I spent over 40 years working as an illustrator, animator, designer, and graduate school professor. ALL of my jobs have recently experienced a 60% decrease in workforce due to Ai, and the universities that I worked for are closing. I have no idea what comes next. Thing is, creative work is something that most who do it are very passionate about. It's what gets us up in the morning. If Ai takes it away from us, then what's the point? How does stealing away the ability to earn an income doing something we love supposed to benefit society?
My concern is when Ai learns from Ai results falsely verified and relays untrue information that we then trust to be true and make bad (potentially dangerous) decisions on.
Chat gpt was already trained from internet informations without tagging negative from positive properly. (at a good percentage). So what you say already happened.
On the contrary that would be the best thing to happen to humanity . Because that means that ai needs humans to verify the results, Sadly i anticipate that this is not the case at all . Were fucked
Wasn’t tech supposed to help humans be more creative, to make decisions and do less labor work… turns out machines will tell us what are the best local labor jobs we can apply to
Its overwhelming how AI slowly overtake all aspects of our life. Starting from google search with chat GPT, finishings with self driving cars like Tesla. Amids excited crowds watching robots cook, babysit and create pieces of art, humanity cannot hear feeble voice of reason and conscience "stop!" I do believe we will see more negative outcomes down the road especially in relation to younger generations engaging AI more and more into their life.
What aspects of are lives are overtaken by AI? Except for google search summarization which recommends you to eat rocks and chatgpt which doesn't know how many "r"s are in strawberry (try it yourself). Relax bro
This already actually happened with some of Facebook's in-house AIs, YEARS ago. They instructed the models to barter for imaginary resources, and before almost any time at all had passed, the AI models invented their own language and started using it to speak to each other in ways the researchers running the project could not decipher at all. They had to shut the whole thing down almost immediately.
It's artificial. It's in the name. We all need to go back to human interaction and stop interacting so much with machines. Our kids are lonely, ignorant, and don't know much about relationships, and AI will completely end what's left of their ability to communicate with others.
Many AI researchers do not agree with the name “artificial intelligence” because there is no universal definition or understanding of intelligence. I’m not anti-capitalism because it’s one of the best economic models available, but capitalism and social media are the reasons why the youth are so lonely and ignorant. The corporations want them to be consumers and social media has been the perfect tool to accomplish that.
Communication with others is only important because of humans lack of understanding. I don't think you understand how close we are to standing in the presence of a god when it comes to AI. Imagine every human on earth knows every bit of information all at once. Communication would become arbitrary as we could all work towards any goal we wanted to effortlessly. AI is the most beautiful thing in our lifetimes. Even if human lives were only a stepping stone for AI, then enjoy the present moment, because it won't last much longer 😊
Zoom out a bit. Clearly we use technology to make up the abilities we lack. Not saying you shouldn’t raise your kids or participate in real life. But it’s clear that our natural instincts are not fit to run a peaceful, greedless world.
Yeah, that's what I always think when I hear about this stuff too. It sounds exciting, but only if we have the right social reform in place first. Otherwise this is going to widen the wealth gap tremendously and destroy our society. Obviously those are two opposite ends of the spectrum, but hopefully as the changes are happening we institute the right social programs. I don't think there's ever been a more important time to be participating in politics.
Unless you get on it early and use it… you should’ve been studying it daily the second GPT-3.5 came out… People using AI will make a massive amount of money very quickly.
This misrepresents the way “agents” are actually thought of in AI. It’s not that agents are experts, it’s simply an AI that can take some kind of action outside of itself. Basically, ChatGPT can currently only interact with the world in a “read only” kind of way, and then tell you about it. But it can’t “write” to the world (in the read/write sense, the way we’d think of a read-only document or an editable document). An agent is simply something that can take action in the world (like you can tell it to order you an Uber), not necessarily an expert.
@@Jerome-iwnl what do you mean AI will do everything for us we’re basically gonna live like rich billionaires. Rich and poor people will all have the same access to everything in the future so it’s actually good for poor people.
@@kevinsm2039 You're delusional. Most will lose their jobs. The few that control the AI will have all the power. Poor people will be stuck on physical labor jobs until cheap robots come along.
AI is apart of data science, machine learning is one of the divisions that encompasses it. It functions basically like statistics right now, probable things are spewed out
@@cesar4729 my comments might be dumb then prove yours is right. people's like you choose money if one has to choose between food and money when you are left with only one choice between the two.
And then the resistance will be created right? Terminator the real deal! We need a time travel like John Conner to come back and and create the resistance! And yes I am nuts 🌰
@Lerppunen vietnam is in a drought & just told rice farmers to slow down because they need the water for ai chip production. yeah we're gonna starve from stupid government decisions
All of the real wealth come from land, agriculture and the extraction of resources from the land. This is what holds the base of the economic system and holds the entire technology sector and the services/entertainment sector as well. This is done through the government. They take the money form land owners and inject into the technological and abstract economy, which is the economy where 90% of people work. Without the government, there wouldn't be any way of taking money from these land owners and there would be no abstract economy. The base economy is real and is a zero-sum game and everything is already taken. The abstract economy is not a zero-sum game, but it's somewhat of an illusion and can only survive from the money taken from the base economy. The problem is that, to accommodate the whole population, the abstract economy must be inflated. It's a very feeble and ephemeral economy and it needs to keep innovating and changing all the time. The abstract economy is so weak, that you could work your entire life and never save enough money to buy a good piece of land (After all, that's where 90% of the population work). When the government can't take as much money from the base economy, or when there are too many people on the country, wages inevitably go down. After all, the abstract economy can't keep alive by itself, so inflation happens. People lose buying power. And now, things are so desperate that we are going to have to work more, in order to earn a living. That's why most jobs nowadays are bullshit jobs. Most jobs and companies are just an excuse to take money from the land owners. It's all just a big play.
Not ironically, studies are being carried out where psychologists "blindly" rate AI responses better than those of psychologists, and only change their minds when they are allowed to see the origin of the response.
5:16 explanation about deep learning is not correct. Deep learning is a set of methods that allows training deep neural networks. It is beneficial to have lot of layers in neural network because deeper layers can learn higher abstraction, but training (learning) algorithms that was available before deep learning methods were invented didn't work well for neural networks with large number of layers (deep networks). It is not separate type of nn, or a way to organize training data, or something like that. It is a set of methods tthat enabled more complex networks training always wanted
I Hit $38k today. Thank you for all the knowledge and nuggets you had thrown my way over the last months. Financial education is indeed required for more than 70% of the society in the country as very few are literate on the subject.
Interesting video. I would add that robotics is definitely an exploding field as well and the cost is dropping quickly, manufacturing jobs will be replaced in 10 years or sooner imo.
Ehhh… still goona be a long while. Look at how unreliable milking robots are, and you still need a human to fix it when it breaks. Then there’s the complexity of the job at hand that needs to be accomplished through brainstorming, and I still haven’t really seen AI show any good ways of addressing complex problems that aren’t necessarily written down to solve. For example a cow jumps a gate and is stuck in an awkward position that might require sawing, and lifts.
@@Ryantrock8888The small detail is that the last 12 months investment in AI and humanoid robotics has escalated insanely. If you manage your expectations based on the pace of progress of the past, you may have a rude awakening.
AI is not only going to change the world but it will define the world in the future ! AI is only becoming more advanced & user friendly. It comes down to us as humans if we will use it to our advantage, to make life better or if it will be used to benefit selected groups while doing a disservice to others. Time will tell ..
I'm a web dev, you can't imagine how helpful claude 3.5 sonnet has been in the past weeks for my work. The large context window changes everything, I can keep on a conversation about an specific thing, while feeding multiple long files for context. It's like having a very experienced web dev sitting at my side all day. Sure, it's not perfect, but I'm amazed at how much it can do. 95% of the stuff is accurate and helpful.
The agent thing is already here. It is why China can have such an efficient factor to consumer model. 90%+ of the communications required for that are automated. It is also using it to remove a lot of the office work that are related to communication both for optimization but also to fight corruption. They are experimenting with smart cities where millions of sensors are put together. Fire alarms that will automatically report directly to the fire department while also informing the traffic light grid so it can clear a way for the firetrucks and that is just the start. Other countries will be doing this to but I expect them to be a lot more secretive about it. Especially in countries where there is very little trust in the government or in situations where governance is actually done by major corporations and investors. It's going to be hard sell if you've been laying about democracy to your people for generations but in order to be competitive on the global stage they will have no choice to use every last bit of data and use it as efficiently as possible.
This video was so awesomely done that Andrei has to have a team behind him helping into producing!? Again loved the information on it, the graphics , music, the edit, top notch!
This is very important knowledge and one of the reasons why buying RENDER is essential. Render is the network which most AI protocols will run on in the near future.
Many experts in the field believe we'll reach AGI by around 2030. Until then we'll make progress but it may not be as mind blowing as people expect. Just so everyone realizes this is a longer term bet than a year or two.
How can customer service reps be replaced if llms hallucinate? This is proving to be a tougher issue to solve because of the probabilistic nature of llms, hallucinations are consequence of their design. Maybe it will be solved in five years, but i know if i need to call my financial institution regarding my investments and speaking to a human is no longer available, then they'll lose my business. I concede, this maybe a futile effort as they'll all adopt ai agents, but it raises a question, should average citizens resist ai where they can?
Andrei, thank you for continually sharing top-quality info and videos! I would also love to hear your advice on what someone can do after maxing out all the generally-advised accounts (like 401k, IRA, HSA, 529, etc) - what would be great next steps ideas before investing into taxable accounts? Thank you for all your hard work and generosity. Your content has changed the game for financial education!
I'm so sick of hearing about AI in the wake of endless tech layoffs. Nothing good will come of this, the rich will get richer and everyone else will be living off food stamps,
I believe u r wrong about politicians. Arnt they just data driven decision makers? As long as they do not use all their power to protect their jobs, they should be replaced by agents... as long as the agents can be trusted to do what is right for humanity and not be controlled by companies for their personal gain. The deep future this feels like it should be an inevitability. But maybe that's my Sci fi brain hoping we can get to someplace better than what we havr now.
It's going to get crazy if we can advance from pre-trained models that use lots of energy and time to train TO live-training models that can learn in realtime on consumer grade hardware.
You missed something >when a model is being trained it doesn’t store the data. imagine listening to every blues song. maybe after you would have some weird shapeless concept of all the music a model has an embedding space the latent space is where it (we don’t know how) sort of creates an impression in high dimensional vectors (12,300 dimensional vectors) there is some lossy part of the space where cats are in every model this is why suing is stupid it is lossy not lossless there is no file our minds can not understand 12,288 dimensions at inference your prompt is fed and the model simulates a world in which an output is typed or spoken in a way a helpful assistant would type even if that is rap lyrics for chaucer as the context window expands this simulation continues until you and the model’s assistant are simulated in the latent space
I think robots will be here faster then people think, think about it, if robots make robots, then they'll get cheaper pretty darn fast, also if it's software it needs, that doesn't seem hard either and once it's done then all the robots will learn it collectively @ the speed of a download, I think 2026 will be mass manufacture of them and by 2028 we will be seeing a lot of them around, what you guys think?
when in a buble, there are chances we overestimate the timeline for all these things to happen. If you take a step back, you realize that most of the things AI is practically used at, are tasks that are allowed to fail ( consistently sometimes ). Which kinda limits their practical usage.
As cse engineer I can say that all things he mentioned takes Atleast 25 years to develop agents are like phase one of robots. Self learning ai we didn’t develop anywhere in the world like terminator movie . All we have based on assumption we give advice that would be majority of time true. 😅😅also trained model you said some one writes in note when you ask questions it checks the data base related content and give you answer in comprehensive way. Ai have so many areas we can only develop these things only. Self thinking and give answer Robots or Ai takes Atleast 50 years lol😂😂😂 I like to listen so many things.
Not to mention they need to somehow stop these LLMs from hallucinating complete bs. Until they do so they are practically useless, if I can't trust the information it comes out with, what is the point of asking it 😂
"They would start communicating in new language and we have to stop the experiment" Like ask them not to invent any new languages, lol. Really, what on earth would motivate you to make such strange predictions? Except the news article from a couple of years ago.