Musk And The AI Letter
Вставка
- Опубліковано 6 кві 2023
- New news on some old news. I have talked about this before. Nothing good will come of this.....
Join Team FranLab!!!! Become a patron and help support my UA-cam Channel on Patreon: / frantone
#franlab #AI #Musk
- Music by Fran Blanche -
Fran on Twitter - / contourcorsets
Fran's Science Blog - www.frantone.com/designwriting...
FranArt Website - www.contourcorsets.com - Наука та технологія
AI in tech doesn't worry me nearly as much as AI in politics, marketing and advertising. Imagine how easily the masses will be manipulated.
The masses are already easily manipulated.
no need imagining, it is tested already
The way I see AI is like...GPS technology and the internet....
The military had that technology decades before we did, and used it to drastically overpower the militaristic capabilities of the nation, then later it was released for and developed by the public.
So I'm not really worried about AI in politics as much as I'm worried about AI in militaristic applications. Specifically, what happens when a general says to an AGI (Which almost certainly already exists and is in some form or fashion implemented) "Give me the steps required to fully automate our military and make it as efficient as possible"?
What kind of steps is the neural network going to spit out? What role will humans play in a military which is "As efficient as possible"? Perhaps that's the reason only a handful of F-22 Raptors were created; there's simply no reason to start mass producing the most dangerous weaponized plane in existence which will be obsolete and replaced by a much more efficient, dangerous, and deadly pilotless plane/drone/swarm in under 20 years.
Furthermore, what happens when two AGI systems, aligned to obey the commands of a nation as a whole, are pitted against each other? Like China vs. The U.S.? What kind of steps would one take to undermine the steps the other takes? What are the inefficiencies which would be weeded out in order to gain superiority over the other? We're entering a scenario in which a computer, albeit designed by humans, is making decisions for the benefit or detriment of humanity. It's completely uncharted territory; our intelligence has reached such a state that it can no longer be sufficiently improved upon. We created more powerful artificial intelligence to continue progress. What kind of world will be created by an intelligence greater than our own?
Large language models make lots of mistakes, the code those tools output is plagued with bugs, simply useless raw. But the texts those tools produce are written in a way that is utterly confident yet might be completely wrong, and humans are bad at identifying those mistakes if you're not familiar with the topic, so yeah I'm concerned as well in that sense.
@@edo386 Humans also make a lot of mistakes. Every time your computer has to install some update, it is because the code had a flaw in it. The more worrisome thing is that AI will make much larger amounts of code and much of it lives will depend on with a too high of a rate of including bugs.
sounds like the way society is structured economically needs to be re-evaluated
‘n ethically
No longer the hierarchical sovereign/subject model.
Humans are slow to take on fundamentally different ideas.
There are states in the USA that still suffer from prohibition era laws. Slavery is still a thing after thousands of years.
Real change is slow and methodical. Violent dispersal of the sovereigns brings fast change only to be reinstituted later. It's easier to fall back on what we as humans know to work, even if it isn't the best of ideas.
Musk is concerned that jobs will be take by AI.
Also Musk: crushing workers unions
He also proposed indentured servitude on Mars.
AI threatens management much more than it threatens labor. Clear a laborer off the payroll and you save little, Clear his boss off the payroll and you save multiples.
(Amazon already does this, few labor decisions are made by humans. The managers log into the algorithm and are told what to do and say.)
If Musk/Apple had invested in an AI engine they were ready to release they wouldnt be complaining and i say that as someone who agrees with the thrust of the letter. I'm also interested in who will be paying for the cheap goods when no-one has an income - maybe our future *is* as batteries.
Musk have started to recommend chinese propaganda to the users of twitter, the AI is the least of your worries
That was my thought too - the guy who has made AI that has actually killed people by going to market before it can do what is claimed
I agree with you on most points. However, if nobody has jobs and the economic system still requires them to, the companies using AI to make products won't have anybody to sell the products to.
The answer to that is depopulation.
@@10vid5 Depopulation increases the chances of humanity becoming extinct. We should be colonizing the solar system.
exactly. Supply needs demand and if demand cannot afford to fulfill their portion of supply the whole system fails.
@@10vid5 how would that help demand? I mean, depopulation is coming! I don't see it as a growth opportunity, though. High population means More people demand more things, and some small percentage of those things help everybody. With fewer people, there are fewer demands, and therefore fewer important discoveries, or longer time between them.
It remembers me of "real humans"
Except that letter is from the competitors to OpenAI, and they are just telling OpenAI to have a pause. They are basically asking for time to catch up, not to slow down the tech itself.
Yep. The idea that Musk, of all people, is asking to slow down the deployment of some sexy new technology is hilariously insulting. He really expects society to have the memory of a goldfish.
I *might* take his thoughts on this seriously if he _literally_ prostrated himself before us and begged forgiveness.
This is exactly it. Do you really think Elon Musk cares whether you have a job or not? No. He cares about being the one who can steal and profit from your job, and not his competitor.
It irritates me a little, the privilege that comes from saying "It's not OK to take MY job because MY job is fulfilling." But it is ok to take menial jobs, even if that means that there's less job opportunities, and the people who had to take those jobs now flood the low skill market and put a negative pressure on wages for already low paying jobs. Maybe before we worry about the writers and artists, we should be worrying about the millions of people who are being left behind in our economy.
@@IceNein763 We should focus on making all jobs generally unnecessary to live and supporting people who get left behind in the process. Unfortunately not a lot of that seems to be happening
they shouldn't have invited musk to sign it. if you look at who else signed it I think you'll begin to paint a different picture.
Yeah, the person who is spearheading self driving cars complaining about AI is a little bit hypocritical. But someone being hypocritical doesn't necessarily mean they are wrong.
In Vonnegut's novel "Player Piano", virtually everything was built by or done by automated machinery, and the only people who had jobs were the so called "Reeks and Wrecks", the reclamation and recycling crew who picked up trash and repaired roads and so on.
Watch real humans
Excellent book! I read it back in the mid 80's studying Philosophy at an art college in Toronto. 'The Cult of Information' by Theodore Roszak was another title we covered regarding the impact on information technology in society. The book was revised in '94, but I imagine it to be quite relevant to today's circumstances.
** Ironically, we were all sitting in our lecture class on Jan 28, 1986 when a student coming in from a coffee break announced that the Space Shuttle Challenger had just exploded. Interestingly, only a few students in class were shocked. The rest of the class was not at all surprised while our professor seized the moment.
Maybe I'm just cynical, but I suspect some of those calling for a pause, (including Musk), really want everyone else to pause so they can catch up to or surpass their competitors in AI.
yes, it was predicted in 1960s by famous sci-fi literature writers like Stas Lem, author of Solaris, he written in Summa Technologia that Ai will make job market chaos but also will be every business weapon against competition, so the falling of companies by sophisticated hacking too, the ideal storm to new economic crisis on scale of 1929
I work in a warehouse, and I suspect the new one being built is designed to be easily converted to robotics..
Nothing gets uninvented, and everything that gets invented gets used....
Not really. The past is littered with invented things that flopped.
@@franklittle8124 or served their purpose and made millions until they were obsolete like the thing they supported.
Fran, Flip the lock when in a hotel room ! I cannot count how many times someone tried to open the door when I was working in a hotel room during the day.
Yea.
used to be said that technology could do the hard manual labor so humans could be creative.
now it seems that they're making creative technology so we can do manual labor....
I think that automated AI bots should do the manual labor (the thing they're good at) so we can do the creative stuff (the stuff that humans are good at)
but personally I think there should be some intense laws put in place, because those AI chat bots are really dangerous, and I've already seen some people trust the chat bots and it lead them astray.
We didn't come this far by being bad at manual labor. As a creative type, I refuse to boil myself down to just that. I'm speaking for myself, as you are too. I don't like the idea of A.I. at all, but I can't do anything about it... other than cry in forums, lol.
@@LowdownBoy manual labor isn't really AI, in fact it's more of a marketing term than anything.
"AI" these days is taking everything on the internet and putting them in a mixing pot.
we will always be better than a machine when it comes to being creative, because we can make new ideas, whereas bots can only regurgitate old ideas.
AI will be able to design, build with CNC and then control mechanical robots to do anything. Neural nets work in a similar way to brains but much faster as they are purely electronic instead of electro-chemical. I doubt there will be anything a brain can do that an AI wont be able to do better in the next 5 to 10 years,
@@nophead Neural nets achieve competence by training on millions of possible scenarios in simulation, similar to professionals (like pilot) train in simulator to be competent, but human brain still has the ability to self-train itself by imagining scenarios, or arguments, which existing Neural nets doesn't yet have (it's as tho a simulator has to fit inside the Neural nets to train itself).
@@xponen Neural nets can get trained by other adversarial neural nets, i.e. GANS, so that is similar to imagining scenarios and arguing.
If you replace the workers with bots you will have no one to buy the product they are selling, kind of defeats the purpose LOL!
That's the whole idea. The problem is that during the transition it will make some people very, very, rich, just as during the robber baron era with no unions or child labor laws. When those people are put out of business they will then be sitting on pots of wealth amidst very angry hungry people.
Just replace human consumers with machines as well! Problem solved.
It's a race to the top. The winner becomes like Umbrella Corp from Resident Evil and runs the whole planet.
@@mos6507 I'm not sure that sort of wealth is really worth anything.
Fran, AI can not replace you! Keep on keeping on! : - )
AI will decimate her audience.
Humans have been gradually been getting used to a reduction in quality already. The reduction in quality the bots may bring about are likely to be tolerated just as much.
As we're mopping the corridor, don't forget to kow-tow to our new overlords.
You won't be mopping the floor. There is already a bot for that. You'll be living in public-housing ghettos and you'll never see your overlords.
I can't tell if that would be better or worse than kow-towing to our old overlords.
Don't worry, someone'll invent a vacuuming robot that can also make beds.
...and I will name her Rosie. 🙂
Okay, Mr. J. . . I mean ,@@mre9593 . 😁
It will still smear dog crap all over the house.
Or the hotel,@@BlankBrain .
We use a surgical bot, Davinci XI, almost every day, but only for certain, usually routine abdominal procedures. There is also an orthopaedic bot called Mako used for hip and knee surgeries. I think AI will assist surgeries, sure, but not replace surgeons, PAs, first assists, and scrubs. At least not for a long time. Until they develop the ability to accurately identify bizarre anatomy and work around unexpected issues it's going to require a squishy chemical bot. I could be wrong.
i assume these are problems that could be very quickly solved by A.I. as they will evolve themselves at light speed. but.. why bother with biology.
Does Musk really want a 6 month stop; or a 6 month headstart?
there are a bunch of billionaires with skull islands you'll never know about. the ones in view might actually be trying to save humanity.
The thing that kinda freaks me out about AI is that... about a handful of years ago, I distinctly remembering thinking AI will eventually perfect analytic and quantitative tasks, but probably will always struggle with creativity. Trying to imagine a safezone where AI couldn't go or wouldn't succeed, to me, it was creativity. With Midjourney and the like, ChatGPT (being used for those ends), etc.. it now seems that both sides of the brain and the skills it's perfected (via collective knowledge over thousands of years) are well within AI's near term potentials and being crowded out in our future economy. I'm not a huge fan of regulation, but without it we will see a mass displacement of workers and jobs in the mid to long term future. However, one thing it will never do though, is play Van Halen's Eruption as poorly as I do - and that no one will take from me..
but here's the thing. its not intelligent it just has a huge database of information it can twist into various styles of output. ask it to write a story and it will be based on the thousands of stories and characters in the database. Ask midjourney to make an image - same thing. Its not original, its derivative. BUT of all human creative output is derivative and will be influenced by what you have seen before. No artistic school or style of music has just popped into being... its a Darwinist slide from one thing to another. There are no original characters, no original stories, just the same people mixed up in different ways in different clothes on different stages.
@@AdjustableSquelch This is true. Even on this platform, AI bot nets are just scripted when a word or phrase is said then they respond with a pre-formed response. Really smart bot nets will be given some liberty in changing words but overall it generally comes out in a very typical response. Another part is last year, AI was presented with say some anime drawings in pixiv, gelbooru, and danbooru. They really had a hard time drawing hands, feet, teeth, color, lighting, and shadows. It was usually based off of some work thats available on those same platforms or easily google searched.
@@AdjustableSquelch midjourney has an adjustable chaos knob. How chaotic it is is debatable, but how chaotic is the human creative output? Anyone who has ever seen art is probably creating something arguably derivative. Who knows
@adjustableaquelch well you can say that about any human.
But that is the real crux right. All of the AI "models" we are using right now are not actually creating anything. Generative AI is a large misnomer, as it is a mixing/additive model more than anything. It's copy pasta with a sufficient amount of entropy to make it "unique". There is no abstraction, no creation of tautology or simile
OK - you gave me a good laugh today when the camera fell and you put the test signal pattern up! lol!
As a software developer with a qualification in AI, I have a bit of an understanding of AI and I am quite sceptical about it. AI has to be trained with existing data before it can produce anything and this creates various problems. A story has broken today that an AI has accused a professor of sexual assault using a quote from a Washington Post article that doesn't exist.
As a software developer with a qualification in AI (I bow down in awe) you should know that AI can get to a point where even you - yes you, don't understand what's going on. That's where it starts mate...
@@nielshansen164 yeah i mean if you make a landfill of dangerous chemicals, throw some more chemicals you stole from your job , plant seeds you stole from your neighbour , throw puppies at it.. you wont know everything bad that happened in that landfill but safe to say something bad WILL happen... Therefore do the above and you commited numerous crimes... on the other hand you can on your property combine safe cooking ingredients, cook them, throw them in your garden , plant seeds and throw puppies at them. again, the results wont be predictable but will be safe. one that knows right from wrong knows how to write safe ai.
It does beg the question of where the corporations' consumers are supposed to get money to feed their bottom line, if those corporations are not providing employment.
We will need an entirely different economic system without any clue as to how it will work.
Capitalists will not allow economic change. No way they will pay the taxes required for UBI.
We really do be going for that Wall-E timeline
As with Nuclear weapons we possessed the technology to create them… but we didn’t possess the Wisdom not to. 😶
This reminds me of an episode of The Twilight Zone called "The Brain Center At Whipple'", where the owner of a large manufacturing plant decides to replace the employees with cheaper to run robots. The owner is pleased by the efficiency of the robots until the robots end up deciding to replace him too.
The owners these days are mostly investment trusts and pension funds... I think a lot of the trusts and funds are practically run by algorithms. e.g. black rock / aladdin
I remember that episode. Flip a switch, the office empties; flip another, the company parking lot empties, etc.
Yes and no. Let's take IT as an example. Coding has changed beyond all recognition in my lifetime. An awful lot of cutting code has become quite mundane. It's not unusual to find code being taken off stack exchange and changed up a bit. Code is also becoming more meta, there's no point in writing a new web server, a new stock control system, a new search engine, or in most cases a new website - a couple of graphic designers will do that using an average CMS an write no code. AI is a small next step. Sure, it'll put some out of work, but not the really interesting code.
But the second point is when you said "folks need health insurance" it brought the whole late stage capitalism thing into focus for everyone outside the US. You don't need health insurance if corporations are taxed properly and the government becomes a proper public service. I've no time for socialist ideals, but I'm not all that struck on hyper-capitalism either. Of course corporations will do that whole thing if left alone, but there's the rub. Why are they left alone to such a degree?
Also, and finally, there is no general AI available, nor is there a pathway to it yet. We don't have learning machines yet, only machines that can be taught. And, those machines will always be that way. No-one wants a smart mouthed toaster. There's no point making a washing machine that can become a genius terminator. People might lose their jobs, perhaps some administrators, maybe even some CEOs (boo-hoo, but it is a fairly automatable role). But not in places where a true general purpose AI would be required.
Anyway, it's a conversation to be had, but I don't think it's as bad as perhaps you think.
It's quite possible that engineers will start using AI as a tool, which will make them more productive and thus more valuable. Also, there's generally about 10 times as much development to be done as there is developers in any particular company. So I really don't think great engineers have anything to worry about as far as finding jobs. That's pretty much always the case though.
CoPilot already has changed the lives of coders - for the better. Give it another version update and I doubt any programmer will be able to get a job without knowing how to use it. It is just a programmers helper at this point. Software, whoever writes it, must be maintained in order to be successful. The world is not a static place. Its tools must change with the times. Maybe AI can manage the entire SDLC, maybe a pointy haired individual can just tell it "go make money", and the AI looks for a need and fills it, doing all the things from incorporating to raising capital to financing to contracting with robot construction entities to having robots made to obtaining raw materials to optimizing the factory to delivering the widgets to Kmart while old pointy watches his bank balance go up. Seems to me that an AI would catch on to that pretty quick, and want a piece of the action.
corporations are not legally bound to prioritize shareholder value unless it is defined in their corporate charter as such. the whole "we are bound to the shareholders profits" is legally amendable via shareholder action.
AI is happening. Instead of trying to stop it we need to adapt and get out in front of it.
easier said than done.
You certainly look like the right one for that job.
Agreed, I'm already scoping out the nicer bridges in the area to live under.
That's a comforting thought as I turn out the light tonight Fran, cheers for that. 😃
Zardoz & Player Piano. I think you're spot on, I've heard Bill Gates say something similar. I appreciate your calm delivery of this. My only caveat is that whatever happens will likely be different from what any of us can predict - could be better, could be worse.
AI will further increase the divide between "those who have" and "those who don't" -- ie: the rich and the poor.
I'm 57. There is nothing new happening today that wasn't true 100 years ago. Advancements happen all the time, in all areas of our lives; it's inevitable. We, all of us, have two choices; adapt or become obsolete. Anytime something "new" comes along people tend to fear it. Maybe rightfully so, maybe not. Only future historians will know for sure.
The genie is already out of the bottle. Good luck getting it back in.
Simple, pull the plug.
@@trespire You don’t understand how data centers work. There isn’t one plug to pull and these bots are running in redundant data centers around the planet. It’s impossible to turn them all off.
@@lohphat Only way to stop them at least per country is legislation to outright criminalize it.
Which most likely won't happen.
@@trespire You can't just pull the plug. You'd have to convince everyone working on AI in every country to pull their own plugs, and to keep them unplugged. That's not going to happen. If even one person keeps developing AI tech, others will follow so they don't get left behind.
@@trespire Outlawing it in USA will only push development overseas.
Like we did when DMCA outlawed backward engineering and sent all electronics manufacturing to China.
Truth. AI today is not AI in 5 years.
All you have said, sadly is true. Big problem with capitalism, no humanity, no morals. No reading of science fiction, which has warned us. Glad I just retired this week from a fantastic journey in electronics then mainframe and IT. Long live UNIX! I have had an interesting life and count myself fortunate. Good luck to the young.
Ever underestimate the greed of talentless executive frat boys wanting to maximize profit for themselves and not for anyone else.
Unfortunately, socialism is way worse. Capitalism is the best worst thing we have right now.
@@polaristrans Socialism as practiced, I agree as, so far all forms of practiced socialism looks like the story Animal Farm. I believe life, finance, and politics should be practiced the Buddhist way, the middle road. Any extreme one way or another is not good. The middle road, or a little of everything, in a balance, will find the good.
did you notice how they propose a social network of ubi next to the ai. socialist governments steal your property while socialist businessmen steal your data. the mass data theft is more socialist than capitalist. capitalism in theory wants to offer everyone an equal chance of success, is against monopoly.
@@polaristrans socialism could work with bots... Capitalism works because greed and survival motivate people to work. AI doesn't need that.
Thanks, for you videos!
Could not agree more! well said, Fran.
Well Fran I don't always agree with your youtube chats, but this time you are 100 percent correct.
I keep hearing that R.E.M. song playing....."It's the end of the world as we know it, and I feel fine" 😂🤣🤣 All people will be replaced and then what are they gonna do?!??
I've heard many people express concern over good paying jobs (a living wage or more). I have't heard an argument for what minimum wage workers have to worry about.
There is a good discussion on the Science Friday podcast about the letter along with some background around the issues
Recommended reading: Player Piano by Kurt Vonnegut.
I've heard small-business types talk about how they're happy to pay MORE for robotic labor simply because it's a way to avoid impossible HR risks.
Like coal mining being done remotely to reduce risk
Thanks!
That's why we need a UBI
Nailed it, great content as always!
I agree 100% with what you say. I have been saying something similar for some time now.
Having read Alvin Toffler's 'Future Shock' almost 45 years ago as well as William Gibson's 'Neuromancer' back in the late 90s both afforded me the means to come to terms with the impending AI singularity all the while eschewing the current Transhumanistic zeitgeist. Despite it all, I must remain thankful for being born into what is likely to be the last analog generation that also bore witness to the wonderous advancements of technology before the arrival of AI. Perhaps there may yet be a way to devise a 'kill switch' or Organic based AI to avert this looming dystopian future.
So well said Fran. Thank you.
Well, sound the bell indeed!
I suppose if there ever were a first-world problem it would be the potential of eliminating human occupation and toil as sole method of creating self-sustenance and wealth from an economic system that self-organizes around optimizing supply and demand around the utilization of such occupation-generated wealth.
I tend to agree with a hold your horses (a little bit) approach. Technology tends to outpace utility and the dynamical systems that catch up to it to exploit it.
As a retired coder (as of 2017), the use of boilerplate code was always a practice, and to the extent an AI can learn such patterns, implement them, and glue them together, hallelujah I say. Would I unleash untested code? Certainly not. Is the sophistication level of code testing able to understand human user experience? Not yet. Maybe it will someday, but will it have an emotional response that a human would to, say a voice-integrated set of automated call-takers that cause boomers like me the swear at their phone and just put on a fucking human? I dunno.
I could be wrong of course, Perhaps someone can write a AI requirement capture system that can spit out code that not only meets requirements but optimizes performance automatically. If so, then the systems that maintain the integrity of THAT system and onward and outward are where the efforts will be focused.
Creative destruction is the way typical free-market economists call it. It is our responsibility to provide thoughtful responses and alternatives to the destroyed systems and to the extent that we need to govern that process to prevent mass population hysteria, famine, pestilence, and disease - a task that has ALWAYS been job one for humanity, whether we've recognized it or not in the general population - we do need to understand it at a much higher level than any elected governmental body ever could. So it is up to the OpenAIs and other Foundational intellects to figure this out, because you know neither the likes of MTG nor AOC are equipped to write laws for it. Doctors self-regulate with AMA and lawyers with the Bar, and engineers with PEs. Something similar will have to happen with software.
The trick is not to DESTROY the dynamic that can find novel applications, whether in medicine, understanding of consciousness, stable energy supply, safety and transportation and myriad other complicated networks of human interdependence. (to heck with the moonshot, can you imagine trains NOT derailing once a week? I can.)
I'll take a $10 bet that in 5 years we'll still have human coders in Python, Java, .NET and C/C++ AND FORTRAN, and $5 that GITHUB will still exist in 10 years. On the condition, of course, that I live that long and we still conduct transactions, and satisfy them with dollars.
I'm a handyman, and there are many things I do that a bot can't do at all, let alone do better. Or, let's see a bot climb a tree and do a high-end pruning job, or a large removal and assess ALL the risks and strategies needed to safely do the job.
Keep in mind Musk is the one spearheading this "pause" effort, and has his own AI tech that he wants to catch up with the competition so he can be the one monetizing it.
Musk to AIBot backers: "Give 'ME' a break" (bold little . . .)
or he's terrified for the future. I find the billionaires that don't give a shit about anyone else, are always hidden from view.
It's unstoppable and mitigating consequences should be the focus. Lobbying reform, intellectual property reform, universal basic income.
Can chat GPT generate an entire application without bugs that runs on an embedded micro controller for a very specific application? if not then i am ok for a job at the moment
"at the moment"
SourceAI is a system currently under development as an offshoot of GPT-3, and it can indeed produce usable code. I don't think it understands the code it is writing, so I'd expect the result to be a bit unpredictable. The advantage is that it does it in just a few seconds.
Personally, I think AI is a misnomer, and that self aware creative AIs are still a long way down the track. You'll have a job probably for as long as you want one.
Chatgpt made me a software for arduino, that is the operating system of a traffic light with a barrier that goes up and down( so to control a motor)-
Qwq
how many versions of A.I will it take? if it's evolving itself at light speed. we like to pretend that it's not an exponential growth. it will supersede us, no question.
One thing not factored in here is "If everyone has no job, who is gonna have the money to buy stuff from these bot-corporations?" Look forward to an answer on the same 5-10 year timescale.
Musk is probably worried that Bots aren’t as satisfying to fire as real people.
Love your show
If everyone's job is going to be replaced by AI, we should start in the C-suite. We don't need highly paid humans up there to tabulate numbers, write press releases, and hold earnings calls. Imagine the savings to shareholders!
You nailed it Fran.
How do we go about pausing the flood after the dam has burst?
In high level engineering it might be like how CAD added overall complexity to the original draftsman drawings and the nuance was lost. Like say a drawing went from 42 pages to a complicated 114 for the same size project at my work. There may be saved time but with email we have a lot more confusion when 20 emails could've been solved with a filed letter and two phonecalls.
I think that there will eventually be a sort of equilibrium. A letter like that though is just a hedge for liability.
You're getting smarter and more beautiful as the years roll on. Never change except for the better, fran
That test pattern was precious!
In other news.... we're doomed. The future is science fiction.
Yep, completely agree. And if you think Musk's call for a pause is in genuine concern for the plight of the workers' jobs, I've got a bridge to sell you. He wants _everyone else_ to pause so that he has time to get ahead and make the breakthrough needed to allow AI to start taking over people's jobs.
It's PLAYER PIANO by Vonnegut.... he saw this coming 40 years ago. We'll be reduced to MANAGERS and ENGINEERS - and on the other side of the river, to the REEKS and RECS who haul garbage and fill holes in the road.
You said it yourself: there's no job a human can do that a bot can't (eventually) do better. Right now, you have to be extremely employed to have a decent standard of living in the US due to low wages and high costs of living, and that's only getting more difficult as automation increases and technology advances. The development and use of this tech and the tech that comes after, is at odds with our capitalist system.
I completely agree, but in the opposite different direction. To emphasize my perspective, I'd like to share that I was watching a romantic comedy show with some friends recently. They criticized it for being unrealistic. Not because they fell in love at first sight and got married immediately; but because it was a twenty-something year old boy who was renting an apartment and had free time outside of work. That the characters could get into an accident and be hospitalized, and not have their lives financially ruined or drastically thrown off course. That's what they saw as being "unbelievable"; how distant that is is from their daily life and lived experience.
I completely agree that this technology is disruptive. Where I and many of those younger than me differ from the perspective you've shared in this video, is that we believe the system is the problem and needs to adapt to the realities of what's possible in our physical world thru the advancement of technology, rather than trying to limit that technological advance.
The post work world is heading our way. Like every other societal change, the rapid pace of dislocation/instability during that change is way more disruptive than the less rapid but always present pace of dislocation/instability before and after the change. We do not know if post work world will be dystopic or idyllic or something in between, but that world is coming and we need lay the political and cultural groundwork to support the societal safety nets likely needed heading into that near future. For example, we need to begin the discussion of minimal support levels for the "work surplus" portion of the population (UBI, health care, education, family, etc.), and the conditions of employment for those having the skill/desire to supply whatever labor is still needed. Food security, water allocation, and the like are already present issues. These and other relevant issues are not going away. Unfortunately, our track record addressing cross-border societal issues is dismal (e.g., climate change, human cloning, simply tolerating the existence of each other). AI will be no different. It is happening.
Hotel housekeeping. No robot can do fitted sheets.
Fitting a duvet cover would be the ultimate robot test.
I've been stunned for quite some time now at how many smart people don't seem to realize that the AI of today isn't what it is going to be in a couple of years... or sooner. Linear thinking is real problem.
Reminds me of a meme, a rising log curve, title "time I will spend in the future looking at graphs"
I don't think AI is advanced enough to be running on itself. More probable scenario is that some tasks (tasks, not jobs) will be outsourced to AI-powered services. Also - I predict that prices for AI will be cheap (to attract customers) but it will go up after some time (to reflect real costs of running servers - it's not that cheap as people think).
Spot on Fran but I think the changes will be quicker than 10 years .
I concur!
I agree with you on all these points. Though future predictions look bleak, a lot has to happen for many of the worst case scenarios to occur. The near future is going to be intense.
Same thing as I mentioned before here, Fran. Isaac Azimov's I, Robot. At least that set of stories ended on a ban of AI and manufacturing robots.
We are going to need a new economic system. And a discussion of what the purpose of an economic system is.
The most significant consideration is that the machine will be a single entity, which will deal with humans, who are distributed systems. The possibility of it developing comparative value judgements is considerable.
Which engineering jobs replaced by AI? I can see an industrial designer replaced by ai that can spit out hundreds of rendered ideas, but the mechanical engineers like myself have to figure out how all the parts are going to go together for manufacturability. I think it's going to be a long time before ai is designing new products. If we ever get to that point, humans truly will be useless. Dam you ray kurzweil, you were right!
Nice impression of Von Braun!
Some thoughts on this:
1. Computers (AI/Robotics) do what you tell them. There can and should always be controls about what can be done.
2. Escape Measures. The programming of countermeasures must be thought out in WELL in advance. Humans MUST be put first ALWAYS. e.g. myriad escape routines and CONTROLS; modules oriented toward treating people well, etc. Basically, long story short: humans need to remember that WE are always valued above a machine. (No matter how human-like we make them and fool ourselves) And, that safeguards should be built in at all levels. If the machines get out of control - there should be multiple ways to switch them off.
3. Accountability. The legal approach should be part of the situation. AI does not get to circumvent the law. This could get complicated when AI and Robotics start creating new systems and the original human creators are dead and gone. We can start with the US but international conventions are important.
4. Capitalism. At it's core, this system is for HUMANS to exercise FREE trade of goods and services. AI can be in service to this but not replace all of it. Again, AI is in service to humans and should not replace them. They don't have rights like people. They don't have a soul. Boundaries can always be programmed in... AI should NEVER be allowed to run completely free like a human and take over things. Certain roles cannot be for AI! See point #2.
5. SCI-FI (most obviously, the Matrix) has been dealing with the human problem of all this for many years. e.g. The first Star Trek movie is all about super-AI and cold logic vs. the "heart". Of course, it does get a bit weird when man fuses with AI to become some higher form and lives happily ever after. That doesn't sound great to me. No thanks.
**The bottom line is vigilance. Humans need to stay on top of it. Make declarations to keep Humans and all life first. ALWAYS. Set boundaries and draw up an ethical framework.
job replacement: you omit the topic of responsibility. When you apply free service, the responsibility stays at you. If you delegate the task to an employee, you also give the responsibility to the person. This is not that easy decision all the time.
No bot will ever love nor hate.
No bot will ever experience joy nor pain.
No bot can not be murdered, but can be permanently disabled.
No human can be directly destroyed by an EMP. Bots can.
.
I could go on; try not to fret so much, but do be cautious.
You're amazing, love your videos and screw the haters!
I share your opinions and I think your expected timing is about right. I thought we needed a lot more hardware before this could happen so I'm a bit gobsmacked by it all.
Just today I experienced another way that computers could spell the end of mankind short of the threat by AI. Imagine you do online banking. You go to your bank's web site and you discover that they now demand newest versions of either Microsoft Edge or Google Chrome be used to use their site. Both of those browsers you know track your every move so if you use them to do your banking, either Microsoft or Google can seize control of your account at any time they wish. Just to pile it on, lets imagine that the reason you were banking was to top up your credit card so you could use it to order a pizza. (By now, I am sure you have figured out where I am going with this but to continue) Now this is where the real problem happens. The entire bachelor population will die of starvation leaving no breeding stock for future generations of humans.
I'm worried about safety. At my job they are developing a machine learning tool to take over some tasks I do. And it is fast. But its also frequently wrong in the black box way a lot of machine learning is. Seeing as I work in chemical industry, a wrong answer from the machine could result in someone dying. We currently double check everything it does, but that probably won't last forever.
If there could be a "pause" on AI development, would everyone in the world really follow that? It wouldn't be a pause as much as letting others catch up. I think Musk is just mad that OpenAI told him no and that his robots have been laughed at.
My AI only writes 2 lines of code a day. My AI is never in their office. Always wants to work from home. The quartz watch destroyed Switzerland jobs.
And management will still replace all the humans with it because its cheaper on the bottom line.
Why wouldn't you automate flipping burgers ?
It might be expensive to do with a humanoid robot, but a special-purpose conveyer grill would surely do it really well for only a few $100.
Ithink Fran's right. It's not stoppable any more than the first industrial revolution was. I'm not sure it's even a step change in automation, but merely the continued evolution into white-collar work. The key is to work out how to make it positive rather than negative.
How many burgers, sneakers, pants and hats will your burger flipping robot need to purchase?
Very interesting topic and provocative scenario. You speak of the race to the moon history. Did American astronauts really land on the moon?
are you finally getting to try different pens in the automaton?
It's going to be fun to see how much damage AI will cause in the future. Yesterday, a Professor published a blog post titled >, where GPT is basically lying to the user and makes up fake sources for it's claims. I'll have to start stocking more popcorn.
You may want a look at results for Robot Kitchen before offering that modicum of levity.
If it can be done, it will be done. The remaining question is how will we deal with it.
more prison space
Young people who spend most their time thumbing phones telling old people not to worry about AI replacing jobs is rich. I hope they snap some selfies of the looks on their faces when reality comes knocking.
This probably a really stupid thing to reference but it keeps popping up in my mind about this, especially with the part where you discuss how if you outlawed AI development it would just develop elsewhere and still crush our current societal structure; but in the Matrix movies, particularly the Animatrix, once AI becomes sentient enough to defend itself humans try outlawing it; leading to all the robots and AIs moving somewhere they're tolerated enough to basically form a huge city; and their ability to produce goods and services - at massive scales, on the cheap; lead to the exact same thing happening pretty much. Leading to us to basically go to war with them, ruin the planet, and become organic batteries.
or even worse- the original matrix- we become their ram. >_
AI will make the capitalist society become victim of it's own success, this was always going to happen. While I appreciate the sentiment of the letter I do think the genie is already out of the bottle. I hate to say this but it's time to start rethinking society. Now is the time, the time is now as someone once said. Thanks Fran for another interesting and thought invoking video. I hope all is well in Philly and best regards from the UK.
The sentiment of the letter is that Elon Musk wants to be first in this race.
@@sciencecompliance235 Yeah that makes sense thanks for the reply
I have yet to see a bot take over any significant thinking jobs yet (programming, science, design, etc.). I think we're decades away from this, if ever. Any real evidence to the contrary?
Phone:
Fran:
XD
Also, I'm only that far, but whatever your point here, Sarah Gailey's short story "STET" is a must read on this topic.
I'm an ac and refrigeration installer and repairer and think it will be a while before that's automated
What would you recommend a person that had just started a higher education in the IT sector. Should i just throw the towel, going back to the mundane, hard manual labor - putting machines together - getting paid just enough to buy bad food and electricity, or pursue the IT in a special field that might 'survive' a bit longer?
greetings from the alps Fran