Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Humans vs. Machines - this time is real
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4  Next  
Reply to topic    Gentoo Forums Forum Index Off the Wall
View previous topic :: View next topic  
Author Message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Sat May 09, 2020 4:15 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
ff11 wrote:
richk449 wrote:
I don't think there is any great mystery about how current machine learning techniques work. Sometimes the results, or the effectiveness, can be surprising, but it can always be traced back to understandable principles.

Only for simple task that humans can do (to be replaced by the machine). But like in the movie of AlphaGo, we have the computer team talk like "we don't understand enough of Go to know...". And is we that are learning from the AI now.

Fair enough. Although all the individual steps are understandable, the program can still reach conclusions that we can’t reason about directly.

Even with the individual steps, sometimes we have hard time trying to understand it.
Well, people love to do it anyway, the Google DeepMind AlphaStar, playing StarCraft II, had a lot of popularity because of this.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
richk449
Guru
Guru


Joined: 24 Oct 2003
Posts: 345

PostPosted: Sat May 09, 2020 9:14 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
richk449 wrote:
ff11 wrote:
richk449 wrote:
I don't think there is any great mystery about how current machine learning techniques work. Sometimes the results, or the effectiveness, can be surprising, but it can always be traced back to understandable principles.

Only for simple task that humans can do (to be replaced by the machine). But like in the movie of AlphaGo, we have the computer team talk like "we don't understand enough of Go to know...". And is we that are learning from the AI now.

Fair enough. Although all the individual steps are understandable, the program can still reach conclusions that we can’t reason about directly.

Even with the individual steps, sometimes we have hard time trying to understand it.

Sure, to the extent linear algebra and calculus are hard, the individual steps in machine learning can be hard to understand. If you are proficient at the math though, there is nothing that complicated.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Sat May 09, 2020 9:27 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
...
Sure, to the extent linear algebra and calculus are hard, the individual steps in machine learning can be hard to understand. If you are proficient at the math though, there is nothing that complicated.

Well, like i said before, "i'm one computer scientist too", and we use a lot of Math. And although i believe that everything is possible to those who have faith, i and no one that i know, can make that kind of statement now. Could you help me with this? Could you point some academic journal and some books that can help me to get to this level? Because i fell that i'm missing something here.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
richk449
Guru
Guru


Joined: 24 Oct 2003
Posts: 345

PostPosted: Sat May 09, 2020 9:40 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
richk449 wrote:
...
Sure, to the extent linear algebra and calculus are hard, the individual steps in machine learning can be hard to understand. If you are proficient at the math though, there is nothing that complicated.

Well, like i said before, "i'm one computer scientist too", and we use a lot of Math. And although i believe that everything is possible to those who have faith, i and no one that i know, can make that kind of statement now. Could you help me with this? Could you point some academic journal and some books that can help me to get to this level? Because i fell that i'm missing something here.

Are you familiar with deep neural networks? If you are, and still think that there is something mysterious, then I'm not sure what to say. If you aren't familiar with them, then sure I can give you some suggestions to get started (not that I am anything close to an expert though).
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Sat May 09, 2020 9:51 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
...
Are you familiar with deep neural networks? If you are, and still think that there is something mysterious, then I'm not sure what to say. If you aren't familiar with them, then sure I can give you some suggestions to get started (not that I am anything close to an expert though).

Yes, i'm very familiar with deep neural networks (and the code behind it). I started to study it on the year of 2004 (wow, the time pass fast). It's very useful to me, i have learned many things with its "random nature".
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
richk449
Guru
Guru


Joined: 24 Oct 2003
Posts: 345

PostPosted: Sat May 09, 2020 10:01 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
richk449 wrote:
...
Are you familiar with deep neural networks? If you are, and still think that there is something mysterious, then I'm not sure what to say. If you aren't familiar with them, then sure I can give you some suggestions to get started (not that I am anything close to an expert though).

Yes, i'm very familiar with deep neural networks (and the code behind it). I started to study it on the year of 2004 (wow, the time pass fast). It's very useful to me, i have learned many things with its "random nature".

okay, so then your turn to explain. NNs just map inputs to outputs in a deterministic way. Why do you find that mysterious?
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Sat May 09, 2020 10:16 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
...
okay, so then your turn to explain. NNs just map inputs to outputs in a deterministic way. Why do you find that mysterious?

Well, i don't understand why is my turn to explain, you only made affirmations without explanations till now.
But anyway, i don't think that the implementation is mysterious in any way. But watching someone like Mark Sagar (CEO of Soul Machines) talking with his BabyX, gives me a very mysterious feeling, like the implementation wasn't as important as it used to be.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Sun May 10, 2020 12:23 am    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
Yea, originally I was just trying to make the point that computers weren't "winning", if winning means being able to do the most sophisticated tasks. They are definitely automating away the easy stuff, but in a way, even that could be a win for humans, if they play their cards right.
We probably agree on a lot of the current / near future aspects of AI. My primary concerns are: the potential for a proportionally larger loss of jobs than with previous displacements (buggy whip, etc.); and accelerated loss of knowledge or skills as a result of job displacement or "assistive" use of AI.

I think the first one is obvious, so with regards to what I mean by the second point, consider a couple of scenarios. First, my earlier game engine reference. The skill set in how to create a game engine may be beneficial, but if the knowledge is lost. The impact there is relatively unimportant, but I think it demonstrates the issue. Another example is with tiered skill sets: entry, mid, senior. If entry level jobs are replaced with an AI, how does one gain experience to step into a mid level position? A parallel is in outsourcing. Get rid of the entry positions and, feigned surprise, there are no qualified people for mid positions. At least with outsourcing the argument can be made that the job didn't go away, it was just made unavailable to an "undesirable" pool of candidates.

With all of that out of the way, I'd ask for more details on how you see a win for humans and what cards need to be played and by whom? As you specified 'humans,' my presumption is a big picture that likely includes "buggy whip" losses. To be clear, I'm not arguing against any losses, only that I think rapid advancement could see a scale of losses that is unsustainable.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Sun May 10, 2020 12:27 am    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
That's make the Video #2 perfect to make the point that many people are ignoring: That we are making one revolution now, using something advanced that we don't understand how it works at all (it's literally coming out of random), but we don't care either, we care about the advancements, the money, and the potential, no mater the costs.
That was not apparent to me in the portion I watched. It didn't seem to say much of value. I would not personally recommend it to anyone.

ff11 wrote:
It's just the History repeating itself.
Which happens a lot. What specific part of history repeating is concerning to you that relates to AI? As yet, I don't think this is particularly different than the "buggy whip." Because of the nature of our job base, I believe there is high potential to put a lot of people out of work in a relatively short period of time. But I don't believe that is going to happen in less than 12 months; longer than 12 months, because I think there is potential for a breakthrough which could happen at any point.

ff11 wrote:
If you think about this, let's say, hypothetically, that some god make a lot of humans to make one specific task (at random), then after a large number of humans, this god found one that can make this task, and after that, this human will do only this task, then can we really think that this human can do only this task, nothing more?!
I don't agree with humans and AI being equivalent. I think it is very possible for an AI to be created that is not capable of creating a new AI. I do think it will be possible that an AI could exist that creates other AIs.

ff11 wrote:
So i will recommend that you read this first (only this part2 is enough): https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html
because by the time we can claim that AI has gained awareness, it may already be too late.
What may be too late? The reference seems a bit long, and the start seemed questionable. What would you say are its primary points? I don't think there is a short term doomsday scenario related to AI. I don't rule out the possibility of that being an issue some day, I just don't think it is on my short list of worries for the average person. I think the average person should be concerned with ethical matters, but at least in the US, they don't tend toward that direction. No one seems to care about data breeches and various other crimes relying on personal information.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Sun May 10, 2020 1:05 am    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

pjp wrote:
That was not apparent to me in the portion I watched. It didn't seem to say much of value. I would not personally recommend it to anyone.

Well, the entire video is an explanation of how to make a bot without having a clue how it works. And i think that is a very good one. Perhaps it's a cultural issue here.

pjp wrote:
Which happens a lot. What specific part of history repeating is concerning to you that relates to AI? As yet, I don't think this is particularly different than the "buggy whip." Because of the nature of our job base, I believe there is high potential to put a lot of people out of work in a relatively short period of time. But I don't believe that is going to happen in less than 12 months; longer than 12 months, because I think there is potential for a breakthrough which could happen at any point.

Any historical big economic revolution (there is always a lot of excluded people). And yes, it won't have a big impact anytime soon, with just a few months. However, if we add the current economic problems, everything can turn into a snowball, and this revolution starts faster than we think.

pjp wrote:
I don't agree with humans and AI being equivalent. I think it is very possible for an AI to be created that is not capable of creating a new AI. I do think it will be possible that an AI could exist that creates other AIs.

Neither do I, but it's also not a simple thing either. Well, I can say that we are doing our best to create one AI that is capable of creating and controlling other AIs (eventually one will appear, if it hasn't already).

pjp wrote:
What may be too late? The reference seems a bit long, and the start seemed questionable. What would you say are its primary points? I don't think there is a short term doomsday scenario related to AI. I don't rule out the possibility of that being an issue some day, I just don't think it is on my short list of worries for the average person. I think the average person should be concerned with ethical matters, but at least in the US, they don't tend toward that direction. No one seems to care about data breeches and various other crimes relying on personal information.

Well, the simple loss of one job along with the inability to find a new one, would be too late for me. But there are hypothetical scenarios that great business leaders are concerned about as well. Then let's just say that the more reasoning we have on the issue, the better (even if some ideas seem crazy). As I said earlier, I also don't believe in major impacts in the coming months. But precisely because of how everything is going on, it's like saying that the main agents are allowed to do whatever they want (even if it break the economy even more). And I just like to explore all the scenarios, you can ignore any part you want. In the end, I don't really care that much about the issue myself, I'm more interested in other people's opinions on the subject.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Mon May 11, 2020 6:53 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
Well, the entire video is an explanation of how to make a bot without having a clue how it works. And i think that is a very good one. Perhaps it's a cultural issue here.
I forced myself to watch the whole thing. In my opinion, it only makes a statement about "not having a clue how it works" and doesn't demonstrate that to be true. I believe they are making a statement that has a technical truth to it, but that the statement is a disservice to the issue and partly misrepresents that underlying issue. I believe anyone watching that video who comes away with a significantly negative opinion about machine learning, or who has more than a passing concern about machine learning as a result of watching that video, demonstrates the problem with their statement as superficially presented.

ff11 wrote:
However, if we add the current economic problems, everything can turn into a snowball, and this revolution starts faster than we think.
That seems like worrying for the sake of worrying. The problem in this scenario would be the snowballing of multiple issues, not some potential problem with machine learning (I think it is incorrect to call the current state of ML "intelligent.")

ff11 wrote:
Well, the simple loss of one job along with the inability to find a new one, would be too late for me. But there are hypothetical scenarios that great business leaders are concerned about as well. Then let's just say that the more reasoning we have on the issue, the better (even if some ideas seem crazy). As I said earlier, I also don't believe in major impacts in the coming months. But precisely because of how everything is going on, it's like saying that the main agents are allowed to do whatever they want (even if it break the economy even more). And I just like to explore all the scenarios, you can ignore any part you want. In the end, I don't really care that much about the issue myself, I'm more interested in other people's opinions on the subject.
I agree there is a problem where people will lose jobs and be unable to pivot. That isn't new. My primary concern is the possible scale at which that will happen.

Referencing the video that talks about the 45% of the workforce that is at risk, if we consider a relatively quick loss of only 25% of those jobs, that's 15.75M people losing a job. That alone would be a 11.25% unemployment rate on top of whatever the baseline would be at that time. If we use a range of 3 - 10%, that's 14.25 - 21.25% unemployment. I believe it was the same video that referred to the Great Depression having a 25% unemployment rate.

That's my concern. And what, if anything, corporations or government(s) will or should do to stop that from happening as quickly. A great depression isn't going to do the companies who create the problem any favors, independent of any government attempt to "help." This is what I think is behind the talk of Universal Basic Income. But UBI isn't a solution as I don't believe most people want to exist without a purpose. And despite some significant efforts to demonstrate otherwise, participating in social media isn't a (healthy) purpose.


Oh, I did watch the AlphaGo video (I skipped the parts not in English). Interesting. There's a point at which one of the AlphaGo team members briefly explains how it works. In some ways, it is a lot like DeepBlue. But I still don't think there is any evidence that there is an underlying intelligence operating with anything that resembles sentience.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Mon May 11, 2020 7:35 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

pjp wrote:
I forced myself to watch the whole thing. In my opinion, it only makes a statement about "not having a clue how it works" and doesn't demonstrate that to be true. I believe they are making a statement that has a technical truth to it, but that the statement is a disservice to the issue and partly misrepresents that underlying issue. I believe anyone watching that video who comes away with a significantly negative opinion about machine learning, or who has more than a passing concern about machine learning as a result of watching that video, demonstrates the problem with their statement as superficially presented.

We need some adaptation there, but i think that is possible to get the basic ideia. And for one video that have (at this day) 325 MIL likes and only 2,2 MIL dislikes, i kinda like how the job is done. Maybe not a video for "computer professionals", but the same level access requirements to find the video is apply to this forum thread too (ins't private, nor for some group of professionals too).

pjp wrote:
That seems like worrying for the sake of worrying. The problem in this scenario would be the snowballing of multiple issues, not some potential problem with machine learning (I think it is incorrect to call the current state of ML "intelligent.")

Well, what "intelligent" means can be a very long philosophical question, so I don't intend to discuss this here.

pjp wrote:
I agree there is a problem where people will lose jobs and be unable to pivot. That isn't new. My primary concern is the possible scale at which that will happen.

Referencing the video that talks about the 45% of the workforce that is at risk, if we consider a relatively quick loss of only 25% of those jobs, that's 15.75M people losing a job. That alone would be a 11.25% unemployment rate on top of whatever the baseline would be at that time. If we use a range of 3 - 10%, that's 14.25 - 21.25% unemployment. I believe it was the same video that referred to the Great Depression having a 25% unemployment rate.

That's my concern. And what, if anything, corporations or government(s) will or should do to stop that from happening as quickly. A great depression isn't going to do the companies who create the problem any favors, independent of any government attempt to "help." This is what I think is behind the talk of Universal Basic Income. But UBI isn't a solution as I don't believe most people want to exist without a purpose. And despite some significant efforts to demonstrate otherwise, participating in social media isn't a (healthy) purpose.


Oh, I did watch the AlphaGo video (I skipped the parts not in English). Interesting. There's a point at which one of the AlphaGo team members briefly explains how it works. In some ways, it is a lot like DeepBlue. But I still don't think there is any evidence that there is an underlying intelligence operating with anything that resembles sentience.

I think the world today already needs UBI to be implemented globally, just as i also think that massive depression should start to be tackled now. But is only my opinion.
The AlphaGo movie have English subtitle too (if someone else want to watch).
Well, the AlphaGo use human experience but the AlphaGo Zero don't (Zero is only cited on the end of the movie), and now we only know that some specifics neural network can do some task, if it can do something else is another history. There is the some people like Mark Sagar that is creating babyX as if it were his own daughter (it's worth following his work).
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Mon May 11, 2020 11:57 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

ff11 wrote:
We need some adaptation there, but i think that is possible to get the basic ideia. And for one video that have (at this day) 325 MIL likes and only 2,2 MIL dislikes, i kinda like how the job is done. Maybe not a video for "computer professionals", but the same level access requirements to find the video is apply to this forum thread too (ins't private, nor for some group of professionals too).
I think that summarizes the problem with the video. Prey upon people without enough knowledge to instill unwarranted fear.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Tue May 12, 2020 12:22 am    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

pjp wrote:
I think that summarizes the problem with the video. Prey upon people without enough knowledge to instill unwarranted fear.

Well, many Hollywood movies do that too, and I don't see it to be a problem if it helps to make people more interested. But again, it's just my opinion.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Tue May 12, 2020 12:49 am    Post subject: Reply with quote

I don't think that video made any disclaimers or clear indications that it was fiction.
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
richk449
Guru
Guru


Joined: 24 Oct 2003
Posts: 345

PostPosted: Tue May 12, 2020 4:01 am    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

pjp wrote:
richk449 wrote:
Yea, originally I was just trying to make the point that computers weren't "winning", if winning means being able to do the most sophisticated tasks. They are definitely automating away the easy stuff, but in a way, even that could be a win for humans, if they play their cards right.
We probably agree on a lot of the current / near future aspects of AI. My primary concerns are: the potential for a proportionally larger loss of jobs than with previous displacements (buggy whip, etc.); and accelerated loss of knowledge or skills as a result of job displacement or "assistive" use of AI.

I think the first one is obvious

The concern is obvious. It is not obvious that it will actually come to pass. Personally I think it is likely - automation will wipe out a very large number of jobs this round. But I don't think I can justify that, and I know my personal record at predicting the future isn't great.

Quote:
so with regards to what I mean by the second point, consider a couple of scenarios. First, my earlier game engine reference. The skill set in how to create a game engine may be beneficial, but if the knowledge is lost. The impact there is relatively unimportant, but I think it demonstrates the issue. Another example is with tiered skill sets: entry, mid, senior. If entry level jobs are replaced with an AI, how does one gain experience to step into a mid level position? A parallel is in outsourcing. Get rid of the entry positions and, feigned surprise, there are no qualified people for mid positions. At least with outsourcing the argument can be made that the job didn't go away, it was just made unavailable to an "undesirable" pool of candidates.

This one worried me less. If companies need mid-level developers, then they will hire juniors and train them to the level they need. Lot's of industries do this currently - welding, for instance, or CAD modeling, for a more modern example.

Quote:
With all of that out of the way, I'd ask for more details on how you see a win for humans and what cards need to be played and by whom? As you specified 'humans,' my presumption is a big picture that likely includes "buggy whip" losses. To be clear, I'm not arguing against any losses, only that I think rapid advancement could see a scale of losses that is unsustainable.

As I said, I think it is likely that automation will have a pretty big impact on our society in the short term. However, the broader framework is that higher efficiency is always better in the long term. Factory farming, which eliminated a large fraction of farming jobs, made food cheap enough that very few people in non-third world countries go hungry these days. The automobile, which allowed individual travel at a cost working families could afford, greatly improved job opportunities, and helped create the conditions to pull millions out of poverty.

And this current round of automation promises the mother of all efficiency increases. It seems likely that we will look back 100 years from now and talk about the ML revolution the way we talk about the industrial revolution now. So that is what I mean by playing our cards right. To capture the benefits of automation without the incredible disruption that could occur.

I don't exactly know how to play our cards. I don't like UBI. I prefer to invest heavily in education. Make college standard, and push back the age at which people typically start working - that will reduce supply of employees. But that is just a small piece of the puzzle. Invest seriously in job retraining, perhaps. Reduce deficit spending during economic growth, so that the government has the ability to step in when necessary. But now we are getting into motherhood and apple pie.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Tue May 12, 2020 10:08 am    Post subject: Reply with quote

pjp wrote:
I don't think that video made any disclaimers or clear indications that it was fiction.

Well, there is one "Footnote video" on the footnotes for this. That by the way, have more footnotes videos with even more precise information, that someone really interested can easily follow. So, I sill don't understand your complain.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Tue May 12, 2020 8:38 pm    Post subject: Re: Humans vs. Machines - this time is real Reply with quote

richk449 wrote:
The concern is obvious. It is not obvious that it will actually come to pass. Personally I think it is likely - automation will wipe out a very large number of jobs this round. But I don't think I can justify that, and I know my personal record at predicting the future isn't great.
I agree. I don't know if ML/AI has replaced an actual job at this point, but it seems among the closest to guaranteed bets a person could make. How many jobs is the part I think is unpredictable. General automation has replaced jobs, so how the automation happens seems a semantic argument. I suppose it is possible that ML will not directly cause job losses.

richk449 wrote:
This one worried me less. If companies need mid-level developers, then they will hire juniors and train them to the level they need. Lot's of industries do this currently - welding, for instance, or CAD modeling, for a more modern example.
You're probably correct. It could be that the entry point for a developer becomes higher / more difficult

richk449 wrote:
As I said, I think it is likely that automation will have a pretty big impact on our society in the short term. However, the broader framework is that higher efficiency is always better in the long term. Factory farming, which eliminated a large fraction of farming jobs, made food cheap enough that very few people in non-third world countries go hungry these days. The automobile, which allowed individual travel at a cost working families could afford, greatly improved job opportunities, and helped create the conditions to pull millions out of poverty.

And this current round of automation promises the mother of all efficiency increases. It seems likely that we will look back 100 years from now and talk about the ML revolution the way we talk about the industrial revolution now. So that is what I mean by playing our cards right. To capture the benefits of automation without the incredible disruption that could occur.

I don't exactly know how to play our cards.
Again, you're probably correct. Improvements to farming efficiency seem obvious, and the job loss impact relatively low (except for the person losing the job). The manner of how that was achieved seems to have had a high cost (fertilizers, concentrated animal waste, inflexible supply chain, ...). And farmer's markets continue to seem like a luxury for those who enjoy complaining about factory farming :).

Automobiles seem like they would have provided a net gain in jobs, though I'm sure there were permanent losses related to equine support services. I don't know that it is as obvious that factory farming produced a net gain in jobs.

I don't see where there is likely to be a net gain of jobs due to ML/AI automation, or even a displacement of roughly the same number of jobs. Kiosks alone seem likely to eliminate some jobs. Bank tellers haven't been eliminated, but the ATM and use of credit seems to have reduced the number of tellers I see in banks. Fast food has been incorporating kiosks as well. As far as I know, neither of those rely on AI or ML.

richk449 wrote:
I don't like UBI. I prefer to invest heavily in education. Make college standard, and push back the age at which people typically start working - that will reduce supply of employees. But that is just a small piece of the puzzle. Invest seriously in job retraining, perhaps. Reduce deficit spending during economic growth, so that the government has the ability to step in when necessary. But now we are getting into motherhood and apple pie.
UBI seems like a really bad idea unless it were used as an emergency band-aid due to drastic and unexpected job losses. Education in general is fine, but I don't see that education directly increases job opportunities. I'm specifically excluding arbitrary requirements that a candidate have a certain level of higher education. I'm not excluding obvious needs such that a person must learn to be a medical doctor before practicing medicine. Presuming significant losses in low end jobs, education does not seem a guarantee that jobs will exist if after the person completes their education program. The "gig economy" seems to demonstrate that; at least I'm not aware of an efficiency which resulted in fewer jobs for graduates in the last ~20 years (or whatever range).
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Tue May 12, 2020 8:48 pm    Post subject: Reply with quote

ff11 wrote:
So, I sill don't understand your complain.
I'm not sure how else to clarify. I think it is misleading and was done with the intent to produce a fear based reaction.

You seemed to demonstrate a degree of "fear" in your description ("is at least scare"):
ff11 wrote:
2- How Machines Learn
Well, something that nobody know how it works, and that works to survive, is at least scare.

_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Tue May 12, 2020 10:05 pm    Post subject: Reply with quote

pjp wrote:
ff11 wrote:
So, I sill don't understand your complain.
I'm not sure how else to clarify. I think it is misleading and was done with the intent to produce a fear based reaction.

You seemed to demonstrate a degree of "fear" in your description ("is at least scare"):
ff11 wrote:
2- How Machines Learn
Well, something that nobody know how it works, and that works to survive, is at least scare.

Well, yes, but isn't from the video itself, it's for the first part of the statement: "something that nobody know how it works, and that works to survive". It ins't one description of the video, is just one personal comment.
If is something simple like the "Evolution Simulator code" (that can be found following the footnotes), then there is nothing to be scare, but that is only the basic (just one implemented model), we are using it to complex task now, and we only can be sure that it can do the complex task, if it can do unnecessary things outside the task, nobody seems to care.
Although we already know that it does unnecessary things within the complex task. It do the best until the pass test point then start to do unnecessary things. Maybe it’s just me who thinks this is really weird, and my imagination going wild because of it.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Tue May 12, 2020 10:37 pm    Post subject: Reply with quote

Do you know of an example where machine learning has done unnecessary things outside the task?
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Tue May 12, 2020 11:52 pm    Post subject: Reply with quote

pjp wrote:
Do you know of an example where machine learning has done unnecessary things outside the task?

Not that I can remember now, or at least not that I have personally analyzed it. Only within the task, mainly because we usually limit it, but it isn't a rule, we just do it.
Maybe Mark Sagar have some answers at this point, because it's part of his project with babyX.
Sorry, I have no concrete answers now.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
pjp
Administrator
Administrator


Joined: 16 Apr 2002
Posts: 18558

PostPosted: Wed May 13, 2020 1:19 am    Post subject: Reply with quote

I was just curious what "outside the task" meant in a specific case and why it was concerning. If there are no incidents, then concern seems unwarranted..
_________________
Your lips move, but I can't hear what you're saying.
Back to top
View user's profile Send private message
richk449
Guru
Guru


Joined: 24 Oct 2003
Posts: 345

PostPosted: Wed May 13, 2020 3:38 am    Post subject: Reply with quote

https://cacm.acm.org/news/244846-can-ai-become-conscious/fulltext
Back to top
View user's profile Send private message
ff11
l33t
l33t


Joined: 10 Mar 2014
Posts: 644

PostPosted: Wed May 13, 2020 12:08 pm    Post subject: Reply with quote

richk449 wrote:
https://cacm.acm.org/news/244846-can-ai-become-conscious/fulltext

Good reference, but they are still playing with human words definitions instead of looking at the core of the question, if you know what i mean.

pjp wrote:
I was just curious what "outside the task" meant in a specific case and why it was concerning. If there are no incidents, then concern seems unwarranted..

Well, it can get complex to explain using simple things. But I will try.
First I would like to say that an ordinary person of modern times will not normally feel this fear. So you don't have to worry about "causing unnecessary panic" here. And that we kinda don't need to worry about the incidents either, as I said before, what humans do is much more worrying.
And again, because of modern culture, many of us have a different sense of fear than the natural one. I'll bring up this reasoning again:
ff11 wrote:
...
From: https://waitbutwhy.com/2015/01/artificial-intelligence-revolution-2.html
Quote:
...
Let me draw a comparison. If you handed me a guinea pig and told me it definitely won’t bite, I’d probably be amused. It would be fun. If you then handed me a tarantula and told me that it definitely won’t bite, I’d yell and drop it and run out of the room and not trust you ever again. But what’s the difference? Neither one was dangerous in any way. I believe the answer is in the animals’ degree of similarity to me.

A guinea pig is a mammal and on some biological level, I feel a connection to it—but a spider is an insect,18 with an insect brain, and I feel almost no connection to it. The alien-ness of a tarantula is what gives me the willies. To test this and remove other factors, if there are two guinea pigs, one normal one and one with the mind of a tarantula, I would feel much less comfortable holding the latter guinea pig, even if I knew neither would hurt me.

Now imagine that you made a spider much, much smarter—so much so that it far surpassed human intelligence? Would it then become familiar to us and feel human emotions like empathy and humor and love? No, it wouldn’t, because there’s no reason becoming smarter would make it more human—it would be incredibly smart but also still fundamentally a spider in its core inner workings. I find this unbelievably creepy. I would not want to spend time with a superintelligent spider. Would you??

When we’re talking about ASI, the same concept applies—it would become superintelligent, but it would be no more human than your laptop is. It would be totally alien to us—in fact, by not being biology at all, it would be more alien than the smart tarantula.

By making AI either good or evil, movies constantly anthropomorphize AI, which makes it less creepy than it really would be. This leaves us with a false comfort when we think about human-level or superhuman-level AI.
...

If we add why fear is important to a biological being (NOTE: if someone reading this, think that every fear is unnecessary, then that person should do some extra search about the subject first), then, at least in my view, sometimes not feeling fear can be as problematic as feeling fear for the wrong reason.
The point that we can't forget here is that AI isn't a living thing. Then let's go back to "within the task" using 2 user's case (there are better cases, but I will use the already mentioned ones):
1- The simple Evolution Simulator: After we get some "useful" Creature, we can see that it does the job well, even if it can't complete the job, it keeps doing the best it can; and even though it has already achieved the goal, it keeps doing the best it can; it doesn't forget what it is and start doing unnecessary things, doing the task is what it is, so it does the task all the time that is put to run. You can say "this is obvious", because it is just a program that was made for this, which takes us to the second case:
2- The AlphaGo version of the movie: Here, we can identify 3 modes of operation, with only 1 being desired: The "doing the best" mode, the "already gave up" mode and the "already won" mode. No one wants or expects the "already gave up" and "already won" modes, be it the developers who don't want to be ashamed, or the players who want to see the best game. We all want "Creature from Evolution Simulator" that is always doing the best it can. And in the movie, as well as in many other sources, we hear arguments like "it doesn't care to win in the best possible way, only in winning, to reaching the goal". What?! Now we start talking about what does AI care about?! It has never been a living thing in the first place, and certainly the programmers didn't do it to embarrass themselves, the question should be "why is it clearly not doing the best it can?" (like the "Creature from Evolution Simulator"). One could argue that this behavior is only being imitated based on human data, but the Zero version that does not have human data also does this and has the same 3 modes of operation.
And here, about the AI, I will say that I don't feel any degree of similarity to me, remembering that it isn't a biological being at all. And just like in the movie, at least, there will be a fear of embarrassing myself on a presentation using AI.
_________________
| Proverbs 26:12 |
| There is more hope for a fool than for a wise man that are wise in his own eyes. |
* AlphaGo - The Movie - Full Documentary
"I want to apologize for being so powerless" - Lee
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Off the Wall All times are GMT
Goto page Previous  1, 2, 3, 4  Next
Page 2 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum