts: 12,838
  6. Rep Power: 162504
Originally Posted By AltarOfPlagues⏩
we are a long way away from human level AGI androids, thats correct. thanks for your posts.
Yes. A long way away from this thread being relevant. Entertaining read, though.
The closer we approach the uncertainty of life's ending the more we wish to trade all of the things we have acquired in exchange for all of the things we have lost: wealth for youth, knowledge for fresh curiosity, resignation for hope. We'd trade our wisdom for new experiences, but it is wisdom that will teach us that at the end of the road the only new experience is death.
post 1596717841 01-30-2020, 01:34 PM
-
#377
  1. BOZZ
  2. Banned
  1. BOZZ
  2. Banned
  3. Join Date: Oct 2001
  4. Posts: 4,396
  5. Rep Power: 0
Originally Posted By Scarrdbutsmartr⏩
You make a strong point but the point you are making isnt really fair....It is true that AI can require a few weeks to a few months to learn and make decisions. The infallibility in your statement is that it took your brain years to be able to adapt and respond quickly in the instances you proposed. Your brain, as an infant, didnt instantly start make good decisions or adapting without challenges but that is what you are expecting of AI.

When you want to compare a fully learned AI instance to a human brain, I always wins. If the data store is large enough AI always wins. When you include the capabilities of Quantum computing its game over..
So what stops the AI from developing a bias? Isn't this a known issue with AI? Like when Google photos turned up pictures of people when you searched for gorillas? I did work with a company using facial recognition and it was terrible at identifying non white people.

Just one example but how is this handled? Or is it just accepted that in certain scenarios it's useless?
post 1596717911 01-30-2020, 01:34 PM
-
#378
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By MyGhettoFantasy⏩
Yes. A long way away from this thread being relevant. Entertaining read, though.
its relevant to everyone who isnt constrained by grammatical literalism.

edit- reppin on spread
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596718101 01-30-2020, 01:37 PM
-
#379
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By MyGhettoFantasy⏩
A sincere, concerned doctor can influence patient compliance. Of course this is variable.

People are talking about doctors being replaced. The ability to convince an incredulous patient to fully and truthfully participate in a treatment plan is implied when you're talking about replacing a doctor (and everything that they do/ are capable of doing).

Perhaps the implications are lost on some.
This assumes business decisions are driven by the consumer...Decisions are almost always driven by cost.

Example

Option A - You will have to pay insurance premiums of $600 a month, When you are sick you will have to visit the doctor, pay a copay and there will be an error rate of 45% (this can lead to multiple visits, numerous tests/life expentency). Numerous tests will be scheduled that may or may not identify the problem and you will pay a portion of all of the tests. The time to get well will typically be 4 weeks.

Option B - You will have to pay an insurance premium of $50 a month - You will have the option to use the latest app on your smart phone to fill in your symptoms. You will use a home based analysis station (These exist today). AI will conduct an analysis based on your survey and home based analysis, it will be run through centralized AI. The AI system will identify your health challenges based upon trillions of other data points. The accuracy for this diagnosis will be near 98%. A centralized doctor will review the AI evidence, contact you via a smart phone video app and answer any questions. The two test required to validate your medical issue have been automatically schedule. These two test will have a 99% viability rate for validating AIs diagnosis. The typical time to identify cause and get well is typically 3 days.


This ^^^ is whats coming. It isnt fantasy, its case study. Insurance companies/Governments are already looking at this technology and model. The governments of the world that manage healthcare are going to jump on this...It will work and then the rest of the private healthcare entities will fall like dominoes...This WILL happen.
post 1596718801 01-30-2020, 01:44 PM
-
#380
  1. FAPhaggot
  2. Hawaiian shirt hoarder
  1. FAPhaggot
  2. Hawaiian shirt hoarder
  3. Join Date: Nov 2015
  4. Posts: 24,226
  5. Rep Power: 185354
Originally Posted By BulkingIsHard⏩
AI is actually pretty slow at learning things. Humans are far more adaptable and quicker learning. Plus we utilize previous knowledge and experiences that make us better at other unrelated tasks.

For example, a human can learn linguistics and build off that into literacy. Then go to the gym and learn how to do a clean and jerk from another person or a book. An AI can only be good at one of those, and it needs to train on thousands on samples/scenarios for thousands of times before it can perfect it.

Of course once it reaches the state of convergence, it will consistently perform better than a human every time. If the data is already available, this makes training quite easy... only a few days in realtime or less and the AI is an expert. But through reinforcement learning without simulation training, expect it to struggle for a long time.

AI will continue to progress but the next leap forwards for AI regards general intelligence.
Heh, I actually do research in motor learning. We learn a lot more like an AI than you might think.

Got to leave soon, so I'll hit you with the cliffs:



-Your brain builds models of what it expects from certain sensations or actions.

-When your brain tries to throw a ball, say, or touch a barbell, it has internal predictions built on years of experience. "This is where I expect the ball to go if I angle my arm like this. This is what I expect a steel bar to feel like."

-When your brain is introduced to something new, it takes its estimated most-similar model, compares it to the IRL sensation, and goes "oh chit, I'm off by this much."

-That error is used to correct and adjust your mental model so that's it's more accurate in the future.

-Over time, the expectation and the reality converges and there's no further updating. Once you've grabbed 100 barbells, you know what the 101st is going to feel like pretty accurately.



If that sounds like AI reinforcement learning... it kinda should. Humans aren't faster at learning per se, we've just been grabbing and throwing random chit since we we were 2 years old.

Quickest picture I found on Google... I'll update this later if I find something better.

FA Crew
Always Pick 1 Crew

"Experience is something you get right after you need it."
post 1596719081 01-30-2020, 01:48 PM
-
#381
  1. propreffered7
  2. Registered User
  1. propreffered7
  2. Registered User
  3. Join Date: Mar 2010
  4. Posts: 36,839
  5. Rep Power: 86393
Originally Posted By Scarrdbutsmartr⏩
When you include the capabilities of Quantum computing its game over..
No, no its not. Just because a quantum computer can do exponentially more computations does not mean it replicates the intricacies of the brain. As has been stated many times itt, ive posted lots of problems that, regardless of a computers computational power has no answer for. This is just a few of them. And there are hundreds if not thousands. I see a lot of these proponents then acquiescing and saying "well, they arent really intelligent" but they can do the things we specifically designed them to do. And or then altering the definition of intelligence. Inb4 "we will" with no grounding whatsoever.

here are some issues:
6a. The Subjectivity Epistemological Problem. This is the problem associated with “subjective” versus “objective” perspectives on being in the world, and is highlighted by folks like Andre Marquis' quadrant analysis. A way to think about this is that subjective experiential consciousness is fully "contained" within the individual. This containment results in two important sub problems, which are mirror images of each other. The first (6a1) is the problem of directly knowing another’s subjective experience—the problem being it cannot be done. This is the problem of: “How do I know that you see red the way I see red?” This problem also relates to our knowledge of consciousness in other animals, which we can only know indirectly. This is also related to the philosophical problem of zombies. Indeed, all subjective experiences can only be inferred via behavior from an objective perspective. The second (6a2) issue is the inversion of this problem. This is the problem that, as individuals, we are trapped in our subjective perceptual experience of the world. That is, the only way I can know about the world is through my subjective theater of experience. We can see the nature of this problem when we ask, as Rene Descartes did a long time ago, How do I know the external world is real and that I am not living in a dream arranged by some evil demon? This problem was the basis of the popular movie series The Matrix.

6b. The Neuro-Engineering Problem. This refers to the specific mechanics of how the activity of the brain gives rise to felt experiences of reds, and pleasures and pains. We can also break this up into two different problems, the neuro-correlation problem and the neuro-causation problem. The Neuro-Correlation problem (6b1) refers to how we map the correlations between felt experience and brain activity. Consider, for example, we have long been aware that damage to the occipital lobe impairs vision, whereas damage to the auditory lobe impairs hearing. Even more directly, scientists have homed in on the kinds of brainwave activity that directly corresponds to conscious access. For example, in an exciting line of research on an idea called “global neuronal workspace” theory, Dehaene and his colleagues have identified a P3 “ignition wave” that occurs approximately 300 milliseconds after sensory input that is associated with consciously experiencing a visual stimulus. We are getting better and better at mapping brain correlates of conscious experience and that is leading to clearer frameworks for understanding the subjective conscious mind. The Neuro-Causation problem (6b2) is the question of why and how certain brain processes produce the subjective experiences that they do. Specifically, what causes the experience of redness versus greenness? Why does the architecture of the occipital lobe produce the experience of vision, whereas the architecture of the auditory lobe produces the experience of sound? Similarly, what are the minimal brain processes that generate conscious experiences to begin with and why and how do they do that? This is often referred to as the “hard” problem of consciousness. It is also sometimes called and/or relates to the neuro-binding problem. Currently, the question of “why” specific brain states result in specific experiential states is pretty mysterious.

7. The Self-Consciousness (SC) Problem. This refers to the nature of self-reflective consciousness, which, in its full form, involves self-reflective narration. In the beginning of Horgan’s book, he speaks of a time when he went fishing as a kid with his friends when he all of a sudden had a profound moment of self-reflective awareness and announced, with newfound awe, “I am me!” to his friends. Although his friends were not impressed, Horgan was deeply impacted. What changed in Horgan at that moment was not so much his experiential consciousness (i.e., his sense of the fishing rod in his hand), but of self-reflective awareness of himself as an independent entity. This domain of consciousness involves questions such as: What is self-consciousness, when and why did it evolve? What is human consciousness relative to animal consciousness? How does self-consciousness develop? How does it relate to language and narration?

One key feature of self-conscious narration that is very different from experiential consciousness is that it is not contained inside the nervous system. In contrast to the slight pain in my finger, which I can indirectly report on because it is contained in my subjectivity, I am directly sharing with you my thoughts about the 10 problems of consciousness. Language allows for an explicit intersubjective sharing. That is one key difference between experiential and self-consciousness, and there are several. Four sub-problems associated with self-consciousness are: the problem of the self (SP); the problem of free will (FWP), the problem of personhood (PP); and the strange loop problem (SLP).

7a. The problem of the self. Self-consciousness brings in a new word, self, into the mix. What is the self, what is the sense of the self? Is there a self at the experiential consciousness level? What is the relationship between the self and consciousness at both levels? What is the relationship between related concepts like self-concept, identity, self-esteem? How does the self-consciousness self-system develop over childhood? What are we to make of the Buddha's insight that there is no self? How does that contrast with earlier notions of Atman or other language games that emphasize the central role of the self in human psychology?

7b. The problem of free will. Self-awareness means that I am aware of me and of what I am doing and what I might or should be doing. These are the key ingredients for the concept of free will, which is the idea there is a self which is choosing to act. I experience myself as deciding to write this blog. But what does that mean? Does self-awareness really give rise to the capacity to self-consciously choose one’s actions? Are (or can) those actions be chosen freely? How is that possible if we live in a universe determined by physical processes?

7c. The problem of personhood. Normally, we say a person is a human being. But in the Behavior of Persons, Peter Ossorio makes the compelling argument that this is not quite right. According to Ossorio, a person is an entity who is self-aware and takes reflective responsibility for their actions, whereas a human being is a particular kind of great ape. To make his point that the concept of personhood is different than the concept of a human being, Ossorio points out that there are many science fiction characters (think of beings in Star Wars like Yoda or Jabba the Hut) who clearly are persons in the conceptual sense, but are not human beings. Ossorio's analysis also highlights the intimate connection between language and personhood, such that full fledged exemplars of persons behaving as such involve narrative awareness and a sense of ownership for one's actions.

7d. The problem of the strange loop of self-awareness. Something odd happened to Horgan as he repeated “I am me.” As he narrated that, his conception of his situation changed. Thus, he found himself in a “strange loop” where the cause and effect of his narrative behavior was all tangled up. The cognitive scientist/philosopher Douglas Hofstader has explored the dynamics of being a strange loop. I believe that the strange loop can also be applied to broader issues, like the problem of the double hermeneutic and Sartre’s analyses exploring the nature of human freedom and the fusion of facts and values. The bottom line is that it does carry complicated implications for things like the relationship between causes and effects and describing/explaining and prescribing/influencing things.
Btw. Why do you suggest this is good? Legit want to hear the reasons. If you think ai has this capability cant the autonomous decision maker turn around and use it for evil ?
post 1596719151 01-30-2020, 01:48 PM
-
#382
  1. MyGhettoFantasy
  2. Registered User
  1. MyGhettoFantasy
  2. Registered User
  3. Join Date: Mar 2016
  4. Age: 42
  5. Posts: 12,838
  6. Rep Power: 162504
Originally Posted By Scarrdbutsmartr⏩
This assumes business decisions are driven by the consumer...Decisions are almost always driven by cost.

Example

Option A - You will have to pay insurance premiums of $600 a month, When you are sick you will have to visit the doctor, pay a copay and there will be an error rate of 45% (this can lead to multiple visits, numerous tests/life expentency). Numerous tests will be scheduled that may or may not identify the problem and you will pay a portion of all of the tests. The time to get well will typically be 4 weeks.

Option B - You will have to pay an insurance premium of $50 a month - You will have the option to use the latest app on your smart phone to fill in your symptoms. You willuse a home based analysis station. AI will conduct an analysis, it will be run through centralized AI. The AI system will identify your health challenged based upon trillions of other data points. This is 98% accurate. A centralized doctor will review the AI evidence with you and answer any questions. The two test you require for your medical issue have been automatically schedule two test have 99% viability rate for validating AIs analysis. The typical time to identify cause and get well is typically 3 days.


This ^^^ is whats coming. These are the cases that are being running, ROIs created and the insurance companies/governments are already looking at this model. The governments of the world that manage healthcare are going to jump on this...It will work and then the rest of the private healthcare entities will fall like dominoes...This WILL happen.
Option B sounds great, tbh.

My point was not focused on cost and what drives deciding one treatment plan over another. I am making the point that a great doctor will, for example, influence their patient to lower their A1c through dietary changes before they ever develop diabetes. There are millions of cases where an effective doctor can do something of this nature.

Patient noncompliance is notorious. AI can and will change the way in which we diagnose illness. Getting patients to comply with treatment is a different animal.

Option B really does sound potentially great.
The closer we approach the uncertainty of life's ending the more we wish to trade all of the things we have acquired in exchange for all of the things we have lost: wealth for youth, knowledge for fresh curiosity, resignation for hope. We'd trade our wisdom for new experiences, but it is wisdom that will teach us that at the end of the road the only new experience is death.
post 1596719301 01-30-2020, 01:51 PM
-
#383
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By BOZZ⏩
So what stops the AI from developing a bias? Isn't this a known issue with AI? Like when Google photos turned up pictures of people when you searched for gorillas? I did work with a company using facial recognition and it was terrible at identifying non white people.

Just one example but how is this handled? Or is it just accepted that in certain scenarios it's useless?
Great question...

AI works very well on hard data points right now. AI faces challenges in real time where abstract details are introduced. These challenges present themselves in very prevalent ways when comparing pictures (as you mentioned). These challenges have impacted autonomous driving based off of visuals picture selection. Scientist are looking at finding hard data points within the abstract. In many instances the hard data points are traits you wouldnt think of such the unique identifier. A good example of a unique identifier is a Dorsal fins of a whale. I feel very confident that the scientist will identify unique identifiers for each abstract. It may go deep beyond the abstract data and get down to something of a more technical nature (this is a long explanation that I wont try to make here). All of this will take time because its vertical/industry specific. It will require time and effort. Most industries do not have this challenge, however. They have hard data points and that data is already available and its sitting in database waiting to be siphoned off and used by AI..

When referencing hard data points. The outcomes are being reviewed and anytime there is an anomaly safety check codes/validation/quality checks are being written.
post 1596719731 01-30-2020, 01:55 PM
-
#384
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By propreffered7⏩
Btw. Why do you suggest this is good? Legit want to hear the reasons. If you think ai has this capability cant the autonomous decision maker turn around and use it for evil ?
weak ai is practically a tool that augments reality to make life easier, and thats a debatable generalization. did refined sugars make life easier for us? smartphones? sometimes are boons create this hormetic effects.

super human agi has potential problems way beyond "personhood" and strange loop recursion. what if skynets and yeets us because we asked it to stop pollution?


most likely this is moot, i think its coming regardless.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596719841 01-30-2020, 01:57 PM
-
#385
  1. propreffered7
  2. Registered User
  1. propreffered7
  2. Registered User
  3. Join Date: Mar 2010
  4. Posts: 36,839
  5. Rep Power: 86393
Originally Posted By AltarOfPlagues⏩
weak ai is practically a tool that augments reality to make life easier, and thats a debatable generalization. did refined sugars make life easier for us? smartphones?

super human agi has potential problems way beyond "personhood" and strange loop recursion. what if skynets and yeets us because we asked it to stop pollution?


most likely this is moot, i think its coming regardless.
None of those are nearly close to creating another sentient being.

Like i said, i just want to hear why others would want it

I suspect theres people in here that think doctors make too much money or just hate their prestige
post 1596719931 01-30-2020, 01:58 PM
-
#386
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By MyGhettoFantasy⏩
Option B sounds great, tbh.

My point was not focused on cost and what drives deciding one treatment plan over another. I am making the point that a great doctor will, for example, influence their patient to lower their A1c through dietary changes before they ever develop diabetes. There are millions of cases where an effective doctor can do something of this nature.

Patient noncompliance is notorious. AI can and will change the way in which we diagnose illness. Getting patients to comply with treatment is a different animal.

Option B really does sound potentially great.
You make a very strong point...and AI is absolutely a gateway drug. If we assume the analysis will be AI driven we should expect the prevention to be AI driven. We wait for bad things to happen to us. We wait until we've gone over the waterfall. AI will be preventative...It will monitor you and prevent you from many illnesses. In addition it will confirm you've taken your medicine...How will it confirm you've taken your medicine...Analysis from a smart toilet. How will it confirm your sleep - Your smart bed...etc..etc..etc This is the internet of things and its already here.


People think Im joking about this but this is headed our way...Dont want to comply, Ok...Be prepared to have MUCH higher premiums.


One of the main drivers for AI is to move from reactive to proactive. Let that sink in. Think of all of the bad things that happen to you and think how many of them could have been avoided if you had advanced notice.

In addition think about how this works to improve your life..What if AI decided that you really need a new product because of your data points. The product shows up without you even ordering it. You get to use it for a week to determine if you want it. If you want it, you keep it. If you dont like it AI already knows because you've stopped using the product. A box shows up for you to send it back. The reality is that based upon the data points you already wanted it before you knew you wanted it. The odds you send it back...less than 2%. Cant afford it, no problem but now your conscious mind is aware of the benefits and down the road their is a 99.999999% prediction rate you will end up buying it.
post 1596720011 01-30-2020, 01:59 PM
-
#387
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  1. AltarOfPlagues
  2. yerrrrrrrr meh?
  3. Join Date: Oct 2015
  4. Posts: 14,613
  5. Rep Power: 136737
Originally Posted By propreffered7⏩
None of those are nearly close to creating another sentient being.

Like i said, i just want to hear why others would want it

I suspect theres people in here that think doctors make too much money or just hate their prestige
oh idgaf i think they need it to pay back uncle sam. i just take issue with op saying his diagnosing intellect cant be modeled by a computer bc he caught an HIV diagnosis based on a weird vibe he had.

or because what if a patient is illiterate or doesn't have any hands or something.
do not read my posts and weep, i am not there i do not sleep
i am the thousand greens that rep, i am the ban bet dutifully kept
of memes and trolls in toasted breads, i am not there, i am not dead.
post 1596720771 01-30-2020, 02:12 PM
-
#388
  1. SurgeonBrah
  2. 15 Blade please...
  1. SurgeonBrah
  2. 15 Blade please...
  3. Join Date: Aug 2013
  4. Location: Australia
  5. Posts: 1,699
  6. Rep Power: 5246
Basic summation OTT

Medical/Healthcare crew

- Medicine is complex field requiring humanity at its core to function
- experts in field (surgeons and doctors of the Misc)offer their educated opinion regarding the requirement of emotional aspect to function properly and this is ignored by non-medical background AI proponents
- realize that medicine is more than a business model driven to become more and more efficient at the detriment of above said humanity
- have not been shown any evidence to back up claims of speculation from AI crew

AI crew

- AI learning is beyond scope of medical
Professionals
- extrapolate that because of predictive models suggesting advances that haven’t yet to happen they WILL happen
- continuously ignore the expert views given by medical experts
- give a few examples of how AI can be used in medical field that are way way off being steps towards full automation

Both parties - we will have to agree to
Disagree I think!

both approaching the issue with the same principles of argument - the other side can’t comprehend the complexities of the issue thus the other side is wrong.

Truthfully need an AI implementing expert with a background in clinical medicine to have a proper educated and tempered view of the whole picture.

Cliffs

Arguments on both sides defer to complexities the opposite side cannot comprehend

ends in stalemate ie - misc medicos not persuaded of the AI POV, Misc AI not persuaded of Misc MeD POV...
Plastic/Reconstructive =/= Cosmetic Surgery....

*Mod negged survivor*

Here for a laugh and to escape the grind, don't take too much of what I say too seriously.

Look after your health/body, we only get one shot at it
post 1596720891 01-30-2020, 02:14 PM
-
#389
  1. NVious
  2. Banned
  1. NVious
  2. Banned
  3. Join Date: Apr 2010
  4. Posts: 28,858
  5. Rep Power: 0
More like Luddites vs people who understand history.
post 1596721361 01-30-2020, 02:19 PM
-
#390
  1. Scarrdbutsmartr
  2. Registered User
  1. Scarrdbutsmartr
  2. Registered User
  3. Join Date: Mar 2009
  4. Posts: 20,492
  5. Rep Power: 93529
Originally Posted By SurgeonBrah⏩
Basic summation OTT

Medical/Healthcare crew

- Medicine is complex field requiring humanity at its core to function
- experts in field (surgeons and doctors of the Misc)offer their educated opinion regarding the requirement of emotional aspect to function properly and this is ignored by non-medical background AI proponents
- realize that medicine is more than a business model driven to become more and more efficient at the detriment of above said humanity
- have not been shown any evidence to back up claims of speculation from AI crew

AI crew

- AI learning is beyond scope of medical
Professionals
- extrapolate that because of predictive models suggesting advances that haven’t yet to happen they WILL happen
- continuously ignore the expert views given by medical experts
- give a few examples of how AI can be used in medical field that are way way off being steps towards full automation

Both parties - we will have to agree to
Disagree I think!

both approaching the issue with the same principles of argument - the other side can’t comprehend the complexities of the issue thus the other side is wrong.

Truthfully need an AI implementing expert with a background in clinical medicine to have a proper educated and tempered view of the whole picture.

Cliffs

Arguments on both sides defer to complexities the opposite side cannot comprehend

ends in stalemate ie - misc medicos not persuaded of the AI POV, Misc AI not persuaded of Misc MeD POV...
Im not rebutting anything you've said above but I do have opinions on all of it...

Please be aware that the health insurance companies are one of the largest adopters of AI...

They've already implemented it..Its already running in the background to identify risk and chronic conditions....They dont care about the patient or the doctor. They want to reduce costs

You are debating an end game but the reality is that its already here. Its already being used. There are dozens of case studies you can easily look up that prove AI is churning away without your awareness in the medical community. It will continue to grow as the financials make an undeniable case. The medical community is literally fighting the tide of time and money.. Time and money are relentless opponents and you will be drug out to sea.

and if time and money dont get you the tsunami that you dont see coming is that the patients will demand it...
Quick Navigation Top Misc
Bookmarks
Digg.com
Digg
del.icio.us
del.icio.us
Stumbleupon.com
StumbleUpon
Google.com
Google
Facebook.com
Facebook
Posting Permissions
  1. You may not post new threads
  2. You may not post replies
  3. You may not post attachments
  4. You may not edit your posts