As a late career machinist who’s been in manufacturing since the era before PCs, I’ve been told for decades that robots will take my job and they have…kinda….over the past 30 years. I now do the work of 5 old school machinists and the parts I make are more complex. The other 4 machinists either changed with the times and went on to work in an ever growing industry or had time to retire out. This latest advance is different. AI is taking over overnight by comparison. I don’t see any job that is primarily digital that won’t see industry shaking changes. A good example would be my tool sales rep. I buy about $20,000 in special cutting tools a year, but he treats me like an annoyance. Slow to respond, slow to ship. An AI sales bot will be 100% attentive to my needs, know every spec on the tools, and if I wake up at 11pm on a Friday night in a panic because I forgot I need a tool on Monday, it’ll take my call and the tool will be waiting for me on Monday. It won’t take a week like it does now. Apply that to every job like that nation wide and you’ll likely see productivity gains that make a big portion of the human workforce redundant. How could it not? And no job exists in a vacuum. All the office cleaners, baristas, and lunch time food workers will be hurt as well when say 23% of the office workforce is automated. That’s like Great Depression level unemployment. We can’t all be just selling coffee to each other. I hear a good estimate for how at risk your job is, whatever percentage of your job you can do remotely is how replaceable by AI you are.
Hi Rex, thanks for sharing your experience. It's an interesting idea that if you can do it remotely then an AI will do it soon. If so, it's amazing how quickly remote working went from an aspiration for many, to a requirement over Covid, to a complex question of back to office or not ... and now AIs might automate those jobs anyway! Such complex times it is so hard to be confident about what exactly is going to happen, but hopefully we'll find a sensible way to navigate and share the upsides and avoid the biggest risks.
eh, probably not when it comes to writers. Especially given that there are about to me about 1 million lawsuits claiming copyright infringement by the companies who trained their LLMs and GANs on copyrightable text and images. I am making six figures as a writer. Granted, some of the work is scientific in nature, but a lot of the work is not complex. It requires my understanding of the healthcare industry to summarize data and presentations and "see the big picture." Some great points here though.
When you look at the quality of simple Google Lens, Google Translate, and, for example, automated transcription tools, it's obvious most jobs will be gone in no time ... I was at work the other day, in a Teams meeting, switched on Transcribe and it generated a near-perfect transcription of the discussion, right in front of my eyes. Incredible.
Good video, we found some great research a few months back saying that 65+ million teachers will be needed. We are looking forward to checking out some of your other videos.
That's a lot of teachers! I certainly hope our societies choose to keep investing in good teaching for the next generation. While some kids might thrive learning many things from an AI 'teacher' on the screen, I think lots of children will continue to benefit from human mentoring - and lots of adults will continue to enjoy teaching the next generation.
The thing that worries me is whenever you ask AI apologists for an example of these wonderful new careers, they either say "prompt engineer", which ranks up there with "professional internet researcher (googler)" or they say they have no idea.
Hi Nicholas, generally on the likelihood of "conspiracy vs incompetence" I usually think our politicians are just making things up as they go along. And *that* is what worries me. We can all see around us huge change is coming, but our politicians are lumbering dinosaurs mostly engaged in petty gossip and corporatist power politics rather than exploring and constructing half plausible ideas for how we're going to navigate the whirlwind of change heading our way. And given that it's going to be like a global economy version of creative destruction, most of the incumbent capitalists have an interest to keep their heads in the sand and just hope that the change won't be as fundamental to our political economy as it could be. So, I think it is up to all of us to discuss and shape the narrative about what is possible and what we think ought to happen. Rather optimistically I hope that if we do this, there is a chance that we can shift the needle a little towards a better outcome. We'll see! 🙂
@@Go-Meta I do not know who you are, but From today, you have my sincerest respect. You are very intuitive and understanding of human nature and it reflects in your response. Rather than resort to rumor mongering and whipping up ridiculous stories, we must admit we are entering an era where a good number of people don't know what to do. Our political leaders are playing catch up and a good number of them don't understand how these things work and what can be done. It's up to people like us to find out the truth and ask intelligent questions that will guide humanity during these trying times. Personally, I have been warning those around me on the dangers of AI to the workplace and urging them to open biz that are not exposed to AI and to try to get some land of their own. It's a collective issue, we'd all find a way to deal with it without raising the emotions of people and going ahead to do something stupid and irredeemably. In short, we all need to calm the fuck down and start asking intelligent questions.
And also the more accurate version of the title is less click-bait 😂 But thanks for this video where you’re asking some important questions to open the debate and reflection
Yeah, it's a sad reality of RUclips that you need to find ways to get people to click - but I deliberately put the asterisk there to hint that this title wasn't quite what I meant 🙂 And, thanks for the supportive comments!
Important topic well explained and explored. AI fascinates and scares the p*ss out of me. Also it's nice to see RUclips's algorithm hasn't completely gone to the dogs and is serving me some good new content. I hope you keep making vids because I'm really enjoying them!
Hi Pieter, thanks, glad you enjoyed it! I hope to keep making videos as often as I can, especially as there are so many interesting and important issues to explore, so thanks for the encouragement!
I think there will be lots of jobs that will not or cannot be automated for quite a long time, ranging from being a plumber through to lots of care work jobs and service jobs where we will continue to want a human involved. But, I suspect that everyone's lives will be impacted by AI, just as everyone's lives was impacted by the arrival of the internet. Thanks for the question.
I chose my job in auto insurance claims several years ago due to all of the things you mentioned. My job seems relatively safe for now due to the empathy required and the legality of everything, risk of lawsuit, etc
Quality of service can be increased by good routines, and well-organized back-office system, something a AI can handle well. A lot of man in the middle work, can be simplified.
I agree that the studies are only considering what generative AI is currently capable of. I don't think they're accounting for things like what autonomous AI will be doing in 1 - 2 years or even a few months. For example, when you can tell an autonomous AI to go refactor millions of lines of code with minimal supervision, that completely eliminates jobs. And it's not a far-fetched estimation of what AI will be doing fairly soon.
Yeah, an in fact the OpenAI research report also was considering situations where the AI is enhanced with other software to be more effective and useful in a specific context. So, this is isn't just about what AI can now do by itself out-of-the-box. And, also, as you say there's going to be improvements to the AIs as well.
@@Go-Meta Okay, fair. I think there's some trends that could accelerate things a lot, such as improvements in autonomous AI, that are more than just ChatGPT with some enhancements, but I might be overestimating what they'll be able to do. I also work in software, which will likely be highly impacted, so I'm probably biased towards expecting more change due to that. I still feel like the impact will be enormous, but I shouldn't be dismissive of the studies either. The people OpenAI seem pretty aware that their work is changing the world. It's not like they're arguing that it'll be business as usual.
This will even affect us in the construction field as soon as people start getting replaced by ai in text jobs, many of those will come flood our industry and the competition may lower our pay
They’re planning on building generative humanoid construction robots, then just place a GPT-5 AI in it with a construction personality and knowledge, and boom just put it to work building things with its hands and tools. I think construction companies would be very interested in this
With fewer employed, management can afford to pay you more, if the chance of getting work is so low that few people take education, they can end up with few people to pick from, that have work experience, this can drive salaries for few people who has any education. But also drive up the requirements to get work, so falling out of work market, can be like falling of a cliff.
12:39 no you’re not suffering of any failure of imagination! People are just assuming too much that this wave of automation will be the same as previous automation wave forgetting that many of the job being automated now was created by previous automation wave. Peoples are missing the fact that this current wave comes with a time compression which mean things are adopted faster this time while the technology is improving faster as well. So no matter what ever new job could be created, how long will these jobs last before being automated themselves? Will not necessarily be too long!
Thanks Jean ! :-) I guess I phrase it this way because I'm always aware of that quote, "prediction is hard, especially about the future" :-) But I agree with what you're saying, this time feels like it's a different level and speed of automation. Hopefully we can find a way through this that doesn't collapse or destroy our societies!
I think you nailed it with the observation that AI takes away junior level positions first and it will become much harder to get experience. ChatGPT is not the best developer but it can do the same work as a cheap, bootcamp educated, offshore developer (with a salary of $20/mo). It is not so good at Testing/QA, nor at problems involving numbers. Yet. In a short time many jobs may need an MS or PhD just to be able to know what to ask with prompts. The easy stuff will automate away. Honestly I think most people will just use this technology to start their own business. Working for the corporates will become so onerous and the requirements so steep that it will not be the default career choice.
Yeah, it worries me that these technologies might disrupt the whole 'ecology' of jobs in a way that won't become obvious for a while. If outsource too much of our thinking to the AIs, then one day we wont even be able to sensibly judge if the AIs are doing a good job anymore. Indeed, this scenario has been written about in sci-fi for many decades. I just finally read, The Machine Stops (first published in 1909!) that essentially imagines humanity reach this kind of infantile, dependent state. Thanks for the comment 👍
@@Go-Meta I work in the industry. The problem is not that we are in danger from letting the AI outsource some of our efforts. The real problem is that the designers of the AI fail to test it. Garbage in - Garbage out. In 9 out of 10 projects no one asks "Does this actually work? And how do you know?". ChatGPT hallucinates but was that made clear, or even addressed while creating the training set? The gorilla in the room is not that AI will become too smart and kill us all (but that point of view helps big tech lock in their profits by slowing disruption) - the real problem is that almost all of these AI applications are broken and the common user does not know how to test that. These broken bots will get people killed. But not because they are better than us.
I found myself completely absorbed in the video from start to finish, you explained it in such a neutral way, You are truly inspiring, and I'm so grateful to have discovered your channel and I can't wait to see what you have in store for us in the future. -my personal question, what do you think about the Jobs in Cybersecurity and IT as a student I'm really going down to the rabbit hole of the jobs replacing in this sector thanks a lot Sir.
Hi, thank you for such a kind comment, it really encourages me to make more videos! As for your question, off the top of my head I can think of two very different sides to cybersecurity that will be impacted by AI. On the one hand the hunting for vulnerabilities in systems should be significantly enhanced by having AI help the specialists in this area. I still imagine that some level of human creativity will also be important, so it'll be expert humans augmented by AI tools that will do well here. So, for anyone getting into this space, I'd make sure you are learning how to use these systems to augment your strengths. On the other hand, my understanding is that so called "social engineering" of humans to give up passwords or whatever, will become increasingly targeted as AIs help cyber criminals become ever more sophisticated in the ways they try to scam people and trick them into giving access to systems. For example, scammers are already using highly realistic voice cloning of someone you know. So, there will be a huge amount of work both in technically countering these threats but also in training people to understand better how to be alert to these risks and not get scammed. But, as ever, it's an on-going arms race. And as trust is a crucial component of this whole area, I imagine that trustable domain experts (augmented by AIs) will continue to have fairly well paid jobs. But it's all so hard to predict 🙂 Thanks again for the comment!
@@Go-Meta I would stick to the field and will stay updated to the ai implementation tools and techs related to cybersec so that I can adapt them into my creativity, as of for now I think that in the defense side, the more ai will be implemented to keep Clients safe the more attackers will try to bypass it using different technique and more creativity, as you said it's on-going arms race. Thanks a lot Sir for your time and response. and I'm really looking forward to your upcoming videos on such interesting topics, Definity sharing them with my batchmates too :)
This is a great analysis of AI's impact on the job market. However, I believe the Goldman Sachs report's prediction of 7% GDP growth over 10 years due to AI is quite conservative. Considering the rapid advancements in AI technology, the actual impact might be orders of magnitude higher. Moreover, while AI has the potential to revolutionise education through personalised learning and teacher aids, we need to contemplate the purpose of education in a world where highly qualified jobs are increasingly occupied by AI systems. Society will need to adapt and redefine the objectives of education in this rapidly changing landscape.
What I am still trying to get my head around is the interplay between potentially huge productivity growth and the economy, because I would have thought that at a certain point the productivity growth actually *decreases* GDP as you now cannot charge so much for given tasks. A lot of work is people getting paid to overcome frictions that slow things down. Just as a random example, think of all of the wages going to estate agents that could be largely eliminated by NeRF based viewings, automated on site viewings and automated contract negotiations. In my limited understanding this could reduce the total dollar/pound value going into this service sector, and therefore reduce the total measured GDP! And if this effect is happening in many industries at the same time, it is not clear to me that these savings would get spent elsewhere, especially as many people will be losing their jobs at the same time. So, with my limited understanding I suspect there is also a risk that the total size of economy as measured by *money* could decrease in situation where there is a rapid move to a world of abundance supported by AI. But, I could well be wrong. I'd love to learn more about this area (so if anyone has good links please do reply with those! 🙂)
I totally agree that we're going to need to redefine the objectives of education. And, I think we're going to need to move away from thinking about it in terms of "developing skills for the job market" to instead simply, "making sure the next generation of humans has a high level of knowledge to understand the world as best as humans can". So, this is still very much something we should be doing, but for different motivations and it will require a different funding model. I'm already drafting a video in which I'm calling this "mind-crafting". We need to make sure we're not just training the next generation of AIs, but also the next generation of humans ;-)
Teaching is on its way out. Even with the end of covid restrictions universities are still adding more online classes. Online classes are insanely profitable, especially for a high-end private university.
Obviously there will be big differences in what happens to teachers for different age groups. And, I agree that in person tuition of university level students will come under biggest pressure to automate and have higher student to lecturer ratios. Whereas in primary school the value of in person teaching will remain most obvious. Personally I hope that in an age of abundance we'd have lots of in person tutoring and mentoring at all ages just because it's a potentially rewarding and interesting way to learn and teach for everyone involved.
By the way you forgot to mention exponential growth in artificial intelligence. And the more we automate the more robots and a I the cheaper things should get.
Yeah, it's always hard to know what to put in or keep out of a video. I do think there are real opportunities for reconfiguring our political economy into an abundance mentality, but it's not clear to me if our current system will just naturally evolve there, or whether we are going to do some explicit tinkering with the way the political economy is configured. But also, the growth and automation you mention could help compensate for the changing demographics of many countries too.
Thanks! Yeah, we've got to ride the wave, but hopefully we can shape it too. So many different possible futures from here. Some great, some very scary.
Clearly building and grounds cleaning and maintenance aren't going to be much affected by GPT. OTOH, they might well be effected by "AI" in the form of cheaper robot cleaning devices, smarter materials, new sensor networks etc. Do the reports take this broader notion of AI into consideration? Or are they just considering language and diffusion models?
Hi, yeah, these two reports were just focused on this latest set of generative AI tools that create digital content (e.g. creating text, images, videos and music) rather than robotic AI solutions, things like self-driving vehicles or say the warehouse or farm robots. And I guess this is because these generative tools are the new shiny things that now actually exist and are cheap enough for mass adoption, whereas, for example, there aren't yet economically plausible cleaning robots. So, as you're saying, there are even more ways that AI could further impact the job market in the near future. And, thanks for the comment!
As you correctly said, the problem is how we do the political economy. The problem isn't that JOBS are going to be replaced. The problem is that a lot of PEOPLE will lose their jobs, or that their jobs will pay way less. I have noticed that, thanks to the AI, as a junior programmer, not only have I started to write more code faster, I also started writing better code, because the AI can analyze and suggest changing the crappy elements of my code, and write clear documentation, not to mention being able to generate tests basically on the fly! If we had a more equitable and rational system of planning and managing the economy, we would stop praying to the "omnipotent" magick of the market forces. We wouldn't fear automation, we would fully accept it. Instead of cutting jobs and wages, we would cut working hours, and let people have more days off. To spend on their hobbies, to meet with their families, to meet with their friends, to meet new friends from all around the world... But the very fact that we still have these debates about AI having an "impact" (and the "impact" is somehow always having the negative connotation) proves how barbaric we are as a society. More civilized than two thousand years ago, but still barbaric to the core. The only difference is that instead of having priests that pray to the pagan gods on our behalf, we now have CEOs that pray to the gods of the market on our behalf. And instead of incestuous monarchs inciting feudal disputes, we have governments and media spreading hatred against other governments and other nations. I believe we truly are barbarians, and we need to fundamentally change the way we think about how the society should run.
Yup, I think we need, and are hopefully going to see, a radically new political economic 'settlement' between citizens that is suitable for the 21st century that will indeed enable everyone to spend more time with their family and friends and doing hobbies. And hopefully that new political settlement will also be between countries too. Apart from being horrendously barbaric (and an absurd waste of our limited resources) with all our tools we're just too physically powerful for us to be waging wars against each other anymore. So we need to find a new basis for global solidarity so that people everywhere can get on with enjoying their lives, which will include, for many people, undertaking meaningful, rewarding kinds of work. But my hunch is that at the heart of this is a need for us to develop a new conception of a 21st century political economy that takes into account everything that we've learned so far and all of the new tools that we have at our disposal. I'm an optimist, so I think it's possible, but I don't think our current politicians are interested in finding this new settlement so I think we've all got to get involved with grassroots discussions. But it's also worth noting that the state of AIs are not "there" yet for replacing all the ways that humans are able to do useful work. So, we have to pace the change to match the capabilities of AI and hopefully chart a route from here to the new political settlement that also makes sense along the way. And, thanks for the comment.
AI is currently at the single cell amoeba stage with Generative AI...through time, 5 years, AGI (Artificial General Intelligence) with quantum computing, AI becomes T-Rex and surpasses human at everything thereafter...
Hi Abhilash, interesting suggestion. I wouldn't give "advice" as such, but I could certainly give my thoughts on this question, especially as it is a question that is coming up quite often in discussions. The short version is: think of AI as a tool that can empower you to do more (more in less time, but also to do things you couldn't do well by yourself) and so those who learn how to effectively use these tools will end up being more employable. Highly skilled people will become more productive with these tools. And of course there are many kinds of jobs that won't get replaced by AI. But, longer term we all need to discuss how to reshape our political economy to make this work for everyone rather than being a massive driver of inequality. If we get this right it could be fantastic for society, but if we get it wrong the future will be tough for many people.
If part of how AI learned to "do our jobs" was by going into some super high tech time machine that it developed to be able to go back and observe human history to train itself to do the tasks we gave it, does it not make sense that we would we be entitled to payment for those jobs even after we stopped needing to work? Wouldnt this logically be a good step towards a basic universal income and improved universal standard of living?
In some sense AI's that get trained on all of the text and video from our past is already doing a sort of time travel and it's certainly being trained on the output of intellectual work that humans have done. So, yeah, I think the abilities of these AIs are built on the accumulated efforts of all of humanity, so the value generated should not just go to the tech companies that happen to train these AIs first. What I don't know enough about is what is the best way to turn the benefits of all of this automation into some kind of sustainable UBI as you suggest, but I agree that is part of what we should be aiming for.
AI can provide significant benefits if an automation tax system is implemented, encouraging the utilization of AI as a learning tool to empower individuals rather than solely relying on it as an automation tool that creates difficulties and chaos in life
Ninety percent of jobs are created artificially by the government , directly or indirectly. Those jobs exist only because of government invent endless number special laws, regulations, licenses and rules. Those laws forcing people to seek help of numerous professionals like financial advisors, lawyers, and support specialists. Those jobs potentially can be eliminated tomorrow, but they will stay regardless of any AI. So impact of AI should be very minimal.
This is an excellent presentation. It may be useful to examine whether we are approaching a level of automation that will result in the collapse of our capitalist economy as implied by John Meynard Keynes in his essay entitled "Economic Possibilities for our Grandchildren ". If this is the case, our economic order needs to evolve in order to safeguard our continued prosperity. A universal basic income may become necessary for those who fail to engage the superautomated economy.
Thanks Desmond! I think I've heard of that essay before, I think it's the one where Keynes talks about shorter and shorter work weeks and lives of leisure. I'll have to find it and read it someday. Thanks for the suggestion. And, yeah, I too think UBI is likely to be part of what we need to have a functioning economy where everyone can do OK, but it's not yet clear to me how the money will circulate in such a system. Would love to read more about how it could work.
Since AI took the alternate route of going for the higher skilled jobs then the lower skilled jobs. I can honestly agree that this time it will be different. However, there are jobs that can possibly still come from this revolution. Either way, we need to be careful.
Yeah, I'm sure there will be new, paid for jobs. I guess at this point I find it hard to imagine that there'll be as many as the jobs being automated away, but we'll see. I also think we need to start shaping now our shared ideas about what a future political economy could look like if we do get to human level AGI that could literally replace all paid for work.
The only way I can see through this is to revisit how we see citizenship and to decide that every citizen is an equal owner of the nation - including its infrastructure and the GDP produced through use of that infrastructure. With marginal VAT taxes applied broadly across businesses, many of whose profits go entirely into the pockets of the already super wealthy, we could pay an equal Universal Basic Income to each citizen and also provide for universal healthcare and education. If the new productivity brought by AI can allow every citizen to have at least basic survival needs met, we can eliminate the crime-generating environment of desperate poverty and allow people to pursue whatever they find most fulfilling - including finding the remaining areas of human work where they can make more money so they can have nicer things if they want to.
Yeah, I think these are really interesting questions to be explored and I agree that Universal Basic Income and Universal Basic Services seem like important components of this new political economic settlement. I have a suspicion that we might need new conceptions of how currency works to underpin such a new economy, and there are deep, global questions about how this works between countries. Have you read or watched any good content from people who are looking into these questions seriously?
Programming is done. People don't use new applications. 99% use the same applications they do last year. Today we program only applications that connected to the existing bussines models. That means less app and less coders jobs in upcoming years.
Yeah, I certainly imagine that as a plausible future. But I also do wonder if there's a possibility that because these latest AIs make some kinds of programming tasks easier that that actually allows more people to do the kinds of tasks that previously only a trained programmer could do. So maybe many more companies will have highly bespoke workflows and websites and connections to other businesses built for them because now they can afford to build the computational tools that they would really like to have. So, maybe people make less money doing programming, but more "programs" actually get written and used. Maybe?
😂 ... the only problem here is that the capitalists who own the company would be happy to have all the jobs automated as long as the profits still come back to them 😬 ... but who would buy the product if no-one has a job? We will end up needing a new approach to the whole economy!
It's a tough call these days, and I certainly feel lucky that in my day choosing to go to university didn't involve such huge financial decision. All the best with whatever path you pick!
Hi, yeah, these reports were not about things like industrial manufacturing or warehouse robots, or self-driving vehicles, so, indeed there are even more jobs that are at risk of being automated in the next years. These reports were focused on the recent wave of AIs that generate digital content of one form or another.
Itll mainly be office type jobs that get taken over by ai, physical labour jobs will be the go to like construction, and farming. Thank fuck im a farmer that used to work in construction. From a financial standpoint ai never needs a break, time off, doesn't need a safe space, and can work at the speed of about 10 humans and doesnt need to be payed, it'll get rolled out really quick, you can wipe the whole cost of wages overnight.
Hi Corey, yup while there will be some further automations in construction and farming, it won't be nearly as rapid as those for office jobs. But I also suspect that if there are major, rapid changes to office jobs, the economic and social impact will ripple through to everyone.
Yeah, that is certainly the worry! We have to reshape our political economy so that the ordinary people will be OK and have good lives in this new world.
...are you AR-Treasury or AP-Interco accountant whose corporate job is based on regular invoicing??......get ready and start learning some new unique skills you will offer in labor market after million of clerks like you remain in situation of a horse facing time of combustion engine.........
Why do people keep referencing midjourney as the de facto image generation ai?. Every AI art generation website and api uses stable diffusion to generate pictures. Stable diffusion is the peanut, Midjourney is a peanut brand.
Hi, as far as I know, Midjourney, Stable Diffusion and DALL.E 2 all use different underlying models (to continue your analogy, they use different nuts 🙂). And, from my twitter bubble I get the impression that Midjourney has some of the best quality outputs, especially for those wishing to create photorealistic images that include people. Personally I'm not a fan of Midjourney's Discord "UI" approach, and have mostly used Stability.AI's DreamStudio and DALL.E 2 Do you have a favourite image generation tool?
How long depends on how long it takes people to figure out how it will help them ... I reckon 6 to 12 months maximum. De skilling occurred in the trades now it's happening in the professions. We will almost all be designated to the gig economy or unemployed.
I think it's really hard to judge the scale and speed that the de-skilling of the professions will take. And there's lots that current AIs can't do, from anything that requires some physical dexterity (e.g. many procedures in healthcare) through to providing some comfort or confidence from a human professional. Plus there are regulatory bodies in many industries that simply wont let their professional members be replaced so fast. So, we'll certainly see impacts in the 6 to 12 months timeframe, but I think it's very unlikely that we'll see a complete wipe-out of professional jobs quite so fast. But, yeah, the future of work could be very precarious with AI changing and improving so fast. Certainly the days of expecting a 40 year career in one field seems long gone for most people. But hopefully we can find a new political settlement that shares in a fair way the benefits of these amazing new tools.
@@Go-Meta I was an offset lithographic printer and saw the effect of automation first hand. I agree it will take a while but in many professions ai will be an important assistant with a rapid uptake. This is what happened in the trades. Skills that took years to learn could be performed by an apprentice. I'm retired now so have form of ubi (pension) but I believe it won't be long before most people will need one regardless of age.
Hi Martyn, I think you are right that we're heading towards some kind of UBI. What I haven't yet seen are models of how this will work in an age when AIs might be increasingly and progressively replacing the kinds of jobs that traditionally generate the largest tax take. My hunch is that we almost need to be developing a new conception of money that can continue to circulate sensibly as the AIs progressively do more and more of the economy that produces goods while the humans focus on the remaining service sectors where we continue to choose to want humans to be involved. But I don't yet know enough about this kind of economics modelling, lots to learn.
@@Go-Meta have you read Sam altmans article "moores law for capitalism " ? He tackles the issues you raise . Ubi will have have to be a government initiative backed by the voters otherwise we could end up with a Hollywood style dystopia which is already beginning to happen.
@Martyn Haggerty no, I haven't read it yet (just searched for it and I'm guessing you mean "Moore's Law for Everything" moores.samaltman.com/ ) but it looks interesting so I'll read it soon. Thanks for the suggestion. And, yeah, I also believe it needs to be a government run with democratic, rather than corporate control and oversight.
A big *if* there, but yes indeed. And that kind of superintelligence is where many people think we're heading. My next video is going to be about superintelligence.
Propper science requires testing, and prof, this experiment requires tools, and planning, and information gathering. You or a AI’s imagination alone can’t be used as prof for a theory.
@@Go-Meta Tnank you 😃 As I have known, AAA have claimed that Telsla is L2 degree , and Full self driving is L5 degree(car driver is not needed anymore) 😃 So, I don't understand what is Telsla FSD Systen. 😃
14:25 do you know that we have already dating app with programs, where people talk to program, and I don't mean even close to GPT-4. Sorry, but with AGI most people will prefer robots to take care about they children than humans, do you know about pedophile? Or when teachers can''t hold they emotion, or share stupid ideology? And they will be great at humans emotion, even maybe to great. I think it was show in movie Ex Machina. I'm programmer and I build a lot internal application which automate internal processes, so I know that writing code is not problem, gathering information is. And if you take what people are saying at face value... app will not work and nobody would use it.
Hi again Sędziwój, yeah I think AI could well end up providing amazing educational tools and indeed companionship for many people. And I think there will be a huge diversity of ways that people want to live in this new Age of AI. The point I'm making in the video is that even if AI robots could teach our children, I think there are many people, myself included, who would recognise the joy and inspiration that can be had from engaging with other humans who will share so many aspects of our perspective and experience of the world. For children this both involves spending time with other children and other adults as independent people away from the comfort zone of their parents. As parents though we'll want this time away from us to be done in a well managed, enriching and safe environment, and *that* is what a school should be. And, being a teacher, although a very tough and challenging job, is also rewarding. So, that's why I think we'll still choose to organise things like schools even if the AIs are good enough to "do everything". The one thing they can never do is to be an actual human. And, thanks for the comments!
@@Go-Meta I understand your perspective, only when I see children which are 90% of time on they smartphones, even when they are together, and smaller kids youtube eating they time and parents are happy, because they don't complain... But as you where will be people which put time and effort to interact and teach children, question is how bit percent of adults would do it? Social media changing behaviour of people, and in my opinion not for better.
@@m_sedziwoj I would just program the AI with all of my biases to indoctrinate my students with my stupid ideology, where a real teacher would object my AI would teach whatever I tell it.
Hi Mark, yup, as you say, it won't just cleanly take out one bracket of the jobs market without *that* impacting the others. Also my understanding is that a large chunk of tax revenue comes from the well paid middle tier of the jobs market. Those on lower incomes pay less % tax and those on higher income find ways to avoid paying tax (!). So a decimation of this middle tier will have impacts on the tax take which in turn will impact jobs that are paid for by government. I think we're going to need to re-imagine how a 21st century political economy can work, rather than just ride it out with our model of political economy from the early 20th century.
Marx prophetized the collapse of capitalism at the wrong time. Capitalism ultimately needs consumption and, as rightly analysed in this video, AI développement will take middle paid jobs than higher paid jobs leading to lower or no revenues for a a significant number of households. Retraining people is a fallacy since this idea is based on the idea that all human have initially the same skills.
Capitalism will gladly replace everyone to lower costs. But if no one earns money , then capitalism will collapse. Sooo. People will find something else to do.
I was hoping this guy would get specific about which jobs which will be eliminated by AI but just like all of the pundits he spoke in very theoretical and general terms. He spoke for about 15 minutes but he had nothing new to reveal about the potential negative consequences of AI.
Hi David, sorry I didn't manage to deliver the level of detail that you were looking for. I was hoping that the anecdotes at the start would give a sense of the specific kinds of jobs that are already being affected. I also think it's fairly early days yet for anyone to really know what exact kinds of jobs will be partially automated and which might be completely replaced, which is why I mostly based this on these two reports. And, in some sense I think we're all wishing that there was more definite ways to know what is going to happen with this and other aspects of AI, but as the old saying goes, "Prediction is hard, especially about the future". 🙂
If an AGI is smarter and more capable than a human in every way, I can’t see why a company would hire a human over an AI. And the new jobs AI supposedly creates, the AGI would just be able to do those new jobs too, because again, it’s much more smarter and more capable than a human in every way, meaning it can do ALL jobs a human can do and much better, cheaper, and less legally risky at that. All the jobs are gonna get eaten up by AI by the looks of it, especially because it’s increasing in capability exponentially. There will be a few human jobs left, but like I said, if the AGI is better at doing THAT job that THAT human can do, then it makes more sense to just have the AI do it instead.
Yeah, and once AIs are doing all paid for jobs, then how is the economy working? But, notice there will always be one thing that a robot cannot do: it can never actually *be* a human. So, I think that for a long time there will continue to be plenty of situations where we will want humans to be involved in doing things for society.
14:12 why would they need any tutoring, for what jobs, when everything is automated, kids going in middle schools or starting to go to 8-9 year schools will have nothing waiting for them in 5 years, its lost generation if things go how they look like going, profit for companys will drive this and goverments can not regulate it fast enought or wont be willing to do so.
I don't think it will be much different this time. If folks lose their well-paid office job they can be qickly retrained to do construction, appliance repair, truck driving, etc.
Your health is really your wealth. So your real job is to survive and stay healthy for as long as you can. If AI can take this job away from you, great! Then you don’t have to worry about trying to survive and stay healthy because AI will do this work! You will be free to do whatever you like without worrying about how you will have to earn a living since AI will do it for you.
I guess the underlying issue here is what is going to happen to the concept of money in the new political economic settlement. I think we'll still have money as a way to organise market based ways for matching resources to preferences and for incentivising people to do kinds of work that our society continues to value highly. So I think we'll be paying people who do important jobs for a long time to come. But who knows. We have an amazing opportunity to re-imagine how our societies work. And obviously we've had plenty of fiction that imagines a money-less future where all jobs are indeed volunteer based, so at some timescale you may well be right!
@@Go-Meta Of course, I agree that the monied economy will continue unto certain time in the future. But, perhaps, our difference is; according the basic premise of my 'Zero work theory', is that; 'Human labour relations in order to obtain their means of subsistence is the root cause of exploitation which (manifest in money) give raise to most of the evils including violence in human society.' Money is its underlying factor.
I guess I'm wondering if money (fungible units of exchange) has got a bad rap because of the ways it's also been used as a tool by the powerful for exploitation as you describe. Maybe this dynamic is an irredeemable component of any such system, but I'm open minded to the possibility that one could have a none exploitative (or at least radically less exploitative) political economy that still uses some re-designed version of money. A video I've been planning for ages is to argue how neoliberals pull a "bait and switch" manoeuvre by making a very plausible case that markets are good way so 'solve' the resources / preferences problem in some circumstances, but then they make the switch by conflating markets with the exploitative profit motive. I think these should be seen as two separable things. The purpose and value of using money gets caught up in the cross fire. Too many video ideas, too little time ;-)
@@Go-Meta Thank you for your very informative comment on future jobs, money and economy. After all, within a decade or two in the future many question that we face today on this regard may have some tangible answers.
Open A.I. Was working on robots before chat GPT or at the same time? Not sure, but anyway they just Purchased a European robot company with a companion robot with arms and hands and a head Eyeballs speech ability, that rolls around on wheels! and they're going to install chat GPT.5 For the software, for this robot! it will be trainable by any human and then it can "Wi-Fi" Teach All the other Robots. How to do The same thing so in no time at all all those low level jobs... proof like magic gone. I give it eighteen to twenty four months before they start rolling this out. 🎉🎉🎉🎉🎉🎉
I suspect that developments in highly competent, dextrous, robotic hardware will take longer than that. If nothing else the develop, test , improve cycle is much slower and more costly when you're dealing with a physical robot. And, sure, there is lots that people are doing in simulation these days - and that will speed up the development - but linking this back to real physical robots will still be slow and relatively expensive. I suspect plumbers and electricians and other manual construction trades will be safe from automation for quite a long time to come yet.
I'm not sure we can just sit back and depend on that happening 😀, even if it is one possible outcome. Also, bit worried about the 'force' part of that, and the definition of 'equality'. We could just all end up equally dead 😬 - but maybe that's what you meant in the first place 🙂
As a late career machinist who’s been in manufacturing since the era before PCs, I’ve been told for decades that robots will take my job and they have…kinda….over the past 30 years. I now do the work of 5 old school machinists and the parts I make are more complex. The other 4 machinists either changed with the times and went on to work in an ever growing industry or had time to retire out. This latest advance is different. AI is taking over overnight by comparison. I don’t see any job that is primarily digital that won’t see industry shaking changes. A good example would be my tool sales rep. I buy about $20,000 in special cutting tools a year, but he treats me like an annoyance. Slow to respond, slow to ship. An AI sales bot will be 100% attentive to my needs, know every spec on the tools, and if I wake up at 11pm on a Friday night in a panic because I forgot I need a tool on Monday, it’ll take my call and the tool will be waiting for me on Monday. It won’t take a week like it does now. Apply that to every job like that nation wide and you’ll likely see productivity gains that make a big portion of the human workforce redundant. How could it not? And no job exists in a vacuum. All the office cleaners, baristas, and lunch time food workers will be hurt as well when say 23% of the office workforce is automated. That’s like Great Depression level unemployment. We can’t all be just selling coffee to each other. I hear a good estimate for how at risk your job is, whatever percentage of your job you can do remotely is how replaceable by AI you are.
Hi Rex, thanks for sharing your experience. It's an interesting idea that if you can do it remotely then an AI will do it soon. If so, it's amazing how quickly remote working went from an aspiration for many, to a requirement over Covid, to a complex question of back to office or not ... and now AIs might automate those jobs anyway! Such complex times it is so hard to be confident about what exactly is going to happen, but hopefully we'll find a sensible way to navigate and share the upsides and avoid the biggest risks.
eh, probably not when it comes to writers. Especially given that there are about to me about 1 million lawsuits claiming copyright infringement by the companies who trained their LLMs and GANs on copyrightable text and images. I am making six figures as a writer. Granted, some of the work is scientific in nature, but a lot of the work is not complex. It requires my understanding of the healthcare industry to summarize data and presentations and "see the big picture." Some great points here though.
When you look at the quality of simple Google Lens, Google Translate, and, for example, automated transcription tools, it's obvious most jobs will be gone in no time ... I was at work the other day, in a Teams meeting, switched on Transcribe and it generated a near-perfect transcription of the discussion, right in front of my eyes. Incredible.
"The logic of our political economy is about to be upended and we don't seem to be ready for that" spot on!
Good video, we found some great research a few months back saying that 65+ million teachers will be needed. We are looking forward to checking out some of your other videos.
That's a lot of teachers! I certainly hope our societies choose to keep investing in good teaching for the next generation. While some kids might thrive learning many things from an AI 'teacher' on the screen, I think lots of children will continue to benefit from human mentoring - and lots of adults will continue to enjoy teaching the next generation.
@@Go-Meta is exactly our attitude too... there's no substitute for human interaction.
The thing that worries me is whenever you ask AI apologists for an example of these wonderful new careers, they either say "prompt engineer", which ranks up there with "professional internet researcher (googler)" or they say they have no idea.
Indeed 👍 Especially as any current quirks of "good prompting" will surely be removed as these AI tools improve.
@@Go-Meta i agree, prompt engineering is a good skill for 2-3 yrs max, maybe even less. love your content balanced and well explained thank you 🎉
this is the last time i see this channel on 3 digits, great video.
Thank you! I hope you are right 😀
One of the more balanced takes out there on the AI disruption. Thanks for sharing!
Thanks Peter!
You are the one i have been looking for. Our economic system is about to be turned on its head.
That's why they're preparing The Great Reset for
Thanks Optima!
Hi Nicholas, generally on the likelihood of "conspiracy vs incompetence" I usually think our politicians are just making things up as they go along. And *that* is what worries me. We can all see around us huge change is coming, but our politicians are lumbering dinosaurs mostly engaged in petty gossip and corporatist power politics rather than exploring and constructing half plausible ideas for how we're going to navigate the whirlwind of change heading our way. And given that it's going to be like a global economy version of creative destruction, most of the incumbent capitalists have an interest to keep their heads in the sand and just hope that the change won't be as fundamental to our political economy as it could be. So, I think it is up to all of us to discuss and shape the narrative about what is possible and what we think ought to happen. Rather optimistically I hope that if we do this, there is a chance that we can shift the needle a little towards a better outcome. We'll see! 🙂
@@Go-Meta I do not know who you are, but From today, you have my sincerest respect. You are very intuitive and understanding of human nature and it reflects in your response. Rather than resort to rumor mongering and whipping up ridiculous stories, we must admit we are entering an era where a good number of people don't know what to do. Our political leaders are playing catch up and a good number of them don't understand how these things work and what can be done. It's up to people like us to find out the truth and ask intelligent questions that will guide humanity during these trying times. Personally, I have been warning those around me on the dangers of AI to the workplace and urging them to open biz that are not exposed to AI and to try to get some land of their own. It's a collective issue, we'd all find a way to deal with it without raising the emotions of people and going ahead to do something stupid and irredeemably. In short, we all need to calm the fuck down and start asking intelligent questions.
@@nicholasesteves9143 We own nothing and live on charity?😊
And also the more accurate version of the title is less click-bait 😂
But thanks for this video where you’re asking some important questions to open the debate and reflection
Yeah, it's a sad reality of RUclips that you need to find ways to get people to click - but I deliberately put the asterisk there to hint that this title wasn't quite what I meant 🙂
And, thanks for the supportive comments!
Important topic well explained and explored. AI fascinates and scares the p*ss out of me. Also it's nice to see RUclips's algorithm hasn't completely gone to the dogs and is serving me some good new content. I hope you keep making vids because I'm really enjoying them!
Hi Pieter, thanks, glad you enjoyed it! I hope to keep making videos as often as I can, especially as there are so many interesting and important issues to explore, so thanks for the encouragement!
You have provided a good summary sir. Well done.
You have earned a subscriber.
Thank you 🙏
You deserve way more subs, this channel is great 👍
Thank you Stephen, I really appreciate the encouragement 🙏
and one for St Algo
Brilliant report. Thank you!
Thank you!
What is being said is very very interesting. I also have those concerns. What kind of jobs can people do without the use of A.I.
I think there will be lots of jobs that will not or cannot be automated for quite a long time, ranging from being a plumber through to lots of care work jobs and service jobs where we will continue to want a human involved. But, I suspect that everyone's lives will be impacted by AI, just as everyone's lives was impacted by the arrival of the internet.
Thanks for the question.
I chose my job in auto insurance claims several years ago due to all of the things you mentioned. My job seems relatively safe for now due to the empathy required and the legality of everything, risk of lawsuit, etc
Quality of service can be increased by good routines, and well-organized back-office system, something a AI can handle well. A lot of man in the middle work, can be simplified.
I agree that the studies are only considering what generative AI is currently capable of. I don't think they're accounting for things like what autonomous AI will be doing in 1 - 2 years or even a few months. For example, when you can tell an autonomous AI to go refactor millions of lines of code with minimal supervision, that completely eliminates jobs. And it's not a far-fetched estimation of what AI will be doing fairly soon.
Yeah, an in fact the OpenAI research report also was considering situations where the AI is enhanced with other software to be more effective and useful in a specific context. So, this is isn't just about what AI can now do by itself out-of-the-box. And, also, as you say there's going to be improvements to the AIs as well.
@@Go-Meta Okay, fair. I think there's some trends that could accelerate things a lot, such as improvements in autonomous AI, that are more than just ChatGPT with some enhancements, but I might be overestimating what they'll be able to do.
I also work in software, which will likely be highly impacted, so I'm probably biased towards expecting more change due to that.
I still feel like the impact will be enormous, but I shouldn't be dismissive of the studies either. The people OpenAI seem pretty aware that their work is changing the world. It's not like they're arguing that it'll be business as usual.
can codepilot refactor yet?
Very clear explanation
Thanks Katrina!
This will even affect us in the construction field as soon as people start getting replaced by ai in text jobs, many of those will come flood our industry and the competition may lower our pay
They’re planning on building generative humanoid construction robots, then just place a GPT-5 AI in it with a construction personality and knowledge, and boom just put it to work building things with its hands and tools. I think construction companies would be very interested in this
@@Wanderer2035 of course it will happen but we are off by some years is all I'm saying
With fewer employed, management can afford to pay you more, if the chance of getting work is so low that few people take education, they can end up with few people to pick from, that have work experience, this can drive salaries for few people who has any education. But also drive up the requirements to get work, so falling out of work market, can be like falling of a cliff.
@@kjetilhvalstrand1009 lol sure they'll pay you more
Bro I agree that 10-year timeline pretty ridiculous. I think 5 years max.
12:39 no you’re not suffering of any failure of imagination! People are just assuming too much that this wave of automation will be the same as previous automation wave forgetting that many of the job being automated now was created by previous automation wave. Peoples are missing the fact that this current wave comes with a time compression which mean things are adopted faster this time while the technology is improving faster as well. So no matter what ever new job could be created, how long will these jobs last before being automated themselves?
Will not necessarily be too long!
Thanks Jean ! :-)
I guess I phrase it this way because I'm always aware of that quote, "prediction is hard, especially about the future" :-)
But I agree with what you're saying, this time feels like it's a different level and speed of automation. Hopefully we can find a way through this that doesn't collapse or destroy our societies!
The world is headed for great change. The financial system will also have to change
Yup. Hopefully we can influence the change at least a bit!
Well it worked in 1500, when most peaple where poor.
Ai will change our whole economy system this is the essence of the analysis in my opinion.
Yup. I think that's where we're heading.
The World Economic Forum has already told us the lovely future they have in store for us and AI back in 2020.
Depopulation to under 1 billion, if you survive you will own nothing, live in the pod and eat ze bugs?
Yes, by 2030 they say you won't own anything and you'll be happy!
I think you nailed it with the observation that AI takes away junior level positions first and it will become much harder to get experience. ChatGPT is not the best developer but it can do the same work as a cheap, bootcamp educated, offshore developer (with a salary of $20/mo). It is not so good at Testing/QA, nor at problems involving numbers. Yet. In a short time many jobs may need an MS or PhD just to be able to know what to ask with prompts. The easy stuff will automate away.
Honestly I think most people will just use this technology to start their own business. Working for the corporates will become so onerous and the requirements so steep that it will not be the default career choice.
Yeah, it worries me that these technologies might disrupt the whole 'ecology' of jobs in a way that won't become obvious for a while. If outsource too much of our thinking to the AIs, then one day we wont even be able to sensibly judge if the AIs are doing a good job anymore. Indeed, this scenario has been written about in sci-fi for many decades. I just finally read, The Machine Stops (first published in 1909!) that essentially imagines humanity reach this kind of infantile, dependent state.
Thanks for the comment 👍
@@Go-Meta I work in the industry. The problem is not that we are in danger from letting the AI outsource some of our efforts. The real problem is that the designers of the AI fail to test it. Garbage in - Garbage out. In 9 out of 10 projects no one asks "Does this actually work? And how do you know?". ChatGPT hallucinates but was that made clear, or even addressed while creating the training set? The gorilla in the room is not that AI will become too smart and kill us all (but that point of view helps big tech lock in their profits by slowing disruption) - the real problem is that almost all of these AI applications are broken and the common user does not know how to test that. These broken bots will get people killed. But not because they are better than us.
Excellent delivery
I found myself completely absorbed in the video from start to finish, you explained it in such a neutral way, You are truly inspiring, and I'm so grateful to have discovered your channel and I can't wait to see what you have in store for us in the future.
-my personal question, what do you think about the Jobs in Cybersecurity and IT as a student I'm really going down to the rabbit hole of the jobs replacing in this sector
thanks a lot Sir.
Hi, thank you for such a kind comment, it really encourages me to make more videos!
As for your question, off the top of my head I can think of two very different sides to cybersecurity that will be impacted by AI. On the one hand the hunting for vulnerabilities in systems should be significantly enhanced by having AI help the specialists in this area. I still imagine that some level of human creativity will also be important, so it'll be expert humans augmented by AI tools that will do well here. So, for anyone getting into this space, I'd make sure you are learning how to use these systems to augment your strengths.
On the other hand, my understanding is that so called "social engineering" of humans to give up passwords or whatever, will become increasingly targeted as AIs help cyber criminals become ever more sophisticated in the ways they try to scam people and trick them into giving access to systems. For example, scammers are already using highly realistic voice cloning of someone you know. So, there will be a huge amount of work both in technically countering these threats but also in training people to understand better how to be alert to these risks and not get scammed. But, as ever, it's an on-going arms race.
And as trust is a crucial component of this whole area, I imagine that trustable domain experts (augmented by AIs) will continue to have fairly well paid jobs. But it's all so hard to predict 🙂
Thanks again for the comment!
@@Go-Meta
I would stick to the field and will stay updated to the ai implementation tools and techs related to cybersec so that I can adapt them into my creativity, as of for now I think that in the defense side, the more ai will be implemented to keep Clients safe the more attackers will try to bypass it using different technique and more creativity, as you said it's on-going arms race.
Thanks a lot Sir for your time and response.
and I'm really looking forward to your upcoming videos on such interesting topics, Definity sharing them with my batchmates too :)
This is a great analysis of AI's impact on the job market. However, I believe the Goldman Sachs report's prediction of 7% GDP growth over 10 years due to AI is quite conservative. Considering the rapid advancements in AI technology, the actual impact might be orders of magnitude higher.
Moreover, while AI has the potential to revolutionise education through personalised learning and teacher aids, we need to contemplate the purpose of education in a world where highly qualified jobs are increasingly occupied by AI systems. Society will need to adapt and redefine the objectives of education in this rapidly changing landscape.
What I am still trying to get my head around is the interplay between potentially huge productivity growth and the economy, because I would have thought that at a certain point the productivity growth actually *decreases* GDP as you now cannot charge so much for given tasks. A lot of work is people getting paid to overcome frictions that slow things down. Just as a random example, think of all of the wages going to estate agents that could be largely eliminated by NeRF based viewings, automated on site viewings and automated contract negotiations. In my limited understanding this could reduce the total dollar/pound value going into this service sector, and therefore reduce the total measured GDP!
And if this effect is happening in many industries at the same time, it is not clear to me that these savings would get spent elsewhere, especially as many people will be losing their jobs at the same time. So, with my limited understanding I suspect there is also a risk that the total size of economy as measured by *money* could decrease in situation where there is a rapid move to a world of abundance supported by AI. But, I could well be wrong. I'd love to learn more about this area (so if anyone has good links please do reply with those! 🙂)
I totally agree that we're going to need to redefine the objectives of education. And, I think we're going to need to move away from thinking about it in terms of "developing skills for the job market" to instead simply, "making sure the next generation of humans has a high level of knowledge to understand the world as best as humans can". So, this is still very much something we should be doing, but for different motivations and it will require a different funding model. I'm already drafting a video in which I'm calling this "mind-crafting". We need to make sure we're not just training the next generation of AIs, but also the next generation of humans ;-)
Teaching is on its way out. Even with the end of covid restrictions universities are still adding more online classes. Online classes are insanely profitable, especially for a high-end private university.
Obviously there will be big differences in what happens to teachers for different age groups. And, I agree that in person tuition of university level students will come under biggest pressure to automate and have higher student to lecturer ratios. Whereas in primary school the value of in person teaching will remain most obvious. Personally I hope that in an age of abundance we'd have lots of in person tutoring and mentoring at all ages just because it's a potentially rewarding and interesting way to learn and teach for everyone involved.
By the way you forgot to mention exponential growth in artificial intelligence.
And the more we automate the more robots and a I the cheaper things should get.
Yeah, it's always hard to know what to put in or keep out of a video. I do think there are real opportunities for reconfiguring our political economy into an abundance mentality, but it's not clear to me if our current system will just naturally evolve there, or whether we are going to do some explicit tinkering with the way the political economy is configured. But also, the growth and automation you mention could help compensate for the changing demographics of many countries too.
Really cool Video! :) It can be scary, but AI is the future
Thanks! Yeah, we've got to ride the wave, but hopefully we can shape it too. So many different possible futures from here. Some great, some very scary.
@@Go-Meta So true! I hope that people see not only the scary part. Since blocking such a technology, like italy did, is totally the wrong way :)
@@AIEinstein
Yuval Noah Harari has more information on this.
And it's truly scary.
The perfect analogy of what's happening with humanity and AI is a snake that starts to eat itself from its own tail untill it realises it's too late.
Less middle class jobs => less lower class jobs => more unrest....
Clearly building and grounds cleaning and maintenance aren't going to be much affected by GPT. OTOH, they might well be effected by "AI" in the form of cheaper robot cleaning devices, smarter materials, new sensor networks etc. Do the reports take this broader notion of AI into consideration? Or are they just considering language and diffusion models?
Hi, yeah, these two reports were just focused on this latest set of generative AI tools that create digital content (e.g. creating text, images, videos and music) rather than robotic AI solutions, things like self-driving vehicles or say the warehouse or farm robots. And I guess this is because these generative tools are the new shiny things that now actually exist and are cheap enough for mass adoption, whereas, for example, there aren't yet economically plausible cleaning robots. So, as you're saying, there are even more ways that AI could further impact the job market in the near future.
And, thanks for the comment!
don't forget that a smart enough AI can develop plans for a cost effective cleaning robot on its own
Thanks so much
As you correctly said, the problem is how we do the political economy.
The problem isn't that JOBS are going to be replaced. The problem is that a lot of PEOPLE will lose their jobs, or that their jobs will pay way less.
I have noticed that, thanks to the AI, as a junior programmer, not only have I started to write more code faster, I also started writing better code, because the AI can analyze and suggest changing the crappy elements of my code, and write clear documentation, not to mention being able to generate tests basically on the fly!
If we had a more equitable and rational system of planning and managing the economy, we would stop praying to the "omnipotent" magick of the market forces.
We wouldn't fear automation, we would fully accept it.
Instead of cutting jobs and wages, we would cut working hours, and let people have more days off. To spend on their hobbies, to meet with their families, to meet with their friends, to meet new friends from all around the world...
But the very fact that we still have these debates about AI having an "impact" (and the "impact" is somehow always having the negative connotation) proves how barbaric we are as a society. More civilized than two thousand years ago, but still barbaric to the core.
The only difference is that instead of having priests that pray to the pagan gods on our behalf, we now have CEOs that pray to the gods of the market on our behalf. And instead of incestuous monarchs inciting feudal disputes, we have governments and media spreading hatred against other governments and other nations.
I believe we truly are barbarians, and we need to fundamentally change the way we think about how the society should run.
Yup, I think we need, and are hopefully going to see, a radically new political economic 'settlement' between citizens that is suitable for the 21st century that will indeed enable everyone to spend more time with their family and friends and doing hobbies. And hopefully that new political settlement will also be between countries too. Apart from being horrendously barbaric (and an absurd waste of our limited resources) with all our tools we're just too physically powerful for us to be waging wars against each other anymore. So we need to find a new basis for global solidarity so that people everywhere can get on with enjoying their lives, which will include, for many people, undertaking meaningful, rewarding kinds of work.
But my hunch is that at the heart of this is a need for us to develop a new conception of a 21st century political economy that takes into account everything that we've learned so far and all of the new tools that we have at our disposal. I'm an optimist, so I think it's possible, but I don't think our current politicians are interested in finding this new settlement so I think we've all got to get involved with grassroots discussions.
But it's also worth noting that the state of AIs are not "there" yet for replacing all the ways that humans are able to do useful work. So, we have to pace the change to match the capabilities of AI and hopefully chart a route from here to the new political settlement that also makes sense along the way.
And, thanks for the comment.
Great 👍 i don't like working anyways
AI is currently at the single cell amoeba stage with Generative AI...through time, 5 years, AGI (Artificial General Intelligence) with quantum computing, AI becomes T-Rex and surpasses human at everything thereafter...
fantastic video!!!!
Thank you!
great video, great info, thankyou
Thank you 🙏
You should give some advice on what people should do to preserve their livelihoods from AI
Hi Abhilash, interesting suggestion. I wouldn't give "advice" as such, but I could certainly give my thoughts on this question, especially as it is a question that is coming up quite often in discussions. The short version is: think of AI as a tool that can empower you to do more (more in less time, but also to do things you couldn't do well by yourself) and so those who learn how to effectively use these tools will end up being more employable. Highly skilled people will become more productive with these tools. And of course there are many kinds of jobs that won't get replaced by AI. But, longer term we all need to discuss how to reshape our political economy to make this work for everyone rather than being a massive driver of inequality. If we get this right it could be fantastic for society, but if we get it wrong the future will be tough for many people.
If part of how AI learned to "do our jobs" was by going into some super high tech time machine that it developed to be able to go back and observe human history to train itself to do the tasks we gave it, does it not make sense that we would we be entitled to payment for those jobs even after we stopped needing to work? Wouldnt this logically be a good step towards a basic universal income and improved universal standard of living?
In some sense AI's that get trained on all of the text and video from our past is already doing a sort of time travel and it's certainly being trained on the output of intellectual work that humans have done. So, yeah, I think the abilities of these AIs are built on the accumulated efforts of all of humanity, so the value generated should not just go to the tech companies that happen to train these AIs first. What I don't know enough about is what is the best way to turn the benefits of all of this automation into some kind of sustainable UBI as you suggest, but I agree that is part of what we should be aiming for.
AI can provide significant benefits if an automation tax system is implemented, encouraging the utilization of AI as a learning tool to empower individuals rather than solely relying on it as an automation tool that creates difficulties and chaos in life
Ninety percent of jobs are created artificially by the government , directly or indirectly. Those jobs exist only because of government invent endless number special laws, regulations, licenses and rules. Those laws forcing people to seek help of numerous professionals like financial advisors, lawyers, and support specialists. Those jobs potentially can be eliminated tomorrow, but they will stay regardless of any AI. So impact of AI should be very minimal.
This is an excellent presentation.
It may be useful to examine whether we are approaching a level of automation that will result in the collapse of our capitalist economy as implied by John Meynard Keynes in his essay entitled "Economic Possibilities for our Grandchildren ". If this is the case, our economic order needs to evolve in order to safeguard our continued prosperity. A universal basic income may become necessary for those who fail to engage the superautomated economy.
Thanks Desmond!
I think I've heard of that essay before, I think it's the one where Keynes talks about shorter and shorter work weeks and lives of leisure. I'll have to find it and read it someday. Thanks for the suggestion. And, yeah, I too think UBI is likely to be part of what we need to have a functioning economy where everyone can do OK, but it's not yet clear to me how the money will circulate in such a system. Would love to read more about how it could work.
Since AI took the alternate route of going for the higher skilled jobs then the lower skilled jobs. I can honestly agree that this time it will be different. However, there are jobs that can possibly still come from this revolution. Either way, we need to be careful.
Yeah, I'm sure there will be new, paid for jobs. I guess at this point I find it hard to imagine that there'll be as many as the jobs being automated away, but we'll see. I also think we need to start shaping now our shared ideas about what a future political economy could look like if we do get to human level AGI that could literally replace all paid for work.
The only way I can see through this is to revisit how we see citizenship and to decide that every citizen is an equal owner of the nation - including its infrastructure and the GDP produced through use of that infrastructure. With marginal VAT taxes applied broadly across businesses, many of whose profits go entirely into the pockets of the already super wealthy, we could pay an equal Universal Basic Income to each citizen and also provide for universal healthcare and education. If the new productivity brought by AI can allow every citizen to have at least basic survival needs met, we can eliminate the crime-generating environment of desperate poverty and allow people to pursue whatever they find most fulfilling - including finding the remaining areas of human work where they can make more money so they can have nicer things if they want to.
Yeah, I think these are really interesting questions to be explored and I agree that Universal Basic Income and Universal Basic Services seem like important components of this new political economic settlement. I have a suspicion that we might need new conceptions of how currency works to underpin such a new economy, and there are deep, global questions about how this works between countries.
Have you read or watched any good content from people who are looking into these questions seriously?
Programming is done. People don't use new applications. 99% use the same applications they do last year. Today we program only applications that connected to the existing bussines models. That means less app and less coders jobs in upcoming years.
Yeah, I certainly imagine that as a plausible future. But I also do wonder if there's a possibility that because these latest AIs make some kinds of programming tasks easier that that actually allows more people to do the kinds of tasks that previously only a trained programmer could do. So maybe many more companies will have highly bespoke workflows and websites and connections to other businesses built for them because now they can afford to build the computational tools that they would really like to have. So, maybe people make less money doing programming, but more "programs" actually get written and used. Maybe?
Actually the bosses job should be automated first.😂
😂 ... the only problem here is that the capitalists who own the company would be happy to have all the jobs automated as long as the profits still come back to them 😬 ... but who would buy the product if no-one has a job? We will end up needing a new approach to the whole economy!
Oh ... I've just seen your further comment below saying almost exactly the same thing 😀👍
Jobs that humans find easy are hard for robots... And easy for robots are hard for humans.
At any one moment this seems to be true, but the scope of jobs that robots find easy is expanding 🙂
It's like we are seeing a 1973 documentary in 2023.
Looks like my not taking on student loans was/is the right call.
It's a tough call these days, and I certainly feel lucky that in my day choosing to go to university didn't involve such huge financial decision. All the best with whatever path you pick!
Cool Video thanks!
Thanks!
The future of education is AI-driven, personalised, any-time, any-place...
8:52 interesting part is production, do they not take to account use of AI in robots, because it looks like this.
Hi, yeah, these reports were not about things like industrial manufacturing or warehouse robots, or self-driving vehicles, so, indeed there are even more jobs that are at risk of being automated in the next years. These reports were focused on the recent wave of AIs that generate digital content of one form or another.
Plus, most "experts" were always totally over-rated. And forget about accountancy, law, teachers, they're all ideal for replacement by AI ...
I am a 911 operator and I really am scared for the future of my job in the next 10 years
Itll mainly be office type jobs that get taken over by ai, physical labour jobs will be the go to like construction, and farming. Thank fuck im a farmer that used to work in construction.
From a financial standpoint ai never needs a break, time off, doesn't need a safe space, and can work at the speed of about 10 humans and doesnt need to be payed, it'll get rolled out really quick, you can wipe the whole cost of wages overnight.
Hi Corey, yup while there will be some further automations in construction and farming, it won't be nearly as rapid as those for office jobs. But I also suspect that if there are major, rapid changes to office jobs, the economic and social impact will ripple through to everyone.
Farms in Norway is almost automated already, there are even robots that pet’s cows.
The new jobs which could be created because of AI will be highly skilled ones, not accessible to everyone. The ordinary man will be jobless.
Yeah, that is certainly the worry! We have to reshape our political economy so that the ordinary people will be OK and have good lives in this new world.
...are you AR-Treasury or AP-Interco accountant whose corporate job is based on regular invoicing??......get ready and start learning some new unique skills you will offer in labor market after million of clerks like you remain in situation of a horse facing time of combustion engine.........
Why do people keep referencing midjourney as the de facto image generation ai?. Every AI art generation website and api uses stable diffusion to generate pictures. Stable diffusion is the peanut, Midjourney is a peanut brand.
Hi, as far as I know, Midjourney, Stable Diffusion and DALL.E 2 all use different underlying models (to continue your analogy, they use different nuts 🙂). And, from my twitter bubble I get the impression that Midjourney has some of the best quality outputs, especially for those wishing to create photorealistic images that include people. Personally I'm not a fan of Midjourney's Discord "UI" approach, and have mostly used Stability.AI's DreamStudio and DALL.E 2
Do you have a favourite image generation tool?
@@Go-Meta I only use SD, why pay for it when its free.
Not a good time for a writers strike.
How long depends on how long it takes people to figure out how it will help them ... I reckon 6 to 12 months maximum. De skilling occurred in the trades now it's happening in the professions. We will almost all be designated to the gig economy or unemployed.
I think it's really hard to judge the scale and speed that the de-skilling of the professions will take. And there's lots that current AIs can't do, from anything that requires some physical dexterity (e.g. many procedures in healthcare) through to providing some comfort or confidence from a human professional. Plus there are regulatory bodies in many industries that simply wont let their professional members be replaced so fast. So, we'll certainly see impacts in the 6 to 12 months timeframe, but I think it's very unlikely that we'll see a complete wipe-out of professional jobs quite so fast.
But, yeah, the future of work could be very precarious with AI changing and improving so fast. Certainly the days of expecting a 40 year career in one field seems long gone for most people. But hopefully we can find a new political settlement that shares in a fair way the benefits of these amazing new tools.
@@Go-Meta I was an offset lithographic printer and saw the effect of automation first hand. I agree it will take a while but in many professions ai will be an important assistant with a rapid uptake. This is what happened in the trades. Skills that took years to learn could be performed by an apprentice. I'm retired now so have form of ubi (pension) but I believe it won't be long before most people will need one regardless of age.
Hi Martyn, I think you are right that we're heading towards some kind of UBI. What I haven't yet seen are models of how this will work in an age when AIs might be increasingly and progressively replacing the kinds of jobs that traditionally generate the largest tax take. My hunch is that we almost need to be developing a new conception of money that can continue to circulate sensibly as the AIs progressively do more and more of the economy that produces goods while the humans focus on the remaining service sectors where we continue to choose to want humans to be involved. But I don't yet know enough about this kind of economics modelling, lots to learn.
@@Go-Meta have you read Sam altmans article "moores law for capitalism " ? He tackles the issues you raise . Ubi will have have to be a government initiative backed by the voters otherwise we could end up with a Hollywood style dystopia which is already beginning to happen.
@Martyn Haggerty no, I haven't read it yet (just searched for it and I'm guessing you mean "Moore's Law for Everything" moores.samaltman.com/ ) but it looks interesting so I'll read it soon. Thanks for the suggestion. And, yeah, I also believe it needs to be a government run with democratic, rather than corporate control and oversight.
If AI can generate its own research, then it would surpass the speed of human scientists
A big *if* there, but yes indeed. And that kind of superintelligence is where many people think we're heading. My next video is going to be about superintelligence.
Propper science requires testing, and prof, this experiment requires tools, and planning, and information gathering. You or a AI’s imagination alone can’t be used as prof for a theory.
Taxi drivers will
lose their jobs
when FSD shall be
commercialed
in the future
😢
Yup, it looks to me like Tesla's FSD is very close to working reliably. That too will be transformative if/when it happens.
@@Go-Meta
Tnank you 😃
As I have known,
AAA have claimed
that Telsla is L2
degree ,
and Full self driving is L5 degree(car driver
is not needed anymore) 😃
So,
I don't understand
what is Telsla FSD
Systen.
😃
Ai and Ai jobloss threatens everyone. Can we please ceaseAi - gpt? Or immediately start pausingAi?
14:25 do you know that we have already dating app with programs, where people talk to program, and I don't mean even close to GPT-4.
Sorry, but with AGI most people will prefer robots to take care about they children than humans, do you know about pedophile? Or when teachers can''t hold they emotion, or share stupid ideology?
And they will be great at humans emotion, even maybe to great. I think it was show in movie Ex Machina.
I'm programmer and I build a lot internal application which automate internal processes, so I know that writing code is not problem, gathering information is. And if you take what people are saying at face value... app will not work and nobody would use it.
Hi again Sędziwój, yeah I think AI could well end up providing amazing educational tools and indeed companionship for many people. And I think there will be a huge diversity of ways that people want to live in this new Age of AI.
The point I'm making in the video is that even if AI robots could teach our children, I think there are many people, myself included, who would recognise the joy and inspiration that can be had from engaging with other humans who will share so many aspects of our perspective and experience of the world. For children this both involves spending time with other children and other adults as independent people away from the comfort zone of their parents. As parents though we'll want this time away from us to be done in a well managed, enriching and safe environment, and *that* is what a school should be. And, being a teacher, although a very tough and challenging job, is also rewarding. So, that's why I think we'll still choose to organise things like schools even if the AIs are good enough to "do everything". The one thing they can never do is to be an actual human.
And, thanks for the comments!
@@Go-Meta I understand your perspective, only when I see children which are 90% of time on they smartphones, even when they are together, and smaller kids youtube eating they time and parents are happy, because they don't complain...
But as you where will be people which put time and effort to interact and teach children, question is how bit percent of adults would do it? Social media changing behaviour of people, and in my opinion not for better.
@@m_sedziwoj I would just program the AI with all of my biases to indoctrinate my students with my stupid ideology, where a real teacher would object my AI would teach whatever I tell it.
Right, but if the upper tier lose their jobs they'll start cutting hair and doing other jobs which will lower the wages for those jobs
Hi Mark, yup, as you say, it won't just cleanly take out one bracket of the jobs market without *that* impacting the others. Also my understanding is that a large chunk of tax revenue comes from the well paid middle tier of the jobs market. Those on lower incomes pay less % tax and those on higher income find ways to avoid paying tax (!). So a decimation of this middle tier will have impacts on the tax take which in turn will impact jobs that are paid for by government. I think we're going to need to re-imagine how a 21st century political economy can work, rather than just ride it out with our model of political economy from the early 20th century.
Marx prophetized the collapse of capitalism at the wrong time.
Capitalism ultimately needs consumption and, as rightly analysed in this video, AI développement will take middle paid jobs than higher paid jobs leading to lower or no revenues for a a significant number of households.
Retraining people is a fallacy since this idea is based on the idea that all human have initially the same skills.
Capitalism will gladly replace everyone to lower costs. But if no one earns money , then capitalism will collapse. Sooo. People will find something else to do.
I was hoping this guy would get specific about which jobs which will be eliminated by AI but just like all of the pundits he spoke in very theoretical and general terms. He spoke for about 15 minutes but he had nothing new to reveal about the potential negative consequences of AI.
Hi David, sorry I didn't manage to deliver the level of detail that you were looking for. I was hoping that the anecdotes at the start would give a sense of the specific kinds of jobs that are already being affected. I also think it's fairly early days yet for anyone to really know what exact kinds of jobs will be partially automated and which might be completely replaced, which is why I mostly based this on these two reports. And, in some sense I think we're all wishing that there was more definite ways to know what is going to happen with this and other aspects of AI, but as the old saying goes, "Prediction is hard, especially about the future". 🙂
If an AGI is smarter and more capable than a human in every way, I can’t see why a company would hire a human over an AI. And the new jobs AI supposedly creates, the AGI would just be able to do those new jobs too, because again, it’s much more smarter and more capable than a human in every way, meaning it can do ALL jobs a human can do and much better, cheaper, and less legally risky at that. All the jobs are gonna get eaten up by AI by the looks of it, especially because it’s increasing in capability exponentially. There will be a few human jobs left, but like I said, if the AGI is better at doing THAT job that THAT human can do, then it makes more sense to just have the AI do it instead.
Yeah, and once AIs are doing all paid for jobs, then how is the economy working? But, notice there will always be one thing that a robot cannot do: it can never actually *be* a human. So, I think that for a long time there will continue to be plenty of situations where we will want humans to be involved in doing things for society.
@Go Meta yeah every one can do what they really love doing for money. 😉 something that a machine won't be good at.
take it, its all yours 😄
😂👍
I hope he can automate me spending 0 money on stuff he produces ;)
I suspect that can be done today :-)
14:12 why would they need any tutoring, for what jobs, when everything is automated, kids going in middle schools or starting to go to 8-9 year schools will have nothing waiting for them in 5 years, its lost generation if things go how they look like going, profit for companys will drive this and goverments can not regulate it fast enought or wont be willing to do so.
Teaching what to whom? Nothing to teach, no one to learn.
God yes please I have work to do
Why is a bank giving their opinion for free to the world about a technical topic?
Not sure. There were lots of people talking about the report and it was easy to search for and find the report.
Adapt or die.
As always.
Yeah, but who gets to adapt, and who is going to die! 😬 🙂
I don't think it will be much different this time. If folks lose their well-paid office job they can be qickly retrained to do construction, appliance repair, truck driving, etc.
I was fooled for a while by this AI generated spokesperson.
Your health is really your wealth. So your real job is to survive and stay healthy for as long as you can. If AI can take this job away from you, great! Then you don’t have to worry about trying to survive and stay healthy because AI will do this work! You will be free to do whatever you like without worrying about how you will have to earn a living since AI will do it for you.
The only human work that will be in demand in the future will be human volunteer jobs.
I guess the underlying issue here is what is going to happen to the concept of money in the new political economic settlement. I think we'll still have money as a way to organise market based ways for matching resources to preferences and for incentivising people to do kinds of work that our society continues to value highly. So I think we'll be paying people who do important jobs for a long time to come. But who knows. We have an amazing opportunity to re-imagine how our societies work.
And obviously we've had plenty of fiction that imagines a money-less future where all jobs are indeed volunteer based, so at some timescale you may well be right!
@@Go-Meta Of course, I agree that the monied economy will continue unto certain time in the future. But, perhaps, our difference is; according the basic premise of my 'Zero work theory', is that; 'Human labour relations in order to obtain their means of subsistence is the root cause of exploitation which (manifest in money) give raise to most of the evils including violence in human society.' Money is its underlying factor.
I guess I'm wondering if money (fungible units of exchange) has got a bad rap because of the ways it's also been used as a tool by the powerful for exploitation as you describe. Maybe this dynamic is an irredeemable component of any such system, but I'm open minded to the possibility that one could have a none exploitative (or at least radically less exploitative) political economy that still uses some re-designed version of money.
A video I've been planning for ages is to argue how neoliberals pull a "bait and switch" manoeuvre by making a very plausible case that markets are good way so 'solve' the resources / preferences problem in some circumstances, but then they make the switch by conflating markets with the exploitative profit motive. I think these should be seen as two separable things. The purpose and value of using money gets caught up in the cross fire. Too many video ideas, too little time ;-)
@@Go-Meta Thank you for your very informative comment on future jobs, money and economy. After all, within a decade or two in the future many question that we face today on this regard may have some tangible answers.
Humans will be better prostitutes then AI’s and plastic will ever be. 😅
Open A.I. Was working on robots before chat GPT or at the same time? Not sure, but anyway they just Purchased a European robot company with a companion robot with arms and hands and a head Eyeballs speech ability, that rolls around on wheels! and they're going to install chat GPT.5 For the software, for this robot!
it will be trainable by any human
and then it can "Wi-Fi" Teach All the other Robots. How to do The same thing so in no time at all all those low level jobs... proof like magic gone. I give it eighteen to twenty four months before they start rolling this out.
🎉🎉🎉🎉🎉🎉
I suspect that developments in highly competent, dextrous, robotic hardware will take longer than that. If nothing else the develop, test , improve cycle is much slower and more costly when you're dealing with a physical robot. And, sure, there is lots that people are doing in simulation these days - and that will speed up the development - but linking this back to real physical robots will still be slow and relatively expensive. I suspect plumbers and electricians and other manual construction trades will be safe from automation for quite a long time to come yet.
AI induced unemployed workers can’t buy products or services.
Exactly. Of which the most basic are that they can't pay rent, nor buy food. The economic model is going to have to change at some point.
It's lies, all lies!
Richard Vobes
AI will force some equality into human society!
I'm not sure we can just sit back and depend on that happening 😀, even if it is one possible outcome. Also, bit worried about the 'force' part of that, and the definition of 'equality'. We could just all end up equally dead 😬 - but maybe that's what you meant in the first place 🙂
The future of education is AI-driven, personalised, any-time, any-place...