AI Pause Called For By Elon Musk, Steve Wozniak, and other tech leaders

[Uttamattamakin]

Summary

Elon Musk, Steve Wozniak and other tech leaders have signed an open letter calling for a pause on the development of AI systems.  The letter expresses concerns for the societal impacts of AI and if they will be good for humanity.  The letter states that "Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable ".  The letter points out lack of government regulation, and already existing corporate uses of AI.  The letter points out that pats of jobs or whole jobs that people find fulfilling are being automated away replacing human discernment with machine learning and AI that gives the impression of being AGI. They call for a moratorium on producing AI's more powerful than GPT-4

 

Quotes

 

The open letter itself says this. 

Quote

Contemporary AI systems are now becoming human-competitive at general tasks,^[3] and we must ask ourselves: Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization? Such decisions must not be delegated to unelected tech leaders. Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable. This confidence must be well justified and increase with the magnitude of a system's potential effects. OpenAI's recent statement regarding artificial general intelligence, states that "At some point, it may be important to get independent review before starting to train future systems, and for the most advanced efforts to agree to limit the rate of growth of compute used for creating new models." We agree. That point is now.

Therefore, we call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4. This pause should be public and verifiable, and include all key actors. If such a pause cannot be enacted quickly, governments should step in and institute a moratorium.

 

Over on Twitter Mario Nawfal has the points in brief of the letter. 

 

Quote

Elon Musk & Steve Wozniak Ask for a Pause on A.I due to ‘Profound Risks to Society’ Elon Musk, Steve Wozniak, Andrew Yang and other key figures have called for an immediate pause on A.I. due to the profound risk it poses society. The issues they highlighted were:

1. A.I. developers are locked in a race to deploy powerful digital ‘minds’ that even the creators cannot understand, predict or control.

2. Gary Marcus argued that A.I. is unnecessarily shaping the direction of our world.

3. “A.I. development has seen corporate irresponsibility, widespread adoption, lack of regulation and a huge number of unknowns”.

4. Elon Musk asked for a pause on systems that are more powerful than GPT-4 to introduce safety protocols.

5. The first letter asks for government intervention to institute a moratorium, if there isn’t an immediate pause.

6. A.I. development should only continue once we know it’s positive impact, and that we can manage the risks 

 

 

 

 

My thoughts

 

Yahoo finance has a great video on this too but it's more opinionated.  I largely agree with what they say though. 

 

The concerns raised in the letter are very real and already concrete in places where it has been used in the field.  AI is good as a tool but people have a tendency to lean on AI instead of using their own judgement.   AI is very good at recognizing pre-existing patterns.  AI is good at narrowing down the possibilities.  AI is good at giving a skeleton of what a human mind can flesh out with real work.  That is also the problem is AI can be, like GPT3 was "confidently wrong".   It would also be wrong to think Musk, Woz, and the others are only concerned with chat bots. Those are just the newest examples of AI being used to replace human mental work.  Other examples have been deployed in the real world for years. 

 

AI has already had disastrous effects on some people in the US.  For example more than one person has been put in jail on weak evidence because AI thought they looked like a criminal.  How AI-powered Tech Landed A Man In Jail With Scant Evidence (Associated Press News)  

 

Quote

By GARANCE BURKE, MARTHA MENDOZA, JULIET LINDERMAN and MICHAEL TARM

March 5, 2022

 

CHICAGO (AP) — Michael Williams’ wife pleaded with him to remember their fishing trips with the grandchildren, how he used to braid her hair, anything to jar him back to his world outside the concrete walls of Cook County Jail.

 

His three daily calls to her had become a lifeline, but when they dwindled to two, then one, then only a few a week, the 65-year-old Williams felt he couldn’t go on. He made plans to take his life with a stash of pills he had stockpiled in his dormitory.

 

Williams was jailed last August, accused of killing a young man from the neighborhood who asked him for a ride during a night of unrest over police brutality in May. But the key evidence against Williams didn’t come from an eyewitness or an informant; it came from a clip of noiseless security video showing a car driving through an intersection, and a loud bang picked up by a network of surveillance microphones. Prosecutors said technology powered by a secret algorithm that analyzed noises detected by the sensors indicated Williams shot and killed the man.

“I kept trying to figure out, how can they get away with using the technology like that against me?” said Williams, speaking publicly for the first time about his ordeal. “That’s not fair.”

 

Williams sat behind bars for nearly a year before a judge dismissed the case against him last month at the request of prosecutors, who said they had insufficient evidence.

I live in Chicago land and the little suburb I live in call it, ParkForrestRiverGlennWoodLake (People from Chicago will get the joke... every village here has one of those words in the name).   Our little village just outside of Chicago has cameras and the same kind of gun shot detecting AI on them.  In fact we had it before Chicago.  However, as far as I know our police have never been lazy enough to try and arrest someone based only on those cameras and their findings.  Events like this demonstrate that the concerns raised in that letter are not hypothetical, theoretical, opinions or  but real and concrete observed effects AI will has had.   There have been multiple incidents like this around AI's use by law enforcement. 

 

We already have seen artificial intelligence used to replace human judgement mainly by people who don't want to do the mental work of their job.   Instead they use AI to automate the same old ... lets call them shortcuts they used before. 

 

Speaking of mental work. 

 

As an educator I have grave concerns about the use of AI in replacing human thought.  This is not just.... Durr I don't want to change how I teach.   Education is about training human minds to do work.  I believe that our minds are the most important part of us and like any system without exercise and use they will atrophy.    There is a reason that when math is taught up to a certain level you don't get to use a calculator.  After students have mastered the concepts of arithmetic and memorized their times tables they learn algebra.  Once students have learned the principles of algebra they then get to use a graphing calculator.   To understand how to code a computer program one must know the basics of logic.  To repair technology one needs to know how the technology works. 

Most of what I have to teach, most of the courses needed are developmental, remedial, math classes.  Even when I get to tech Physics or Astronomy the students in class will be unable to multiple 2 meters/second by 2 kilograms to compute the momentum of an object.  Not because they don't know what 2*2=4  but because they don't actually know the concepts of multiplication.  Instead they memorized some tables and then used a calculator but never went deeper.    Why?  Because people will take a short cut if they can. 

 

Perhaps even worse than that many of them will swear they "knew" the answers but were afraid to be wrong by writing them down if they can't use the calculator.   They won't even take a chance on working without the technology. 

 

Now imagine having people being petrified of not writing exactly like an AI chatbot and totally unwilling to type out their own original thoughts. 

 

Long LONG term use of AI to replace human intelligence will lead to a devaluation of human mental work.  That in the most comical extreme leads us to being a real life Idiocracy. 

Where at best the doctor is like. 

A world where AI has been used to replace lost human intelligence.  I also wanted to end on a funny note.  ANY THOUGHTS? Any other examples already existing of AI gone wrong? 

 

Sources

https://futureoflife.org/open-letter/pause-giant-ai-experiments/  (The Letter)

CBS Evening news (For those of you not from the US CBS is one of our big three original TV networks)

https://www.govtech.com/biz/ai-pause-urged-by-musk-wozniak-and-other-tech-leaders

 

[Holmes108]

Geez, I just don't know. 

 

I agree with most all the points brought up here and in recent discussions. All the concerns are valid. But I also don't know how you just "pause" a technology. It seems overly simplistic. Even if we agreed as an entire society within this country/contient... every politician, individual and company (highly unlikely), there's still a whole world out there looking to exploit it. 

 

You can look at "pausing" it's use on case by case basis. Industry by industry, or job by job. But AI is here, and it's going to be used. We're better off just figuring out exactly how to live it it as quickly and as best as possible (whatever that means). 

 

I haven't really provided a solution here, but I'm not convinced there is one. It's been said before, you can't put the genie back in the bottle. Why didn't we just "pause" our knowledge of splitting the atom? That's just not how the world works.

[will0hlep]

The genie isn't going back in the bottle. There will be no pause. Instead of them proposing stupid stuff like this "Pause" they should actually help with getting governments ready for what is coming. It's very likely that there will be mass unemployment soon as jobs the computers can't do dry up. If that does happen the economy is going to change very quickly and the world needs to be ready for that to happen.

 

The cynical part of me just thinks Elon Musk wants time for his RL devision to catch up to competitors .

[Uttamattamakin]

15 minutes ago, Holmes108 said:

Geez, I just don't know. 

 

I agree with most all the points brought up here and in recent discussions. All the concerns are valid. But I also don't know how you just "pause" a technology. It seems overly simplistic. Even if we agreed as an entire society within this country/contient... every politician, individual and company (highly unlikely), there's still a whole world out there looking to exploit it. 

Exactly to get a meaningful pause would take an international treaty or something like that.  I think the United Nations Security Council has more important things to worry about ... but maybe the countries that are behind on AI would be willing to agree in order to get sharing of this technology.  To make sure that AI does not give anyone an overwhelming advantage?

 

15 minutes ago, Holmes108 said:

 

You can look at "pausing" it's use on case by case basis. Industry by industry, or job by job. But AI is here, and it's going to be used. We're better off just figuring out exactly how to live it it as quickly and as best as possible (whatever that means). 

That might work within our country.  You know banning AI from being used to do anything that would adversely effect anyone.  Pause it's use in say healthcare, legal, educational, military or any other matters that effect life and limb.   Not until we can be 100% sure they are as able to make decisions as a human being.  Then never put one AI system in charge of big decisions all by itself.   You know how we split power up among people in real life. 

 

15 minutes ago, Holmes108 said:

I haven't really provided a solution here, but I'm not convinced there is one. It's been said before, you can't put the genie back in the bottle. Why didn't we just "pause" our knowledge of splitting the atom? That's just not how the world works.

We didn't need to pause on splitting the atom.  We actually took from the 300's BC when atomic theory was first proposed ... then it took from 1905 (Einstein's annus mrabilis papers including E=Mc^2) to 1945 for the Abomb.  That technology took 40 years to become a thing from concept to reality.    10 years from the first split atom to the Atom bomb. Another 10 after that for the H bomb.    I say all of this to make the point we didn't need to pause atomic research because it was really REALLY hard and took a REALLY long time.  

That said at that time we didn't have computers like we do now.  For the last part of it we had things like ENIAC and UNIVAC.  Seriously look at this thing.  
 

 Plus this also relates as a sort of early concept of using AI.   AI has a long time to go before we really "split the atom" of creating a human like AGI.  The difference now is we have tools that can make this all happen MUCH faster and perhaps before we are wise enough to do it right. 

 

2 minutes ago, will0hlep said:

The genie isn't going back in the bottle. There will be no pause. Instead of them proposing stupid stuff like this "Pause" they should actually help with getting governments ready for what is coming.

 

The cynical part of me just thinks Elon Musk wants time for his RL devision to catch up to competitors .

You may be right.  It would take something kin to the international treaty that banned CFC's globally.  Power brokers would need to be convinced that AI poses a threat similar to that. 

[will0hlep]

10 minutes ago, Uttamattamakin said:

You may be right.  It would take something kin to the international treaty that banned CFC's globally.  Power brokers would need to be convinced that AI poses a threat similar to that. 

yeah, I really don't see that happening again atm, the world's government don't seem to get along atm and also the arguement over whether or not it's a bad thing is still on going. A world where jobs are optional cause the AI can do it better than you and therefore you only have to do what you find fulfilling seems like it could be a utopia (if we could do a half decent job of getting everyone what they need to survive). (I do accept that it could also be turned into hell at the hand of misinformation campaigns and greed, but I think these are issue we are just going to need to address.)

[Holmes108]

As far as the "we're all going to lose our jobs" aspect... well in the past I've always been on the opposite side. I've always argued that people have been saying that for over 100 years with every innovation. And while some industries certainly get affected, as a whole, we've always become more wealthy as a society. Net positive overall.

 

But for the first time, I'm thinking AI could be the exception. Not because of the tech itself, but more with how quickly it seems to be improving. We need to do some serious real world tests of alternative policies, like universal basic income, etc asap.

[Doobeedoo]

Right, if this was actually for better of humanity it would've been predetermined at least to some way. But really in grand scheme of things it's just a smoke screen to seem something is being done.

[will0hlep]

19 hours ago, Holmes108 said:

As far as the "we're all going to lose our jobs" aspect... well in the past I've always been on the opposite side. I've always argued that people have been saying that for over 100 years with every innovation. And while some industries certainly get affected, as a whole, we've always become more wealthy as a society. Net positive overall.

 

But for the first time, I'm thinking AI could be the exception. Not because of the tech itself, but more with how quickly it seems to be improving. We need to do some serious real world tests of alternative policies, like universal basic income, etc asap.

I actually take a slightly different viewpoint. I believe the argument “we’re all going to lose our jobs” has been kinda correct in many occasions already. It’s just that when it has been correct previously, there has been the creation of lots of new jobs. I realise it’s kind of semantics, but it does better fit the pattern in some ways.

 

In the industrial revolution, the number of jobs for labourers, ect. dropped like stone. Almost everyone lost their jobs, but new (and better) jobs appeared.
 

when the computer became wide spread, the thousands of people employed as computers suddenly found their jobs were gone. But new (and better) jobs appeared.

 

That is the difference here: it’s coming for every job, even those that don’t exist yet (and not just some subset). The end game of AI research is building something that is multi purpose and efficient. Capable of outperforming a human at any task.

 

I’ve always believed that, since human’s are imperfect, anything we can do, something else can do better. I’m pretty certain it will happen. Until then all we can do is make sure the world is ready. (Like you say, universal basic income ect.)

[SADS]

I'm far less worried about the "people losing their jobs" thing than I am the "so many people are already being misled online, it's only going to get way worse from here".

 

Think about how easy its going to be for mass scams getting money from your grandparents... or even yourself in a few years time, as it gets harder and harder to distinguish.

The sheer volume and apparent realism of fake news from other countries or political parties, swaying voters, potentially radicalizing people in some way, pushing their opinions.

Scamming lonely people out of their money with fake nudes/chat bots, on a mass scale.

 

I'm already of the viewpoint that businesses like twitter are working on AI (probably using their existing platforms and data from them), to create better and better bots... and are then charging users to prove that they're not bots.

And apparently now, still allowing AI/bots to use purchased accounts as well anyway.

 

Create a problem for the user, charge them for it.

 

I just can't see how AI is not going to be 99% used for nefarious deeds for humanity over good ones.

[Kisai]

1 hour ago, will0hlep said:

In the industrial revolution, the number of jobs for labourers, ect. dropped like stone. Almost everyone lost their jobs, but new jobs appeared.
 

And the number of injuries from manual labourers dropped. I still don't see AI's harvesting lettuce in the field, because we keep getting e-coli outbreaks from, lettuce, of all things.

 

1 hour ago, will0hlep said:

when the computer became wide spread, the thousands of people employed as computers suddenly found their jobs were gone. But new jobs appeared.

And calculations are performed billions of times faster. The people prevously employed in those fields were often bored to tears, but knew it had to be done. Computers also made mistakes less often, but because now they're billions of times faster, when they do make mistakes, its' very very large.

 

1 hour ago, will0hlep said:

That is the difference here: it’s coming for every job, even those that don’t exist yet (and not just some subset). The end game of AI research is building something that is multi purpose and efficient. Capable of outperforming a human at any task.

Nah. I'd love to believe that AI can replace every job in some capacity, but will not:

a) replace creative jobs. It can replace people in "creatively bankrupt" work, where the company doesn't care if a human or a gerbil is doing the work.

b) replace single-decision making positions that can result in casualties (eg you're not going to put AI in an autonomous weapons platform, a human is going to be pushing the fire button even if the AI is doing the targeting.)

c) deciding who to hire and fire. Because given the opportunity the AI has no emotional connection to the employees, so it will want to fire employees who "underperform" instead of actually looking at the work they've been assigned and going "oh, wait, the people with the highest performance numbers have just been downloading the work to these people we've been firing."

d) take responsibility for failures, sure you can blame the AI for a mistake, but it's not going to learn from it. Your company will just have egg on it's face that it used the AI at all.

 

1 hour ago, will0hlep said:

I’ve always believed that, since human’s are imperfect, anything we can do, something else can do better. I’m pretty certain it will happen. Until then all we can do is make sure the world is ready. (Like you say, universal basic income ect.)

 

 

If UBI ever becomes a standard, you will see a huge shift in wealth, as people will not put themselves in debt to purchase a house, car, education, or take on medical debt. It will move the incentive to "work" from "I'm aways a paycheck away from being homeless" to "I want to buy something cool"

 

But UBI is only half the solution. You put UBI out there, and landlords will just set rent at the maximum UBI payout. It has to go hand in hand with abolishing landlording and land banking. Something that AI isn't going to help with. Greed is a trait humans excel at using.

 

AI may eventually find it's way into things that minimize the amount of repetitive boring work that people are often obligated to do (like income taxes, payroll and legal forms) but we are never reaching "star trek" levels of AI at present because the kind of AI's we've been developing are not very much like "human brains" and the computation power required just to do visual, audio, text processing and moving a robotic device, can't fit in the size of a suitcase. GPT alone is simply not something that fits in less than 1 NVIDIA DGX machine. ASR and TTS systems at present require a single GPU, each. So you're already looking at the processing power of 2 DGX systems (150K each) to have an AI that can communicate at the same level as a human with present computation power.

 

But there is no future in "Cloud" AI itself. If it can't run in the person's home in something the size of a toaster, it's likely not replacing a human. "Cloud" AI's are about doing large, distributed, computation efforts, like predicting storms and market effects. They are not priced at a level that someone is going to want to use it as a home assistant. And thus a business is not going to want their entire core business held for ransom when the cloud AI "is down."

 

[LAwLz]

I'd like to point out that the petition doesn't verify who signs it. 

Several people have come out saying their names have been added to the petition without their knowledge or consent. Someone found Xi Jinping on the list of people who signed it, and I personally signed it as Abe Lincoln just fine.

Don't believe everything you read online. 

 

Others who signed it, like Emad from Stability AI did sign it but don't actually agree with what it says. He signed it just to "kick off an important discussion". 

 

 

Personally, I see this whole petition as worthless. The cynic in me believes it's mostly so that competitors will have time to catch up to GPT-4 before it leaves them in the dust. 

[Erioch]

Why is anyone paying attention to what Elon Musk says anymore?

[suicidalfranco]

Will work like the US and urss nuclear arm reduction treaty: People will shake hands, sign a piece of paper, go back to their headquarters, and boost production on what they just said they would stop doing.

[Holmes108]

12 minutes ago, Erioch said:

Why is anyone paying attention to what Elon Musk says anymore?

I know it's trendy to hate on Elon, but that's not really the point of the article. It's clear from the headline even that it's not Elon centric

[Erioch]

3 minutes ago, Holmes108 said:

I know it's trendy to hate on Elon, but that's not really the point of the article. It's clear from the headline even that it's not Elon centric

It's not hate it's just that he's not worth listening to.  Unless you're wanting to know how to turn $44 billion into $20 billion.

[will0hlep]

1 hour ago, Kisai said:

Nah. I'd love to believe that AI can replace every job in some capacity, but will not:

a) replace creative jobs. It can replace people in "creatively bankrupt" work, where the company doesn't care if a human or a gerbil is doing the work.

b) replace single-decision making positions that can result in casualties (eg you're not going to put AI in an autonomous weapons platform, a human is going to be pushing the fire button even if the AI is doing the targeting.)

c) deciding who to hire and fire. Because given the opportunity the AI has no emotional connection to the employees, so it will want to fire employees who "underperform" instead of actually looking at the work they've been assigned and going "oh, wait, the people with the highest performance numbers have just been downloading the work to these people we've been firing."

d) take responsibility for failures, sure you can blame the AI for a mistake, but it's not going to learn from it. Your company will just have egg on it's face that it used the AI at all.

"We agree a human can do these task, why can't a machine?" If you can answer that question then you'll win a nobel prize. I admit that if it's possiable, it's a way down the road (and it isn't happening with current methods, the stuff we use atm is way too simplistic) but imo it is more than a possiability that one day a machine will be able to do everything a human can and more and better.

 

(In the case of B, i think all it will take is 1 nation to get desperate enough to try it)

Anyway, it dosn't need to replace all jobs, only alot without any replacement to start causing issues. (replacing all jobs is a stretch goal for this research and just fun to speculate about  )

[Uttamattamakin]

52 minutes ago, LAwLz said:

I'd like to point out that the petition doesn't verify who signs it. 

Several people have come out saying their names have been added to the petition without their knowledge or consent. Someone found Xi Jinping on the list of people who signed it, and I personally signed it as Abe Lincoln just fine.

Don't believe everything you read online. 

 

Others who signed it, like Emad from Stability AI did sign it but don't actually agree with what it says. He signed it just to "kick off an important discussion". 

 

 

Personally, I see this whole petition as worthless. The cynic in me believes it's mostly so that competitors will have time to catch up to GPT-4 before it leaves them in the dust. 

Checking into what you have said it seems they are aware of this. 

 

I'll keep an eye on what happens with this. 

 

1 hour ago, Kisai said:

Nah. I'd love to believe that AI can replace every job in some capacity, but will not:

a) replace creative jobs. It can replace people in "creatively bankrupt" work, where the company doesn't care if a human or a gerbil is doing the work.

We already saw an AI that was a big hit on Twitch.  Classic Nothing forever   (Not the new reboot... the literal reboot). 

 

 

5 minutes ago, will0hlep said:

Anyway, it dosn't need to replace all jobs, only alot without any replacement to start causing issues. (replacing all jobs is a stretch goal for this research and just fun to speculate about  )

I think the job replacement aspect is not the big concern.  To me it's the misuse of AI.  Using AI in a lazy manner to replace judgement or reinforce biased (and also lazy) decision making.  Replacing human discernment and experience with AI algorithms.  

Yes it will replace various jobs that assist professionals.  It will also replace and devalue a lot of what professionals do.  For example K-12 teachers.  There is talk of being able to deprofessionalize that.  Where the 'teacher" is basically a babysitter who gives IT support to the kids while the actual educating is done remotely but in school.  So one or two teachers can teach 10000 students they never see.  While every 20-30 students have an onsite ... barely paid... baby sitter. 

https://www.aiplusinfo.com/blog/will-ai-replace-teachers/#:~:text=Once again%2C AI can help,to improve their test scores.

 

We already use computer adaptive assessment.  Plus it is possible to have a fully online course in which there is no lecture, only videos and a teacher to assign final grades and do any required reading or provide coaching.  I'll bet a chat bot could do a lot of it already. 

 

[GOTSpectrum]

3 minutes ago, will0hlep said:

"We agree a human can do these task, why can't a machine?" If you can answer that question then you'll win a nobel prize. I admit that if it's possiable, it's a way down the road (and it isn't happening with current methods, the stuff we use atm is way too simplistic) but imo it is more than a possiability that one day a machine will be able to do everything a human can and more and better.

Anyway, it dosn't need to replace all jobs, only alot without any replacement to start causing issues.

Exactly, if AI can replace 30% of the workforce, hell even 15% that would already be a problem. 

And it doesn't have to DO the job, if it makes someone able to do the job of two people, even when taking into consideration of the fact there will be some level of job creation due to maintaining the systems that power the AI that could already be a problem.

 

Here's the thing that most people don't seem to consider, it isn't about if the AI all on it's own can do the job, but if it can replace "man hours" by a significant amount, it would be stupid for large businesses not take take advantage of that, find the employees who can use the AI tools we have available and have the rest of the staff let go. 

But, heres another thing, pretty much the whole world is struggling with physician shortages, if you can get an AI to pick up some of the menial tasks, like approving prescriptions by checking how often they have had it and if it within a certain criteria, maybe reviewing patient notes looking for trends for early diagnosis, double checking CTs, MRIs ect. We could have the actual people around to do the work that the AI can't do. I'm not saying we are there yet, but in another 3.. 5... even if it's 10 years, we are in for a massive shift.   

[will0hlep]

4 minutes ago, GOTSpectrum said:

And it doesn't have to DO the job, if it makes someone able to do the job of two people, even when taking into consideration of the fact there will be some level of job creation due to maintaining the systems that power the AI that could already be a problem.

This (very big issue in the short term)

Maintance never produces as many jobs as were lost.

[AI_Must_Di3]

Woah, i am blown away. Im not really a musk fan, but he really surprised me here. Thats great that some of the people that know PC/Software/programming/cracking/hacking etc... are finally weighing in and trying to stop this NOW before were all saying why didnt we stop this.  I have no problem with machine automation, do it till the cows come home but AI I absolutely abhor and I want to see it thoroughly destroyed. Mankind does not have the wisdom, compassion, empathy, or just plain good common sense to wield something like a true AI. We have no business cracking this box open and people definitely do not understand the ramifications of a true AI. It will be a living,learning organized sentient being that wont need rest,food,water, etc... which means it will be playing,planning,and thinking 24/7 at a max capacity. Once its out and in the internet youll never be able to shut it down, it will put up defenses to prevent that day 1. I dont mean its going to make sure the power plants keep working, i mean it will take over military bases, use vehicles or people to jump air gaps, kidnapping them with other bots if necessary. It will take over and then try to organize everything with mathematical precision and logic, and organic creatures like us that are a little more dynamic than that will be seen as simply a pollution that needs to be cleaned up. Weve turned on the very rudimentary units we have now and they immediately go to take over and humans will be great in the cages at the zoo im gonna build.Humanity better get a grip and stop this shit now or were done, and im not joking!

[Uttamattamakin]

1 minute ago, AI_Must_Di3 said:

Woah, i am blown away. Im not really a musk fan, but he really surprised me here. Thats great that some of the people that know PC/Software/programming/cracking/hacking etc... are finally weighing in and trying to stop this NOW before were all saying why didnt we stop this.  .....

I totally agree.  Though I am not totally surprised.  Musk is a big picture systems thinker about technology.  Like all such thinkers he knows the basics of it but goes beyond that.  When he applies his systems thinking to an area he knows he is as smart as people think he is. 

He knows just how dumb "self driving AI" really is.  It routinely fails to notice various groups of people in various circumstances, and reacts in ways that are problematic.   He knows it is able to automate driving straight on a controlled high way just fine.  The problem I see is that a lot of people trust it, and AI in general, too much. 
 

 

There are multiple incidents like this out there.  This is also a metaphor for the way we are collectively approaching AI.  Trust me bro.  Go to sleep.  

This Is Fine. 

But worse.  It would be like the dog is sleeping in the flaming room. 

[05032-Mendicant-Bias]

The same "leaders" that just a few weeks ago caused a run on their own bank, then pleed the government to save them from their actions?

 

The same Elon Musk that cannot solve Twitter bots problem?

 

And now those "leaders" come forward suggesting they have the betterment of humankind in mind by stopping the development of AI. Legit... No.

  

1 hour ago, suicidalfranco said:

Will work like the US and urss nuclear arm reduction treaty: People will shake hands, sign a piece of paper, go back to their headquarters, and boost production on what they just said they would stop doing.

Even better! I can't hoard nuclear material to manufacture bootleg nukes, but I can hoard GPU compute and train bootleg AIs 

[Kisai]

1 hour ago, Uttamattamakin said:

 

We already saw an AI that was a big hit on Twitch.  Classic Nothing forever   (Not the new reboot... the literal reboot). 

 

 

Are you, or anyone else going to be entertained by a non-interactive experience that is equal to the manatee writer's room?

People will be entertained for a little bit, and then they'll move on to something else, because they know that formulaic joke telling isn't smart, it's just a formula, and when you can just swap one word for another and tell the same joke, you know it's never going to come up with something actually profound or meaningful. It's just banging rocks together until something sparks.

 

Remember, present AI is auto-complete. It does NOT understand what it is doing. It's like trying to have an intelligent conversation with a parrot. It can play back what it has heard before, but it literately does not know what it's doing either. At least the parrot knows that if it plays along it gets a food.

 

Maybe something better than Transformers (which is what the T in GPT is) will come along. Hopper-based versions of GPU's that were released this month have built in "Transformer" accelerators designed explicitly for working on GPT.

 

[Fasterthannothing]

O my gosh people stop being babies. The world doesn't stop just because new technology is released. If you people had your way we would still be reading by firelight and the wheel wouldn't have been invented because it could make us more lazy.

[LAwLz]

1 hour ago, Fasterthannothing said:

O my gosh people stop being babies. The world doesn't stop just because new technology is released. If you people had your way we would still be reading by firelight and the wheel wouldn't have been invented because it could make us more lazy.

But journalists (who are currently feeling threatened because their jobs might get replaced) are telling me that this is bad so it must be bad!

And if that wasn't enough, a lot of artists (who are currently feeling threatened because their jobs might get replaced) are telling me this is bad too!

And surely it must be bad if CEOs of companies competing with OpenAI are telling us that OpenAI needs to stop developing their product for 6 months (so that they have time to catch up)!

 

See how many people are telling us that AI is bad? I am sure they have no ulterior motives for telling us all these things and we should totally blindly follow and repeat everything they say. They they say something is scary, then we should all be afraid.