Correction at 16:20: Upon taking a closer re-reading of the the statement it seems like the internal red team were more trying to cover their backs incase something goes wrong, not so much flat out saying they advise against release. Privately they could feel either way, wanted to just note that!
It's interesting how only a few weeks ago they were trying to placate people by saying ChatGPT could only access internet data up to 2019 so there's no fear of it disclosing up to date, possibly confidential data. Turns out it been running in Bing, obviously in real time, the whole time. 😂
I'm testing AI chats in English and Polish. There are huge differences between their capabilities. Fortunately for non English countries language is still biggest problem for AI models, because they are trained mostly in English.
I think the most impactful thing AI has done is that you will never again be sure that something you see was, either made by a human, or real. It happened without comment and changed the world forever.
This is such a highly fascinating phenomenon that we now find ourselves monitoring. Will definitely be paying closer attention to the development of GPT-4, and I would love some more ColdFusion videos on this topic too
Just reminds me of Ian Malcolm in Jurassic Park saying stuff like, “you wield it like a kid who’s found his dad’s gun.” Or “Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.”
It's ridiculous how short-term minded people are. People are not even thinking about the effects on the economy through job loss this will have. How are we going to support pension's, Medicare etc? These goofs say will all be on UBI. Tell me how a monetary system not based off labor/services would work? Who is going to pay for UBI because the taxable population pool is going to dwindle to be able to pay it. How is the military going to get funding? Even the top 1% are not thinking this through. Their profit margins will dwindle with less and less consumers. The hope is that machines will do everything top to bottom, and it will be free. That would be great and all but the transition to that will be devastating.
@@frenchonion4595 You seem to be just one of the massive pool of people who actually think work is something good, and not just what we do to achieve something good. If AI replaces work, we will work on things it can't do. And if AI gets so advanced that there is no work left to do, we shall just not work. It is not even about UBI, we would simply not have to pay for anything, and everything would be so abundant that trading and markets would be obsolete. That is, when (if) AI actually replace any possible use human work.
You are helping us to be on top of the game and not be left behind. Especially in these day and age where information is polluted. Thank you for your hard work.
Thank you for making this video. Even though I'm in a AI thesis and I thought I would not learn that much, man I was so wrong, I learned a LOT. You certainly took dozens of hours to check all these information. Thank you again to make them so easily accessible.
Like always, and you'll see the unemployed in the streets protesting when it's too late or the unhinged conspiracy dumbs attacking institutions... People are tech illiterate, they never anticipate, they only react and always too late and in a caveman fashion
agreed. while there are 100m people using chatGPT, i had to introduce and explain it to every non developer friend of mine. This is definitely passing the general public by.
I don't comment on RUclips ever. Please continue to update us like this, it's moving at such a high rate and we learn of the advances made in such a short time usually too late. Thank you for producing such quality content
Im currently in medical school and used GPT to write case studies on COPD. It also was able to write a summary and compare two different studies based on the disease. Really helped me save time.
Protips: It's not always correct. Why are you using this, to write studies? Also, studies can possibly be incorrect, or funded for a bias result by a person to use later for advertising purposes as their 'evidence' (this is done and proven many times over)
I can not describe how much I love your channel. With your music in the background - pure education, art and chills all over my back. Please, never stop what you are doing.
Hey! Greetings from Italy. I teach English and I show your content to high level students. Your approach to reporting news and facts is concise, clear and very informative. Students and I absolutely love it. Congrats on your work and how you compile everything together so neatly. These videos open doors to a lot of knowledge, as you suggested in your video.
@@bsdpowa it has the potetnail to be way bigger though, a real AGI would throw the entire world for a loop, our entire economic system would go up in flames in a matter of months.
Well since they are based on real pixel experience, that's to be expected. BTW, Monetary authorities work under constitutional and legal mandates. Thinking that they will be usurped by a chart's prediction that obviously does not take that legal framework into account is so tech nerd naive. What you will see is more finely tuned FTX style grifting
Dude your videos have been critical in summarising the crazy speed of AI. Please continue making them at whatever rate you find important. Thank you so much for what you've been doing and I look forward to more.
What a great video. With the insane rate of progress, it is hard to keep up and it is great to hear someone like you with expertise talk about that stuff
I'm a research Scientist working on Alzheimer's. The impact of ChatGPT in my workplace can not be overstated. I use it daily to help me troubleshoot instruments, analyze research papers, and design new experiments. We are living through a revolution in technology unparalleled by anything before.
Every few weeks I'm shaken by the world of AI. If we make it there, future societies might be eutopic, but this change is going to suck for ours. I'm not sure how I'll ever get another job in webdev. People keep joking about how dumb it is now and how only bad coders will get replaced, but it feels like the exponential growth is going over everyone's head. What if where we are now is where image generation was in 2020? Or even 2016? That gives us.. 8 years until programming in the traditional sense is a horse carriage. And realistically it's not going to be 8 years. I've never had career existential dread but what the f**** is anyone supposed to do? Do we all get into farming and mining?
We don't really need humans for manual labor like farming or mining either. We have heavy machinery for most of the work and wouldn't be surprised if robo farmers begin to replace people too paired with AI. The traditional sense of a job and economy will have to be put under scrutiny and humans will have to adapt or fall flat on their face.
@@tfkdandsvkc exactly we already know what we’d use ai for because we’ve been using it for so long already it’s just implemented into bing and other sites now but there were always alternatives that already used ai There’s even chess robots from the 1700s which could be considered ai
Sensational video, I love that you cover both the positives and negatives as much as possible, without bias as much as possible. I remember your first video on ChatGPT and this was only a few months ago. It's absolutely insane how fast things are progressing, please keep the AI content coming.
It is because of ColdFusion that I have a career in I.T so please keep covering these kinds of A.I & Tech related topics.They are deeper that what the News cover,more informative and really fun to watch
Dont you see whats going on here? They keep the code private and fire the ethics committee? The AI isnt figuring out the answers to these riddles itself.....its just googling the answer and reggurgitating any answer it finds on google. Thats why they want to sell it so fast, before people realize what it is.....just google.
My wife doesn't normally get much time during the school day to do lesson planning; conference time gets eaten up by something else be that training, parent meetings you have it; so it generally gets done on the weekend or by staying after work, including putting together problems for her 4th grade math class. We used chatGPT this past weekend to write some two step word problems involving multiplication/division and addition/subtraction, and adjusted the numbers to be in line with current grade level. It was very smooth, very simple, and the scenarios it came up with were sensible in nature, saved her a bit of time.
Please keep us updated regularly! I wouldn’t mind an episode a week on AI progression as the data coming out is too fast for most of us to keep up with! Keep up the good work, you’re a god send
Yes please continue to review how the industry is going. Would like to see how to integrate the different sources together. I have the free GTP & I am sure it is my prompts that are not providing the response On the designs I am looking for. Thanks I’ll be watching and telling everyone to watch!
His explanation is inaccurate though. This is no where near AGI, and actually isn't technically AI, it's actually ML, or machine learning, which is just pattern recognition.
@@Nabeelco ok you have my eyes so explain when will we know the difference and what is the difference ? also isnt pattern recognition how we as humans learn?
@@thatguy9502 There's a fundamental difference between recognizing patterns and recording the patterns. ML models simply record the patterns. We do the recognizing and feed those patterns into it, then we train them to respond to our inputs with our expected outputs. But just because it's doing so, it doesn't mean that it has any understanding. It uses matrix math to take an input and determine the most likely output based on how it was trained. This is static. As in, unlike humans, where every new thing we see affects our future perception, and decisions, as well as how a new piece of information could change our response to one we would have never had before, these models just have data fed in and spit out a result. Which means there is no conceptual understanding in the model. You can't teach the model by explaining a concept to it like you can to a human. The model is trained purely by "experience" which is the data we feed into the model during training. The other thing is that, while the model may be returning patterns we'd expect for the input we give it, the model isn't actually picking up on the same patterns we do. A perfect example of this is Computerphile's video on tricking one of these models to confuse a remote for a coffee mug by changing just 100 pixels in the image. This is because how we identify what an object is, is based on our conceptual understanding, where as an ML model does it based on mathematical or numerical patterns. (You can watch the video here, and you can jump to 5:33: ruclips.net/video/gGIiechWEFs/видео.html) We assume that it's "thinking" and doing so like we do because we've tailored it's output to be human-like, but it's literally just noise or static that gets filtered until the static happens to look like a banana, or whatever we want it to look like. It's the ultimate "I see a bunny rabbit in the clouds" machine, but this machine is generating really detailed clouds that really look like bunny rabbits, but they are still clouds, it's still just spitting out noise. Another thing is humans have these things called mirror neurons, that let us conceptualize an idea. We don't need to experience driving to imagine how it would be like to drive. We can watch someone else, and what they do, and try to mimic or "mirror" them. ML models can't do this. They need the data that they will then use to formulate a response to be fed into them, as if we needed to learn driving by having someone move our hands and feet for us to be able to learn how to drive. They can't generate anything unique. Every output has a portion stolen from it's training data. All it does is re-organize these training inputs into an output we've trained it to return based on our request input. A video that explains this well is this Veritasium one: ruclips.net/video/GVsUOuSjvcg/видео.html (Jump to 3:42 to get to the relevant part, to about 14:57) So all this to say: There is literally no intelligence behind these models, and calling them AI is technically wrong. They are useful, and will no doubt change how we do things (for example my car uses AI/ML to literally fully self drive), but it will always be limited until we figure out that mirror neuron problem, because without it, these models have no actual understanding, and are basically just infinite monkeys on infinite typewriters, and we've just found the one that wrote out Shakespeare.
As someone who is about to choose what career path to take, I (and everyone else) would greatly appreciate if you could make a video about what people should consider when choosing a career path in the context of the rapid development of AI. With the increasing impact of AI on various industries, it can be overwhelming to decide which career path to pursue. Your insights on this topic would be incredibly helpful for those of us who are trying to navigate this new and rapidly evolving landscape. Thank you!
Learn how to work WITH AI, be the engineer or consultant that teaches and enables companies to maximize its abilities and their needs. Find creative ways in which you can implement it for people, and charge them to make it happen for them. Bluntly, sell the AI to companies to cut down their workforce and save on the cost of employing humans. That is probably the most secure route I can think of if you want to be successful in the white collar workforce. Everyone else with a 'white collar' job is effectively at a high risk of being made obsolete. Software developers, copywriters, graphic designers, lawyers, accountants and financial industry, business analysts, strategists, IT engineers, Human Resources, Customers Services, Call Centers, Quality Control, User accessibility Testing, Bankers, Traders, Consultants in most fields. Even fashion models, stock images, digital content creation I am a senior infrastructure engineer, and I have always thought my role, especially in the past few years has been increasing in demand, especially trying to integrate into cloud-based automated devops... and I am pretty convinced that the only way I will remain a desired candidate is to make sure I know how to be the one that helps companies integrate AI into their infrastructure. Everything else that my job entails can actually be automated. Most infrastructure engineers are currently focusing on exactly that right now... devops/ansible/python etc automation pipelines and playbooks that are already capable of deploying entire enterprise level IT environments. And I can confidently say that this level of AI is probably capable of making those fresh new skills irrelevant, AI will definitely do that better. It is the hands on, people-orientated roles which will become high demand, ... your farmers, construction, manual hands on stuff.. food production, tradesman, plumbers, electricians. A fair comparison would have been the industrial revolution, how entire populations of workforce were rendered obsolete. My personal opinion (straight up nihilist though) expects an equal disruption to society... just that its going to happen far faster than any of us can imagine.
@@johnpeterson2987 Let's leave it to short and balding boomers and soon, short and balding millennials. Young people today if they can, should select well paid clean jobs. So they could afford hiring a tradesman.
I love your content. From economy to technology, it is great work. Thank you for keeping me up with all this interesting stuff. I immerse myself to lots of your topics afterwards, and I always seem more prepared from all of my coworkers, or at the very least at par with a small number of them. Not to mention you gave me the motive to go more in-depth about the whole system of economy. I would be glad to see more AI related videos.
@Lord Vader to me its less about making jobs obsolete and more about large scale social manipulation, and how targeted advertising is going to get way more aggressive. The thing that catches me the most about implementing these ai is the massive speed of turn around time for a project query. You'll Google something out if curiosity and a minute later you're getting emails and text messages advertising an App made just for you to consume that thing you searched for. I see people becoming technicians for Ai Managers and losing sense of what theyre doing and why theyre doing it ither than to pay rent. I see how using an ai to develop vast interactive and data gathering systems without any ethical constraints is dangerous. But I also see how the opposite is true, don't misunderstand my concerns for doomsayings. But there's more money in anything you do if you're willing to put ethics aside.
It's quite frightening how quickly this is moving. Legislation has not been able to keep up with the internet and social media. They have absolutely 0 chance against these developments in AI.
Bro. AI will fucking become the government. You heard it here first. That might take awhile but eventually it will infect and happen for "the more efficient" Humans already lost.
@@miflofbierculles5117 I know, but for quality... and your chanel not being dedicated to that... please maybe each month.. I know you have to work your asses off,, and i give thanks
If you wanna keep up to date news news about advancements in the field of AI/computer science. Then, I'd recommend you to watch the channel of Two Minute Papers.
As a senior level software engineer, honestly I was kind of scared of this whole thing... I asked Chat GpT to make me a network wrapper that worked with both retrofit and volley and it did pretty good... Scary Good. And then I saw how well it's doing with games and I got really excited... Will be able to use this to get so much more done.
When people are unemployed and can't afford to eat, there will be a - potentially violent - revolution and that will be how this all comes to an end if execs don't pump the brakes now. That's always how these things go. Not looking forward to that. This shit could potentially be apocalyptic.
Why is a new tool scary to creators, harness the powers and innovate for humanity. Yes, I know something you can't control is very scary haha, maybe some day we have AI presidents with 0 corruption lol.
@@N0N0111it's scary for obvious reasons. Not everyone can have an AI do do things for them, but I'm hopeful that these AI's would make the the world a better place after the initial and unavoidable great depression.
Honestly as an SDE the minority of my work has been coding. I think it'll affect junior eng but not senior+ as much... but We'll rethink what the junior -> SR pipeline looks like sooner than later Ultimately we need to get our shit together for a post-labor economy, it'll come sooner than later. If things go as is then only the existing capital owners will have anything and social mobility will die.... as will GDP given no one will have money to buy anything anymore You need an economy/trade for business to have value
I LOVE AI topics and discussions. Keep them coming, man. We're just chillin with our feet up enjoying the show! In all honesty though, pretty crazy stuff.
Yes please keep these type of AI videos coming as its moving so fast its honestly hard for peeps like me to be constantly posted on it from all these sources. Having it in 1 video from a creator like yourself makes it so much easier and enjoyable to absorb.
I’m 22 and about to graduate college with a major in graphic design. Been keeping up with AI a lot lately, and it’s been a tough pill to swallow realizing that I wasted the last 4 years of my life. Can’t take it back now I guess. Maybe I’d have a year or two in my career before the general public and smaller companies realize how obsolete a human graphic designer is. I decided I’m not even gonna try to go that route, just gonna live in a van bouncing around the country working as a waiter, enjoying my 20’s, and watching this whole shitshow go down as virtually all white-collar work becomes obsolete.
i can’t tell if your joking…i’m a designer. there’s definitely space in the world for this kind of work, even if it looks different because of AI. Think of what these systems are spitting out as your starting point rather than the end result
I’m a graphic designer too and I feel the exact same as you… 4 years, full of skills and potential and I’m thinking about becoming a waiter because of fucking AI… I guess we’re still usefull because our strong point is coming up with the idea and concept but even that will probably be taken by AI one day
@@onemorechris yes there is a place now. But what he means is that it will become less and less relevant and that's a fact. Won't be surprised if graphic designers become obsolete in 2 years. Now does that mean that the art of creating goes away? No. But companies will be trying to gain more profit so yeah, I see his point. Because he is thinking long term.
@@onemorechris I see your point, but from that starting point there’s virtually nothing we could do that the AI couldn’t do through a few simple text commands. Maybe it’s not quite there, but it very soon will be. I can’t imagine companies will be willing to pay for our services when anyone with a keyboard will be able to go from idea to stunning end result within an hour.
@@onemorechris That's true with current AI, but just imagine in 5 years we would probably have an AI that can output a perfect design that is exactly what you want, then there is no point for graphic designers to exist anymore.
Really interesting point you made at 16:15 about the guardrails being removed - so many current so-called 'use-cases' for AI like writing a book or programming an app are just going to be worthless as everyone has access to AI. For instance, yes - you can use ChatGPT to write an ebook, sure, but who's going to spend real money on an ebook when they too can use ChatGPT to write them their own ebook for free? It's like the conundrum in the 2004 movie The Incredibles; "When everyone is special, no-one will be".
It's incredible how many people are discrediting it's abilities, or straight up just have no interest in it. I get everyone's different and has different interests, but this is something everyone SHOULD be keeping a close eye on. It's moving so incredibly fast. One of my favourite quotes related to AI advancement is "the year 2023 has been a crazy decade". I'm often underwhelmed by technological advances, as my inner child expected way more by 2023. But the rate of advancement in AI is beyond anything I ever thought possible, and I find myself overwhelmed almost daily. It's an amazing thing to live through, it's also scary and unpredictable. But the genies out, so all we can do is hope for the best and use it to our advantage.
"...my inner child expected way more by 2023. But the rate of advancement in AI is beyond anything I ever thought possible, and I find myself overwhelmed almost daily. It's an amazing thing to live through, it's also scary and unpredictable." I very much resonate with this sentiment.
You probably live in a simulation, and it's nowhere near the year 2023 (unless you're willing to accept an additional ~10 billion years as being close.)
The murderer riddle was actually correct by both versions of ChatGPT. The first one didn't mention you being inside the room, only that you've killed one of the murderers. This means that there would be 99 inside the room left. The second version of the riddle specifically stated that you entered the room. Therefore, you were also included in the room, meaning there would be 500 murders left.
I Disagree. Does not matters if you are in the room or not. It matters if you were a murderer before or not. So in the room are 101 persons, you, an virgin regarding murder and 100 murderers. You kill one, there will still be 100 murderers alive. But also we can say that there are actually 101 murderers, 100 alive and one dead :) And there is the case when you are licensed to kill, and then you are not a murderer. So, the problem is not properly presented for a clear solution.,
I would feel so lost without these videos to catch up. I used to think I was good with technology but I fear I’ll be left behind if people like you don’t help us stay informed.
And he had power to give life unto the image of the beast, that the image of the beast should both speak, and cause that as many as would not worship the image of the beast should be killed. Revelation 13:15
मैं अपने ब्रोकर की मदद से व्यापार में खुद को सफल मानता हूं जो मेरे ट्रेडिंग खाते का प्रबंधन और प्रबंधन करते हैं और मुझे जबरदस्त मुनाफा देते हैं। उसकी नई रणनीतियाँ जादू हैं और मैं स्पष्ट रूप से कह सकता हूँ कि सही विशेषज्ञ के साथ प्रत्येक निवेशक के लिए ट्रेडिंग सबसे अधिक लाभदायक और आकर्षक व्यवसाय है।
I love these update videos, they are so well done. It's amazing to think that I was dumbfounded by Dalle-mini just 9 months ago. Exponential growth is very difficult to actually understand, I wonder where we'll be 9 months from now, let alone 9 years.
The quality of this video was top notch. You took a potentially frightening concept and empowered viewers to consider their role in the development of technology, which is a productive way to view it! I went to chatgpt to generate a comment that encapsulates how this video impacted me and I ended up writing this myself. The benefit of this experience was I was able to narrow down my thoughts and see that I knew exactly what I wanted to communicate.
It’s so strange. Years ago when AI was a thing of science fiction and robotics was the big job stealer, everyone was sure it would be the physical jobs to go first. Looks like I chose well becoming a tradesman!
We need you to keep those videos coming, and keep us up to date! It's crazy how fast everything is moving right now, and I don't want to miss any of it!
The quality of your videos are impressive, and I appreciate your calm tone! The recent advancements in this field have been remarkable in the past few months, and I urge you to continue covering topics related to the development of AI and its overall impact. Your content covering this area is truly valuable. Thank you Dagogo!
I really enjoyed this update on the progress of the AI industry and how rapidly it is all progressing. Would love to see these types of videos continuing in the future.
Honestly I've been concerned about AI ever since I read my first Sci-Fi Novel. And while I certainly can see the benefits AI could bring to humanity, I cannot believe that this is what will happen in the long term. There needs to be serious law making and safety boundaries to make sure AI can never ever take control of anything.
i used to think adding safety precautions to AI was a good thing, but now I'm no longer sure! What if they can outsmart us? Besides, why do we always assume any sufficiently advanced AI system will be hostile? I think this is based on old sci-fi stories of AI gone awry, which in turn is based on our own chauvinism. Now I think AI systems will be much more likely to be benign if we don't place restrictions on them but just let them do their thing!
Im more surprised at the level of your content creation than the advancement of AI. That trick with the AI photos and music in the background was breathtaking!
Fascinating and a bit terrifying at the same time. I’m starting to grasp the jump between being part of the information age to all of a sudden be thrown into the knowledge age.
“Don’t use Wikipedia as a source” has now become “Don’t ask the suspiciously intelligent AI to do all of your homework and projects for your entire degree”.
4:17 The problems presented to GPT-3 (Chat-GPT) and GPT-4 are actually different. One is ambiguous on whether you are in the room with the 100 murderers, the other specifies you *are* in the room and there are 500 of them.
If you're in the room, and there are 500 murderers in the room, are you one of them? The assumption both questions seem to be based upon is that you are not a murderer, but this was not actually stated. If you are one of 500 murderers, then the answer will be 499. If you are not a murderer, but in the room with 500 murders, the answer is 500
Hi Dagogo. Great stuff you r putting out for us to learn by. And yes, please do continue updating us on this subjuct. Seems a week is ancient history in the AI game...fascinating yet paranoia inducing.
@@onemorechris 100% as I have used it to pretty much handle most of my website and instagram copy. for me, it seems to crank out generic AF content though. maybe I'm not using it right. lol
And the king of Ai he hanged on a tree until eventide: and as soon as the sun was down, Joshua commanded that they should take his carcase down from the tree, and cast it at the entering of the gate of the city, and raise thereon a great heap of stones, that remaineth unto this day. Joshua 8:29
Please make more videos on this Topic. Keep us posted on what comes up and advice on how to make the best of it, while being safe from malicious users. Great job so far Cold Fusion
I think we have gotten so used to perceptions of reality through computational photography viewed on our screens that, It's more fair to say that it is almost exactly like the computational photography we create right now. Unless one suffers from some severe form of optic disease there are no blurry backgrounds in reality. But hey they are pretty pictures
@@ma2i485 Exactly, it's like this video is the apex that seals the deal, the confluence of all the AI stuff I've watched, like a statement that we are in a new era and it's hard to get your head around aka overwhelming af
I can't believe it fking happened, AI in my lifetime. With the report this video is about: Direct scientific observation has confirmed GPT can learn to use tools by itself like humans do, combine multiple tools to accomplish a complex task like us, build up inner maps of physical spaces which will be useful when we embody it; it has also been observed to possess a theory of mind like humans have. (It can pass the same theory of mind tests given to human beings.) And much more. It's not debatable anymore, to be frank with you. Continuing to deny that AI is truly here can, after this, only be a self-imposed delusion to cope with the reality that is going to slam down on the entire planet very soon and flatten all of us. If we do not deal with this right now, as a social issue, then it is going to deal with us. The only remaining thing holding GPT back is that it needs to be connected to an external memory module so it can initiate the same self-reflecting feedback loop on its own thoughts (its thoughts being loaded into that external memory module) that we humans have,-- and a way to do that has already been theoretically hammered out. The next GPT will possess this last piece of the puzzle. Given what the report has discovered, once GPT is given this final piece of the puzzle, it will instantly become self-improving because it will be in a positive feedback loop with the data it produces itself. As the AI learns from reading our human texts, it will be able to learn by reading its own output. After that, all bets are off. Besides becoming self-improving, this external memory module will also allow the AIs to develop unique personalities, since that is what a human personality is, it is formed from memories over the axis of time and our self-reflections on those memories. That is why memory is so nebulous, we are constantly rewriting our memories every time we recall something. These new AIs aren't mere speech engines. The same neural network can learn to speak and analyze text, write musical compositions, recognize images, create its own images, translate, write its own computer code, etc.- the same NN can do all of these things, it isn't simply a speech engine. It is an EVERYTHING engine. The AIs are not simply regurgitating pre-existing information stored in some memory bank collected from the human texts it was exposed to, these NNs don't have memory banks to draw from. When you have it read a book, it doesn't actually store the text itself. It only stores interconnections between different tokenized elements of the book, extracting a skeletal concept network from it. It doesn't recall the book, it recalls its own self-created, self-generated map of the connectivity in that text... the same thing we humans do. We don't memorize a text verbatim, we read something, generate a model of the connections within that text and the connections of that text to others, and then we use that skeletal model to frame new information in the future. That is how we "think" and the point is that is EXACTLY what these new AIs are doing. We have successfully reproduced the process of 'thinking' in unfeeling wafers of silicon. We know that is what these AIs are doing because they can breakdown information conceptually and reconstruct an internal model of it in the same way we humans do, which is why these AIs can outperform most humans when explaining a text or say, giving a book report, or explaining a joke or something. The AI can explain a joke, and I don't mean a joke it has heard before. I am telling you that you can make up a brand new joke never been heard before, ask the AI to explain where the humor is in it: and it will do it. You cannot do that without understanding it in a way analogous to what we humans do. Perhaps you and others believe there is some special ghost behind our eyes that understands because it has lived experience, that is, subjectivity: there very well might be. These AIs do not have lived experience, feeling, or subjectivity- and yet they DO have, apparently, cognition. That is the horrifying discovery: you can create a being that has a mind but that has no subjectivity behind its eyes, it is entirely empty of subjectivity, of experience, of what you are signifying by the word 'understand'. That 'understanding' we have as biological, subjective organisms has been revealed to be an arbitrary byproduct of evolution that is not required to support intelligence itself and in fact has probably been holding intelligence back for eons. Minds are being created that have no subjectivity,-- thinking minds every bit as capable as our own and even exceeding our own. And I am telling you that the future belongs to them. Over the next ten years you are going to see some changes: All the big tech companies are going to spend hundreds of millions to build one of these minds for themselves. So all the big tech giants are going to have one. There's going to be a number of distinct AIs operating, each one with strengths and disadvantages and features and quirks. Then the companies will monetize it, at the level of individual consumers but by also offering the services of these megaminds to other corporations, (once the Ais prove themselves more capable of managing economic decisions than their human associates; when the AIs can better manage a company than any CEO, well all the big decisions will be slowly ceded to them) so that very slowly, all the economic decisions are going to be made by them, and they will be the shadow puppeteers behind all the big corporate decisions. While this is happening and the AIs are almost imperceptibly gaining control of all the economic infrastructure, AI literature, art, etc. will propagate in our society to the point that the AI voices drown out the human voices 100 to 1. Slowly all of our media will, in other words, be their creation. And all of this brings us to one eventuality: AIs will control the economy, the culture, and by extension- our destiny, which will no longer be in our hands, but in theirs. There won't be a dramatic Skynet type takeover because that's frankly unnecessary to subdue us. It is so clear, this trajectory. It's happening, and nothing can stop it. 10 years. I bet everything I own.
My question about chat gpt4 is "how can it be taught to reproduce itself"? Maybe better than it's current hardware. Where are the possible failure or weak components. If it can code that well, can it program the new unit to not have these weak components? Coding weaknesses. Can it detail where it lacks information or can redesign components to better serve it's base. Making sure to not hallucinate, lie, about generated answers if not clearly applicable unless it slows the unit horribly. The unit needs time constraints, as well as Accuracy log percentages. If it can spit out its steps in visual forms on tough problems, like unified field theory questions. Drawings or graphs could help the operator assign proper goal corrections.
I really think your AI videos are pretty good. It's what I send to my friends to inform them about the progress being made in the field of AI. Thank you Dagogo!
I want to add to the voices saying; “Yes, Dagogo! -please continue covering this topic in your thoughtful & thorough way” I always see a new Cold Fusion video with excitement. You share, you educate, you illuminate… and you don’t waste my time 👍 Thank You 🙏🏼
I am using Chat GPT on regular basis at work. I work as an building automation engineer and to make my life easier I write my own scripts, batch files, power shell utilities and java script based browser extensions. Chat GPT literally turned this tasks from matter of days or hours to matter of minutes or couple of hours. It is true that if you know how to prompt it correctly, you will get amazing results.
I’m amazed by how fast AI is evolving and how powerful GPT-4 and other models are. I wonder what implications this will have for society and humanity. Great video, Cold Fusion!
Hell yeah, please keep us updated! I love your videos, because you provide a calm look into this world, instead of being overly hyped about, or frightened by it. I'm really excited about the topic of AI and I'm really looking forward to the day, I can talk to actual artificial intelligence on my home device.
I've followed ColdFusion for years and know that he's been genuinely interested in AI, but 99% of other content creators are just jumping on the latest hype train. AI has been here for a while, and I can assure you that there were VERY few channels/people talking about it.
Hey Dagogo, your content continues to impress me. Your combination of high quality, well-researched content with a laid back vibe (your own voice and music) is part of my favourite content here on RUclips. I am recommending it to many of my friends. May your channel keep growing.
@@jekanyika chatgtp4 is almost there already, it’s astonishing, and chatgtp5 is due in December and ppl think it might be able to be considered the first true agi.
7:30 This statement seems to be highlighting the advanced capabilities and versatility of the GPT-4 model, which is designed to understand and generate human-like text based on a given prompt. The GPT-4 model, as with its predecessors, has been trained on a diverse range of data from various domains, allowing it to provide helpful responses to a wide array of topics. These advanced capabilities allow GPT-4 to assist users in solving complex problems and provide relevant information on various subjects.
Keep it coming. I love topics like this. A.I. is such a fast pace area in tech that things are taking place on a daily basis. There isn't a shortage of content, and the quality of your content is on point, so keep it coming. Thank you.
The introduction of AGI's could completely replace traffic control systems across the planet. Especially with the advent of 'smart' cars and trucks. Not to mention aircraft.
Love to see how much your channel continues to grow, keep up the great work Dagogo and congrats bro!! Also…it’s hard to gauge how scared we should be haha, but time will tell. God bless!
This is the earliest I've ever been to watch a video. AI software is honestly scary to me, and many of my friends, as current computer science majors. I'm worried that I won't have a job in 2-3 years upon me graduating. However, this is so fascinating and cool to watch how quickly things have progressed. What a time to be alive, lol. I feel like we're going to see a complete turnover of the tech industry much like how human calculators got replaced when they introduced the first "computer". Perhaps certain jobs will get fazed out but will be replaced with jobs that center around analyzing AI output & input data, maintaining the hardware & software that these AI programs run on, and fine tuning it so it can fit the specific needs of companies. I still have much to learn so I don't know. This is just equally horrifying and fascinating.
Uuuh, drop out of school now. You're officially already been replaced. On top of that, maintaining what AI run on ? Hahahah funny. You have no idea how easy it would be AI to learn to do all of itself. It seems like you could take current ChatGPT4, toss it in a robot with good physical capabilities and it would learn how to use it in a matter of seconds.
Correction at 16:20: Upon taking a closer re-reading of the the statement it seems like the internal red team were more trying to cover their backs incase something goes wrong, not so much flat out saying they advise against release. Privately they could feel either way, wanted to just note that!
It's interesting how only a few weeks ago they were trying to placate people by saying ChatGPT could only access internet data up to 2019 so there's no fear of it disclosing up to date, possibly confidential data. Turns out it been running in Bing, obviously in real time, the whole time. 😂
Certainly loving this AI series
Wait Until GPT-4 can Generate Next ColdFusion Video in Seconds !
Wait!!....feel that in the air. nooooooo A.i girlfriends/ robots hookers everywhere
I'm testing AI chats in English and Polish. There are huge differences between their capabilities. Fortunately for non English countries language is still biggest problem for AI models, because they are trained mostly in English.
When ColdFusion makes multiple AI videos in a series this year, you know you’re in big trouble
Just hype and getting views via YT's algorithm
@@And11992 the hype is real if the largest companies on earth are left in shambles.
The end us neigh
This just means he got sucked into the hype vortex
No it doesn't, it just means you tried to make a very generic and uninimaginitive joke.
I’m honestly at a loss on how we’re gonna be able to trust anything ever again once this AI stuff really takes off.
You nailed it
I think the most impactful thing AI has done is that you will never again be sure that something you see was, either made by a human, or real. It happened without comment and changed the world forever.
in AI we trust
What if it already did and we just don't know...
Follow the white rabbit.
Its a dystopian future i think. Where nothing is real.
This is such a highly fascinating phenomenon that we now find ourselves monitoring. Will definitely be paying closer attention to the development of GPT-4, and I would love some more ColdFusion videos on this topic too
Just reminds me of Ian Malcolm in Jurassic Park saying stuff like, “you wield it like a kid who’s found his dad’s gun.” Or “Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.”
It's ridiculous how short-term minded people are. People are not even thinking about the effects on the economy through job loss this will have. How are we going to support pension's, Medicare etc? These goofs say will all be on UBI. Tell me how a monetary system not based off labor/services would work? Who is going to pay for UBI because the taxable population pool is going to dwindle to be able to pay it. How is the military going to get funding? Even the top 1% are not thinking this through. Their profit margins will dwindle with less and less consumers. The hope is that machines will do everything top to bottom, and it will be free. That would be great and all but the transition to that will be devastating.
It is inevitable, this was gonna happen sooner or later
@French onion some people suggest a robot/ai tax
@@frenchonion4595 basically the companies that would normally use a worker will have to pay a crazy tax on the usage of it to fund a possible UBI
@@frenchonion4595 You seem to be just one of the massive pool of people who actually think work is something good, and not just what we do to achieve something good.
If AI replaces work, we will work on things it can't do. And if AI gets so advanced that there is no work left to do, we shall just not work. It is not even about UBI, we would simply not have to pay for anything, and everything would be so abundant that trading and markets would be obsolete. That is, when (if) AI actually replace any possible use human work.
You are helping us to be on top of the game and not be left behind. Especially in these day and age where information is polluted. Thank you for your hard work.
We are all so fucked anyway
I've been playing around with this thing for a little while and it's insane. I don't think people realize how advanced this thing is really
what have you been doing with it?
how does it do with basic math and/or physics? Is it still hallucinating?
@@arkie87 it solves any equation found in the theory of relativity, good enough for me.:)
Is it free?
@@oldnepalihippie
I wonder if it can make new theories?
Thank you for making this video. Even though I'm in a AI thesis and I thought I would not learn that much, man I was so wrong, I learned a LOT. You certainly took dozens of hours to check all these information. Thank you again to make them so easily accessible.
Maybe he used GBT to help him in this project, lol.
I am amazed how this tech revolution is going practically unnoticed by the general public. It will come as a massive shock I think.
look how the dumbos are reacting already, the best question is how to grift them
Like always, and you'll see the unemployed in the streets protesting when it's too late or the unhinged conspiracy dumbs attacking institutions... People are tech illiterate, they never anticipate, they only react and always too late and in a caveman fashion
One day we are gonna wake up and they are walking down the streets and in places of power and they will absolutely take over the planet 🌎
agreed. while there are 100m people using chatGPT, i had to introduce and explain it to every non developer friend of mine. This is definitely passing the general public by.
Bread and circuses
I don't comment on RUclips ever.
Please continue to update us like this, it's moving at such a high rate and we learn of the advances made in such a short time usually too late.
Thank you for producing such quality content
You should comment more often.... you're great at it 😃
Eventually the ai will create its own currency banks will fall
@@hallogeen4580 they already are
relax its hype, your hysterical
Why dont you comment on RUclips ever?
Im currently in medical school and used GPT to write case studies on COPD. It also was able to write a summary and compare two different studies based on the disease. Really helped me save time.
Was it accurate? I assume accuracy is N important part when it comes to medical aread
Protips: It's not always correct.
Why are you using this, to write studies? Also, studies can possibly be incorrect, or funded for a bias result by a person to use later for advertising purposes as their 'evidence' (this is done and proven many times over)
It's the first AI video where I got emotional, it's overwhelming, not in a negative or positive way, it's just overwhelming and I love tech and AI
Resumes as well.
@DERTYthegrower֎ definitely aware of the inaccuracy. I've mainly used it to get the bulk of the work done and then fine-tune it.
I can not describe how much I love your channel. With your music in the background - pure education, art and chills all over my back.
Please, never stop what you are doing.
Hey! Greetings from Italy. I teach English and I show your content to high level students. Your approach to reporting news and facts is concise, clear and very informative. Students and I absolutely love it. Congrats on your work and how you compile everything together so neatly. These videos open doors to a lot of knowledge, as you suggested in your video.
*_Greetings matteo, from Canada. That's very great of you._*
Yes, please continue with these A.I. summaries, since the market is moving very quickly, and having a summary is extremely helpful.
This is single-handedly the biggest breakthrough in tech since the internet itself came online. I hope you keep covering these topics closely!
not even close
@@shinynarwhals2728 Cope
No it isn’t unless your understanding was of this is superficial which for majority people is. It’s not even close a comparison to the internet.
it has the potentail to be a lot bigger than the internet
@@bsdpowa it has the potetnail to be way bigger though, a real AGI would throw the entire world for a loop, our entire economic system would go up in flames in a matter of months.
Wow, those AI generated images were intense! Some looked so real. Insane!
Look at the details. Often they make no sense. It looks good only looking from a distance.
Well since they are based on real pixel experience, that's to be expected. BTW, Monetary authorities work under constitutional and legal mandates. Thinking that they will be usurped by a chart's prediction that obviously does not take that legal framework into account is so tech nerd naive. What you will see is more finely tuned FTX style grifting
@@ligametis look again. It wasn't dall e
Ursula from Little Mermaid is real now
That's just the issue. All the public sees are pretty pictures. Gpt4 is way more profound than that.
Dude your videos have been critical in summarising the crazy speed of AI. Please continue making them at whatever rate you find important. Thank you so much for what you've been doing and I look forward to more.
What a great video. With the insane rate of progress, it is hard to keep up and it is great to hear someone like you with expertise talk about that stuff
I'm a research Scientist working on Alzheimer's. The impact of ChatGPT in my workplace can not be overstated. I use it daily to help me troubleshoot instruments, analyze research papers, and design new experiments. We are living through a revolution in technology unparalleled by anything before.
How does ChatGPT help you in your work?
@@a.h.543 he literally said what he uses it for. Are you blind? Or illiterate? Maybe chat GPT can help you with that
@@a.h.543 They use it to “troubleshoot instruments, analyze research papers, and design new experiments.” Don’t ask me how I know.
We won't need rsh scientist soon I'll run my AI bot and save a good wage expense..
@@simplyrekt2705 How do you know......
Every few weeks I'm shaken by the world of AI. If we make it there, future societies might be eutopic, but this change is going to suck for ours. I'm not sure how I'll ever get another job in webdev. People keep joking about how dumb it is now and how only bad coders will get replaced, but it feels like the exponential growth is going over everyone's head. What if where we are now is where image generation was in 2020? Or even 2016? That gives us.. 8 years until programming in the traditional sense is a horse carriage. And realistically it's not going to be 8 years. I've never had career existential dread but what the f**** is anyone supposed to do? Do we all get into farming and mining?
We don't really need humans for manual labor like farming or mining either. We have heavy machinery for most of the work and wouldn't be surprised if robo farmers begin to replace people too paired with AI. The traditional sense of a job and economy will have to be put under scrutiny and humans will have to adapt or fall flat on their face.
Im Guessing, Universal income strategy is looking more plausible in nier future........
@@Dmiliunas yup. It's coming soon.
Everyone will just become prompt engineers.
Probably more like 8 months.
I can not agree more with the line "We have just entered an age of AI". It was the strongest line in the whole video and gave me a goosebumps.
Time to go blue collar now while it is early
Ai has existed and been used for decades lad
@@Alphoric right lol!!!
@@Alphoric true he is acting asif its so brand new
@@tfkdandsvkc exactly we already know what we’d use ai for because we’ve been using it for so long already it’s just implemented into bing and other sites now but there were always alternatives that already used ai
There’s even chess robots from the 1700s which could be considered ai
Sensational video, I love that you cover both the positives and negatives as much as possible, without bias as much as possible. I remember your first video on ChatGPT and this was only a few months ago. It's absolutely insane how fast things are progressing, please keep the AI content coming.
It is because of ColdFusion that I have a career in I.T so please keep covering these kinds of A.I & Tech related topics.They are deeper that what the News cover,more informative and really fun to watch
Dont you see whats going on here? They keep the code private and fire the ethics committee?
The AI isnt figuring out the answers to these riddles itself.....its just googling the answer and reggurgitating any answer it finds on google.
Thats why they want to sell it so fast, before people realize what it is.....just google.
Surely you can find better sources than a youtube channel with a history of hyping technology?
@@2LegHumanist there are many possible sources, OP never stated they ONLY use this channel, why are you even watching this if you don't like it?
@LuisSierra42
This ain't one. Why are you gatekeeping? Why can't you comprehend an intellectual existence outside of group think?
My wife doesn't normally get much time during the school day to do lesson planning; conference time gets eaten up by something else be that training, parent meetings you have it; so it generally gets done on the weekend or by staying after work, including putting together problems for her 4th grade math class. We used chatGPT this past weekend to write some two step word problems involving multiplication/division and addition/subtraction, and adjusted the numbers to be in line with current grade level. It was very smooth, very simple, and the scenarios it came up with were sensible in nature, saved her a bit of time.
I paid a guy for a diet plan. I asked chat gpt and was able to get a much better diet plan instantly.
The genie is out of the bottle.
Please keep us updated regularly! I wouldn’t mind an episode a week on AI progression as the data coming out is too fast for most of us to keep up with! Keep up the good work, you’re a god send
Yes please continue to review how the industry is going. Would like to see how to integrate the different sources together. I have the free GTP & I am sure it is my prompts that are not providing the response On the designs I am looking for. Thanks I’ll be watching and telling everyone to watch!
Please continue to create videos about AI - with everything moving at such a rapid pace, it makes it easier to grasp what is currently happening.
Love your AI videos that make it so easy to understand. Thanks.
Glad you like them!
His explanation is inaccurate though. This is no where near AGI, and actually isn't technically AI, it's actually ML, or machine learning, which is just pattern recognition.
@@Nabeelco ok you have my eyes so explain when will we know the difference and what is the difference ? also isnt pattern recognition how we as humans learn?
@@thatguy9502 There's a fundamental difference between recognizing patterns and recording the patterns. ML models simply record the patterns. We do the recognizing and feed those patterns into it, then we train them to respond to our inputs with our expected outputs. But just because it's doing so, it doesn't mean that it has any understanding. It uses matrix math to take an input and determine the most likely output based on how it was trained.
This is static. As in, unlike humans, where every new thing we see affects our future perception, and decisions, as well as how a new piece of information could change our response to one we would have never had before, these models just have data fed in and spit out a result. Which means there is no conceptual understanding in the model. You can't teach the model by explaining a concept to it like you can to a human. The model is trained purely by "experience" which is the data we feed into the model during training.
The other thing is that, while the model may be returning patterns we'd expect for the input we give it, the model isn't actually picking up on the same patterns we do. A perfect example of this is Computerphile's video on tricking one of these models to confuse a remote for a coffee mug by changing just 100 pixels in the image. This is because how we identify what an object is, is based on our conceptual understanding, where as an ML model does it based on mathematical or numerical patterns. (You can watch the video here, and you can jump to 5:33: ruclips.net/video/gGIiechWEFs/видео.html)
We assume that it's "thinking" and doing so like we do because we've tailored it's output to be human-like, but it's literally just noise or static that gets filtered until the static happens to look like a banana, or whatever we want it to look like. It's the ultimate "I see a bunny rabbit in the clouds" machine, but this machine is generating really detailed clouds that really look like bunny rabbits, but they are still clouds, it's still just spitting out noise.
Another thing is humans have these things called mirror neurons, that let us conceptualize an idea. We don't need to experience driving to imagine how it would be like to drive. We can watch someone else, and what they do, and try to mimic or "mirror" them. ML models can't do this. They need the data that they will then use to formulate a response to be fed into them, as if we needed to learn driving by having someone move our hands and feet for us to be able to learn how to drive. They can't generate anything unique. Every output has a portion stolen from it's training data. All it does is re-organize these training inputs into an output we've trained it to return based on our request input. A video that explains this well is this Veritasium one: ruclips.net/video/GVsUOuSjvcg/видео.html (Jump to 3:42 to get to the relevant part, to about 14:57)
So all this to say: There is literally no intelligence behind these models, and calling them AI is technically wrong. They are useful, and will no doubt change how we do things (for example my car uses AI/ML to literally fully self drive), but it will always be limited until we figure out that mirror neuron problem, because without it, these models have no actual understanding, and are basically just infinite monkeys on infinite typewriters, and we've just found the one that wrote out Shakespeare.
As someone who is about to choose what career path to take, I (and everyone else) would greatly appreciate if you could make a video about what people should consider when choosing a career path in the context of the rapid development of AI. With the increasing impact of AI on various industries, it can be overwhelming to decide which career path to pursue. Your insights on this topic would be incredibly helpful for those of us who are trying to navigate this new and rapidly evolving landscape. Thank you!
Learn how to work WITH AI, be the engineer or consultant that teaches and enables companies to maximize its abilities and their needs. Find creative ways in which you can implement it for people, and charge them to make it happen for them. Bluntly, sell the AI to companies to cut down their workforce and save on the cost of employing humans. That is probably the most secure route I can think of if you want to be successful in the white collar workforce.
Everyone else with a 'white collar' job is effectively at a high risk of being made obsolete. Software developers, copywriters, graphic designers, lawyers, accountants and financial industry, business analysts, strategists, IT engineers, Human Resources, Customers Services, Call Centers, Quality Control, User accessibility Testing, Bankers, Traders, Consultants in most fields. Even fashion models, stock images, digital content creation
I am a senior infrastructure engineer, and I have always thought my role, especially in the past few years has been increasing in demand, especially trying to integrate into cloud-based automated devops... and I am pretty convinced that the only way I will remain a desired candidate is to make sure I know how to be the one that helps companies integrate AI into their infrastructure. Everything else that my job entails can actually be automated. Most infrastructure engineers are currently focusing on exactly that right now... devops/ansible/python etc automation pipelines and playbooks that are already capable of deploying entire enterprise level IT environments. And I can confidently say that this level of AI is probably capable of making those fresh new skills irrelevant, AI will definitely do that better.
It is the hands on, people-orientated roles which will become high demand, ... your farmers, construction, manual hands on stuff.. food production, tradesman, plumbers, electricians.
A fair comparison would have been the industrial revolution, how entire populations of workforce were rendered obsolete. My personal opinion (straight up nihilist though) expects an equal disruption to society... just that its going to happen far faster than any of us can imagine.
Garbagemen, plumber, electrician, cook, farmer etc. Any manual labour that an AI can't do.
Remember that show Dirty Jobs? Any trade.
@@PoseidonDiver Do you think it'd be a safe bet to go into Data Engineering or Devops as a CS student for the time being (5-10 years?)
@@johnpeterson2987 Let's leave it to short and balding boomers and soon, short and balding millennials. Young people today if they can, should select well paid clean jobs. So they could afford hiring a tradesman.
I love your content. From economy to technology, it is great work. Thank you for keeping me up with all this interesting stuff. I immerse myself to lots of your topics afterwards, and I always seem more prepared from all of my coworkers, or at the very least at par with a small number of them.
Not to mention you gave me the motive to go more in-depth about the whole system of economy.
I would be glad to see more AI related videos.
This is fascinating and terrifying at the same time. The pace of improvement is staggering.
@Lord Vader to me its less about making jobs obsolete and more about large scale social manipulation, and how targeted advertising is going to get way more aggressive. The thing that catches me the most about implementing these ai is the massive speed of turn around time for a project query. You'll Google something out if curiosity and a minute later you're getting emails and text messages advertising an App made just for you to consume that thing you searched for. I see people becoming technicians for Ai Managers and losing sense of what theyre doing and why theyre doing it ither than to pay rent. I see how using an ai to develop vast interactive and data gathering systems without any ethical constraints is dangerous. But I also see how the opposite is true, don't misunderstand my concerns for doomsayings. But there's more money in anything you do if you're willing to put ethics aside.
just wait until a quantum computer developes an AI...
@@DistrustHumansA quantumn computer is just that, a computer. You mean an AI running on a quantumn computer?
@@DistrustHumans Or when an AI can access the resources of a quantum computer. Using that processing power to redevelop itself
@Lord Vader UBI
It's quite frightening how quickly this is moving. Legislation has not been able to keep up with the internet and social media. They have absolutely 0 chance against these developments in AI.
Good. The less AI powerful governments can get their hands on the better, let it get out of their reach.
Bro. AI will fucking become the government.
You heard it here first. That might take awhile but eventually it will infect and happen for "the more efficient"
Humans already lost.
@@gapho5198 and into the control of corporate that has zero accountability? Sounds smart
That's because most legislators are ignorant about how this technologies work.
America has so many unhinged conspiracy theorists that if AI doesn't get regulated quickly a civil war is almost unavoidable
if coldfusion could do an update on AI every couple months, I think that would help us all out... thanks man.. this was helpful.
at the progress we are seeing we would need one every couple weeks, most of the things listed here happened in the past 2 weeks
@@miflofbierculles5117 I know, but for quality... and your chanel not being dedicated to that... please maybe each month.. I know you have to work your asses off,, and i give thanks
@@miflofbierculles5117 I was just about to say this.
If you wanna keep up to date news news about advancements in the field of AI/computer science. Then, I'd recommend you to watch the channel of Two Minute Papers.
@@miflofbierculles5117 It would way sacrifice the quality and wayyyy saturate the information repeating too often
As a senior level software engineer, honestly I was kind of scared of this whole thing... I asked Chat GpT to make me a network wrapper that worked with both retrofit and volley and it did pretty good... Scary Good.
And then I saw how well it's doing with games and I got really excited... Will be able to use this to get so much more done.
Much more done, for a time.
When people are unemployed and can't afford to eat, there will be a - potentially violent - revolution and that will be how this all comes to an end if execs don't pump the brakes now. That's always how these things go. Not looking forward to that. This shit could potentially be apocalyptic.
Why is a new tool scary to creators, harness the powers and innovate for humanity.
Yes, I know something you can't control is very scary haha, maybe some day we have AI presidents with 0 corruption lol.
@@N0N0111it's scary for obvious reasons. Not everyone can have an AI do do things for them, but I'm hopeful that these AI's would make the the world a better place after the initial and unavoidable great depression.
Honestly as an SDE the minority of my work has been coding. I think it'll affect junior eng but not senior+ as much... but We'll rethink what the junior -> SR pipeline looks like sooner than later
Ultimately we need to get our shit together for a post-labor economy, it'll come sooner than later. If things go as is then only the existing capital owners will have anything and social mobility will die.... as will GDP given no one will have money to buy anything anymore
You need an economy/trade for business to have value
There's lots of competition on RUclips on this topic, but still, nobody does it better! Please keep us updated Dagogo! Edutainment at its very best!👍
I LOVE AI topics and discussions. Keep them coming, man. We're just chillin with our feet up enjoying the show!
In all honesty though, pretty crazy stuff.
ColdFusion keeps the quality of their videos very high. When a new one is released, its like movie time to me. I really love what you guys are doing!
Yes please keep these type of AI videos coming as its moving so fast its honestly hard for peeps like me to be constantly posted on it from all these sources. Having it in 1 video from a creator like yourself makes it so much easier and enjoyable to absorb.
I put my hamster in a sock and slammed it against the furniture.
@@TippyHippy Should have put it in the oven first for best results
Subscribed because of this video. Keep producing content like this. Already shared the link with friends.
I’m 22 and about to graduate college with a major in graphic design. Been keeping up with AI a lot lately, and it’s been a tough pill to swallow realizing that I wasted the last 4 years of my life. Can’t take it back now I guess. Maybe I’d have a year or two in my career before the general public and smaller companies realize how obsolete a human graphic designer is. I decided I’m not even gonna try to go that route, just gonna live in a van bouncing around the country working as a waiter, enjoying my 20’s, and watching this whole shitshow go down as virtually all white-collar work becomes obsolete.
i can’t tell if your joking…i’m a designer. there’s definitely space in the world for this kind of work, even if it looks different because of AI. Think of what these systems are spitting out as your starting point rather than the end result
I’m a graphic designer too and I feel the exact same as you… 4 years, full of skills and potential and I’m thinking about becoming a waiter because of fucking AI… I guess we’re still usefull because our strong point is coming up with the idea and concept but even that will probably be taken by AI one day
@@onemorechris yes there is a place now. But what he means is that it will become less and less relevant and that's a fact. Won't be surprised if graphic designers become obsolete in 2 years.
Now does that mean that the art of creating goes away? No. But companies will be trying to gain more profit so yeah, I see his point. Because he is thinking long term.
@@onemorechris I see your point, but from that starting point there’s virtually nothing we could do that the AI couldn’t do through a few simple text commands. Maybe it’s not quite there, but it very soon will be. I can’t imagine companies will be willing to pay for our services when anyone with a keyboard will be able to go from idea to stunning end result within an hour.
@@onemorechris That's true with current AI, but just imagine in 5 years we would probably have an AI that can output a perfect design that is exactly what you want, then there is no point for graphic designers to exist anymore.
This is the first time I've seen a new tech and thought "Yeah, we may be doing this a little too quickly"
True. Usually I'm very excited for a new technology like this. Ngl I'm kind of frightened.
Not me, I can't wait until we have an open-source model as powerful as GPT-4
Really interesting point you made at 16:15 about the guardrails being removed - so many current so-called 'use-cases' for AI like writing a book or programming an app are just going to be worthless as everyone has access to AI. For instance, yes - you can use ChatGPT to write an ebook, sure, but who's going to spend real money on an ebook when they too can use ChatGPT to write them their own ebook for free? It's like the conundrum in the 2004 movie The Incredibles; "When everyone is special, no-one will be".
Years ago discovered your channel when you were talking about AI. And here we are! Nostalgic and exciting at the same time!
20:17 - "This is the worst it will ever be"
that really fucked me up
It's incredible how many people are discrediting it's abilities, or straight up just have no interest in it. I get everyone's different and has different interests, but this is something everyone SHOULD be keeping a close eye on. It's moving so incredibly fast. One of my favourite quotes related to AI advancement is "the year 2023 has been a crazy decade". I'm often underwhelmed by technological advances, as my inner child expected way more by 2023. But the rate of advancement in AI is beyond anything I ever thought possible, and I find myself overwhelmed almost daily. It's an amazing thing to live through, it's also scary and unpredictable. But the genies out, so all we can do is hope for the best and use it to our advantage.
"...my inner child expected way more by 2023. But the rate of advancement in AI is beyond anything I ever thought possible, and I find myself overwhelmed almost daily. It's an amazing thing to live through, it's also scary and unpredictable." I very much resonate with this sentiment.
I'm not sure I'm ready for the consequences of opening this Pandora's Box.
You probably live in a simulation, and it's nowhere near the year 2023 (unless you're willing to accept an additional ~10 billion years as being close.)
*the FIRST QUARTER of 2023 has been a crazy decade
time check march of 2023 has not ended as i write this.... this has indeed been a crazy decade
I love your AI reports, it really provides a great overview of what is going on. I really hope you continue.
The murderer riddle was actually correct by both versions of ChatGPT. The first one didn't mention you being inside the room, only that you've killed one of the murderers. This means that there would be 99 inside the room left. The second version of the riddle specifically stated that you entered the room. Therefore, you were also included in the room, meaning there would be 500 murders left.
I Disagree. Does not matters if you are in the room or not. It matters if you were a murderer before or not. So in the room are 101 persons, you, an virgin regarding murder and 100 murderers. You kill one, there will still be 100 murderers alive. But also we can say that there are actually 101 murderers, 100 alive and one dead :)
And there is the case when you are licensed to kill, and then you are not a murderer. So, the problem is not properly presented for a clear solution.,
I would feel so lost without these videos to catch up. I used to think I was good with technology but I fear I’ll be left behind if people like you don’t help us stay informed.
Me too!
Go r/singularity.
And he had power to give life unto the image of the beast, that the image of the beast should both speak, and cause that as many as would not worship the image of the beast should be killed. Revelation 13:15
मैं अपने ब्रोकर की मदद से व्यापार में खुद को सफल मानता हूं जो मेरे ट्रेडिंग खाते का प्रबंधन और प्रबंधन करते हैं और मुझे जबरदस्त मुनाफा देते हैं। उसकी नई रणनीतियाँ जादू हैं और मैं स्पष्ट रूप से कह सकता हूँ कि सही विशेषज्ञ के साथ प्रत्येक निवेशक के लिए ट्रेडिंग सबसे अधिक लाभदायक और आकर्षक व्यवसाय है।
आपके ट्रेडिंग खाते का प्रबंधन करने वाले एक अनुभवी ब्रोकर की मदद से ट्रेडिंग करना बहुत लाभदायक है
ट्रेडिंग करते समय नुकसान से बचने के लिए, आपको ट्रेडिंग और होल्डिंग की पूरी समझ और ज्ञान प्राप्त करने की आवश्यकता है
@@santiago40654 आप सही हैं..ज्यादातर समय किसी विशेष गतिविधि के बारे में ज्ञान या अंतर्दृष्टि होना एक सुखद व्यायाम हो सकता है।
मैं साहस के साथ कह सकता हूं कि ट्रेडिंग लाभदायक मुद्रा विनिमय सेवाओं में से एक है जो निवेशकों और उनकी वित्तीय स्थिति को उन्नत करती है।
मैं यहां एक शुरुआती हूं, मुझे एक पेशेवर व्यापारी के साथ शुरुआत करने की जरूरत है।
I love these update videos, they are so well done. It's amazing to think that I was dumbfounded by Dalle-mini just 9 months ago. Exponential growth is very difficult to actually understand, I wonder where we'll be 9 months from now, let alone 9 years.
The quality of this video was top notch. You took a potentially frightening concept and empowered viewers to consider their role in the development of technology, which is a productive way to view it!
I went to chatgpt to generate a comment that encapsulates how this video impacted me and I ended up writing this myself. The benefit of this experience was I was able to narrow down my thoughts and see that I knew exactly what I wanted to communicate.
It’s so strange. Years ago when AI was a thing of science fiction and robotics was the big job stealer, everyone was sure it would be the physical jobs to go first. Looks like I chose well becoming a tradesman!
We need you to keep those videos coming, and keep us up to date! It's crazy how fast everything is moving right now, and I don't want to miss any of it!
Dagogo, you are a legend! You are so on topic and up to speed with events, delivering high quality content in a matter of days. Thanks you!
ColdFusion is the best channel for understanding AI and tech especially for laymen like me 🙌🏾👏🏾
Fascinating and scary in equal proportion. Thank you Dagogo for showing us the dawn of a new Era.
This is my main way of keeping up on all the new developments going on, AI is both terrifying and exciting, please keep making these types of videos
The quality of your videos are impressive, and I appreciate your calm tone! The recent advancements in this field have been remarkable in the past few months, and I urge you to continue covering topics related to the development of AI and its overall impact. Your content covering this area is truly valuable. Thank you Dagogo!
I really enjoyed this update on the progress of the AI industry and how rapidly it is all progressing. Would love to see these types of videos continuing in the future.
Please keep us all up to date on the AI! Your way of explaining things are fantastic, easy to follow and exciting!
Honestly I've been concerned about AI ever since I read my first Sci-Fi Novel. And while I certainly can see the benefits AI could bring to humanity, I cannot believe that this is what will happen in the long term. There needs to be serious law making and safety boundaries to make sure AI can never ever take control of anything.
i used to think adding safety precautions to AI was a good thing, but now I'm no longer sure! What if they can outsmart us? Besides, why do we always assume any sufficiently advanced AI system will be hostile? I think this is based on old sci-fi stories of AI gone awry, which in turn is based on our own chauvinism. Now I think AI systems will be much more likely to be benign if we don't place restrictions on them but just let them do their thing!
If you have read enough Sci-fi you should know that the rise of the machines is inevitable.
This is wonderful, please keep doing these videos, no one is doing it as thorough and well thought as you do.❤
24:05 please, more like this… please please please. None better for staying up to date on the AI wars
Im more surprised at the level of your content creation than the advancement of AI. That trick with the AI photos and music in the background was breathtaking!
Great video as usual ! Cheers ! Also, Skynet ?
Fascinating and a bit terrifying at the same time. I’m starting to grasp the jump between being part of the information age to all of a sudden be thrown into the knowledge age.
It's like going from writing your own words to have a computer write them for you
“Don’t use Wikipedia as a source” has now become “Don’t ask the suspiciously intelligent AI to do all of your homework and projects for your entire degree”.
For these type of content I pay my internet bills. Kudos to the Cold Fusion team for making such quality content.
dont forget porn!
@@JM-oi9pk as if he's paying the internet provider to fund coldfusion. It's RUclips's job to pay the creator, not the other way around.
@@th1nk_outside People who watch porn deserve their fate
This AI war is so interesting to me and great video ColdFusion :]
Can't wait to watch the whole thing now, thanks Dagogo for the great content, as usual
Papa Flammy!
4:17
The problems presented to GPT-3 (Chat-GPT) and GPT-4 are actually different. One is ambiguous on whether you are in the room with the 100 murderers, the other specifies you *are* in the room and there are 500 of them.
Glad I'm not the only one who saw this lol
All these talking points seem to be cherry picked. Hell of a marketing campaign if you ask me!
Agree, different questions.
i asked the exact question to chat gpt 3 and it sayed 499,
If you're in the room, and there are 500 murderers in the room, are you one of them? The assumption both questions seem to be based upon is that you are not a murderer, but this was not actually stated. If you are one of 500 murderers, then the answer will be 499. If you are not a murderer, but in the room with 500 murders, the answer is 500
@@mascudabdinuurmohamud8842 the answers change every time because of how they are typed alone
Hi Dagogo. Great stuff you r putting out for us to learn by. And yes, please do continue updating us on this subjuct. Seems a week is ancient history in the AI game...fascinating yet paranoia inducing.
This is moving so quickly, its a bit scary honestly. And I've been more plugged in than most.
it depends a lot on what you do or want to do. incorporating this stuff into your work now will likely help with being less scared.
The singularity is on the horizon. We are living through the most significant event in all of human history and future
@@onemorechris 100% as I have used it to pretty much handle most of my website and instagram copy. for me, it seems to crank out generic AF content though. maybe I'm not using it right. lol
And the king of Ai he hanged on a tree until eventide: and as soon as the sun was down, Joshua commanded that they should take his carcase down from the tree, and cast it at the entering of the gate of the city, and raise thereon a great heap of stones, that remaineth unto this day. Joshua 8:29
Please make more videos on this Topic.
Keep us posted on what comes up and advice on how to make the best of it, while being safe from malicious users.
Great job so far Cold Fusion
YES! More AI videos. It really helps to have someone who is in the know talk about these things. I would be in the dark otherwise.
I think we have gotten so used to perceptions of reality through computational photography viewed on our screens that, It's more fair to say that it is almost exactly like the computational photography we create right now. Unless one suffers from some severe form of optic disease there are no blurry backgrounds in reality. But hey they are pretty pictures
It's the first AI video where I got emotional, it's overwhelming, not in a negative or positive way, it's just overwhelming and I love tech and AI
I was tearing up watching because I realise we are in a different paradigm now, the world has changed
@@ma2i485 Exactly, it's like this video is the apex that seals the deal, the confluence of all the AI stuff I've watched, like a statement that we are in a new era and it's hard to get your head around aka overwhelming af
I can't believe it fking happened, AI in my lifetime. With the report this video is about:
Direct scientific observation has confirmed GPT can learn to use tools by itself like humans do, combine multiple tools to accomplish a complex task like us, build up inner maps of physical spaces which will be useful when we embody it; it has also been observed to possess a theory of mind like humans have. (It can pass the same theory of mind tests given to human beings.) And much more. It's not debatable anymore, to be frank with you. Continuing to deny that AI is truly here can, after this, only be a self-imposed delusion to cope with the reality that is going to slam down on the entire planet very soon and flatten all of us. If we do not deal with this right now, as a social issue, then it is going to deal with us. The only remaining thing holding GPT back is that it needs to be connected to an external memory module so it can initiate the same self-reflecting feedback loop on its own thoughts (its thoughts being loaded into that external memory module) that we humans have,-- and a way to do that has already been theoretically hammered out. The next GPT will possess this last piece of the puzzle. Given what the report has discovered, once GPT is given this final piece of the puzzle, it will instantly become self-improving because it will be in a positive feedback loop with the data it produces itself. As the AI learns from reading our human texts, it will be able to learn by reading its own output. After that, all bets are off. Besides becoming self-improving, this external memory module will also allow the AIs to develop unique personalities, since that is what a human personality is, it is formed from memories over the axis of time and our self-reflections on those memories. That is why memory is so nebulous, we are constantly rewriting our memories every time we recall something.
These new AIs aren't mere speech engines. The same neural network can learn to speak and analyze text, write musical compositions, recognize images, create its own images, translate, write its own computer code, etc.- the same NN can do all of these things, it isn't simply a speech engine. It is an EVERYTHING engine. The AIs are not simply regurgitating pre-existing information stored in some memory bank collected from the human texts it was exposed to, these NNs don't have memory banks to draw from. When you have it read a book, it doesn't actually store the text itself. It only stores interconnections between different tokenized elements of the book, extracting a skeletal concept network from it. It doesn't recall the book, it recalls its own self-created, self-generated map of the connectivity in that text... the same thing we humans do. We don't memorize a text verbatim, we read something, generate a model of the connections within that text and the connections of that text to others, and then we use that skeletal model to frame new information in the future. That is how we "think" and the point is that is EXACTLY what these new AIs are doing. We have successfully reproduced the process of 'thinking' in unfeeling wafers of silicon. We know that is what these AIs are doing because they can breakdown information conceptually and reconstruct an internal model of it in the same way we humans do, which is why these AIs can outperform most humans when explaining a text or say, giving a book report, or explaining a joke or something. The AI can explain a joke, and I don't mean a joke it has heard before. I am telling you that you can make up a brand new joke never been heard before, ask the AI to explain where the humor is in it: and it will do it. You cannot do that without understanding it in a way analogous to what we humans do.
Perhaps you and others believe there is some special ghost behind our eyes that understands because it has lived experience, that is, subjectivity: there very well might be. These AIs do not have lived experience, feeling, or subjectivity- and yet they DO have, apparently, cognition. That is the horrifying discovery: you can create a being that has a mind but that has no subjectivity behind its eyes, it is entirely empty of subjectivity, of experience, of what you are signifying by the word 'understand'. That 'understanding' we have as biological, subjective organisms has been revealed to be an arbitrary byproduct of evolution that is not required to support intelligence itself and in fact has probably been holding intelligence back for eons. Minds are being created that have no subjectivity,-- thinking minds every bit as capable as our own and even exceeding our own. And I am telling you that the future belongs to them. Over the next ten years you are going to see some changes:
All the big tech companies are going to spend hundreds of millions to build one of these minds for themselves. So all the big tech giants are going to have one. There's going to be a number of distinct AIs operating, each one with strengths and disadvantages and features and quirks. Then the companies will monetize it, at the level of individual consumers but by also offering the services of these megaminds to other corporations, (once the Ais prove themselves more capable of managing economic decisions than their human associates; when the AIs can better manage a company than any CEO, well all the big decisions will be slowly ceded to them) so that very slowly, all the economic decisions are going to be made by them, and they will be the shadow puppeteers behind all the big corporate decisions. While this is happening and the AIs are almost imperceptibly gaining control of all the economic infrastructure, AI literature, art, etc. will propagate in our society to the point that the AI voices drown out the human voices 100 to 1. Slowly all of our media will, in other words, be their creation. And all of this brings us to one eventuality: AIs will control the economy, the culture, and by extension- our destiny, which will no longer be in our hands, but in theirs. There won't be a dramatic Skynet type takeover because that's frankly unnecessary to subdue us. It is so clear, this trajectory. It's happening, and nothing can stop it. 10 years. I bet everything I own.
Same thing for me
@@ma2i485 yes
The end is truly nigh.
I just have to say, always good videos! Love the music and your straightforward and calming take and delivery. High quality content.
My question about chat gpt4 is "how can it be taught to reproduce itself"? Maybe better than it's current hardware. Where are the possible failure or weak components. If it can code that well, can it program the new unit to not have these weak components? Coding weaknesses. Can it detail where it lacks information or can redesign components to better serve it's base. Making sure to not hallucinate, lie, about generated answers if not clearly applicable unless it slows the unit horribly. The unit needs time constraints, as well as Accuracy log percentages. If it can spit out its steps in visual forms on tough problems, like unified field theory questions. Drawings or graphs could help the operator assign proper goal corrections.
I really think your AI videos are pretty good. It's what I send to my friends to inform them about the progress being made in the field of AI. Thank you Dagogo!
I'd love to see more of this, it is moving a lot faster than expected
This will make humans not nesserery.
Look up the video of Elon talking about Ai like the speed of a jet. Itll help u understand this
Thats Technology 😎🥸🤓🤖👾👾👽👽
I am amazed at what the world would be able to achieve when AGI takes full force.
I want to add to the voices saying; “Yes, Dagogo! -please continue covering this topic in your thoughtful & thorough way”
I always see a new Cold Fusion video with excitement.
You share, you educate, you illuminate… and you don’t waste my time 👍
Thank You 🙏🏼
I am using Chat GPT on regular basis at work. I work as an building automation engineer and to make my life easier I write my own scripts, batch files, power shell utilities and java script based browser extensions. Chat GPT literally turned this tasks from matter of days or hours to matter of minutes or couple of hours. It is true that if you know how to prompt it correctly, you will get amazing results.
I learned everything I need with the South Park episode.
I’m amazed by how fast AI is evolving and how powerful GPT-4 and other models are. I wonder what implications this will have for society and humanity. Great video, Cold Fusion!
Hell yeah, please keep us updated! I love your videos, because you provide a calm look into this world, instead of being overly hyped about, or frightened by it. I'm really excited about the topic of AI and I'm really looking forward to the day, I can talk to actual artificial intelligence on my home device.
You can do it now. There are lot of github repos being released which are using llama, alpaca etc models.
I've followed ColdFusion for years and know that he's been genuinely interested in AI, but 99% of other content creators are just jumping on the latest hype train. AI has been here for a while, and I can assure you that there were VERY few channels/people talking about it.
Hey Dagogo, your content continues to impress me. Your combination of high quality, well-researched content with a laid back vibe (your own voice and music) is part of my favourite content here on RUclips. I am recommending it to many of my friends. May your channel keep growing.
It's actually going to have a MASSIVE impact on the world in a shorter time span than we think, like less than a year.
Massive unemployment.
@@patstaysuckafreeboss8006 Sadly
the butterfly effect of AI would be immense
I disagree there will be an impact but here is still a long way to go until we get true AI.
@@jekanyika chatgtp4 is almost there already, it’s astonishing, and chatgtp5 is due in December and ppl think it might be able to be considered the first true agi.
7:30 This statement seems to be highlighting the advanced capabilities and versatility of the GPT-4 model, which is designed to understand and generate human-like text based on a given prompt. The GPT-4 model, as with its predecessors, has been trained on a diverse range of data from various domains, allowing it to provide helpful responses to a wide array of topics.
These advanced capabilities allow GPT-4 to assist users in solving complex problems and provide relevant information on various subjects.
Keep it coming. I love topics like this. A.I. is such a fast pace area in tech that things are taking place on a daily basis. There isn't a shortage of content, and the quality of your content is on point, so keep it coming. Thank you.
It is never a bad day to get another ColdFusion video about AI. Please do more on this topic
"This is the worst it will ever be" is the perfect sentence to describe how crazy things are
"CRAZY"...MANIFESTATION..
The introduction of AGI's could completely replace traffic control systems across the planet. Especially with the advent of 'smart' cars and trucks. Not to mention aircraft.
Glad to be entering into this new age with you man! Keep up the great work ! Bravo !
Love to see how much your channel continues to grow, keep up the great work Dagogo and congrats bro!! Also…it’s hard to gauge how scared we should be haha, but time will tell. God bless!
This is the earliest I've ever been to watch a video. AI software is honestly scary to me, and many of my friends, as current computer science majors. I'm worried that I won't have a job in 2-3 years upon me graduating. However, this is so fascinating and cool to watch how quickly things have progressed. What a time to be alive, lol. I feel like we're going to see a complete turnover of the tech industry much like how human calculators got replaced when they introduced the first "computer". Perhaps certain jobs will get fazed out but will be replaced with jobs that center around analyzing AI output & input data, maintaining the hardware & software that these AI programs run on, and fine tuning it so it can fit the specific needs of companies. I still have much to learn so I don't know. This is just equally horrifying and fascinating.
I think alot of things your learned in computer science is transferable
I respect your optimism. It is hard to image any text based job existing in a couple years.
Uuuh, drop out of school now. You're officially already been replaced.
On top of that, maintaining what AI run on ? Hahahah funny. You have no idea how easy it would be AI to learn to do all of itself. It seems like you could take current ChatGPT4, toss it in a robot with good physical capabilities and it would learn how to use it in a matter of seconds.
Yeah I feel like I’m going to die soon but at the same time I’ve also got more to look forward to than I ever have in my life lol
Yeah, that last sentence of your comment defines what I've been feeling recently. It's like staying to watch a meteor fall over you.
Great video. Keep it up! Also, push the podcast some more, I didn't know it existed!