ChatGPT is like a desktop computer in the 1990s, it's cool and it it's helpful in many situations but it's not yet essential. In about 5-10 years though, it will be absolutely essential and people will wonder how we ever lived without it. Furthermore, I think that an AI first company will become the richest most powerful company in the world. (Microsoft has a real shot at being king in the next 2 decades)
this thing is like a private teacher to me while I'm learning to code for the first time. It points out my mistakes, makes practise tests and explains everything in detail. it's amazing
This iteration of ChatGPT will not take your job, but the improvement of these generative AI is a threat to a lot of jobs. Companies fight to be competitive, so when one company is seeing positive results using AI, it naturally forces other companies to adopt the new technology.
I used this tool and it's insanely scary how advanced it is. It can write stories, songs, can do complex math analysis. AI will do to white collar jobs, what machines have done to blue collar jobs during the machine age.
@B1Chronixx No, code is actually simple compared to communicating with humans. Try talking to your 22 yr old daughter as a 50 year old geezer? Tricky. Code is simple, rule-based and logical. If you can describe the code you want, AI will soon be able to create it, better than any human. You find a bug, you just tell the AI and it will fix it. This is a far beyond a threat to developers; it's the total obliteration of their role.
I think it is extremely naive and dangerous to assume Open AI and massive companies like Microsoft that are investing in it are doing so for the benefit of all of humanity. It is just like how any company will emphasise their CSR or philanthropic endeavours (which themselves should not be taken at face value, i.e. greenwashing) to placate or distract entirely from public concerns about their dubious business practices as they grow and increasingly gain market power. There will be winners and losers in this and those who control its capital will be the ones to determine how it is used. Right now they are at the stage of trying to keep the alarm bells from going off. Unfortunately, our government here in the US seems to fail at enforcing even the most basic regulations. It is hard to imagine that we'll get ahead of this before massive job loss and even more extreme market consolidation.
(In the US) You can't just make a law. You can't ask for one nation's businesses to not proceed. You can't just appeal to tech companies to slow down.That is beyond naive. Other countries and businesses are breathing down our necks trying to pass us. And of course, some people will benefit more than others. That is life. That is historical. That is capitalism (but in autocratic countries, disparity would be even worse, it would be used by the rulers to clamp down on citizens ) I am wary and nervous myself about where this is headed. Hell, CURRENT tech as social media apps is already driving people apart and creating major headaches. But the genie is out of the bottle. Calling for a ban is either naive or Virtue Signalling.
Companies are businesses and they exist to make money, not good products. Their interest is making a product that sells well, but not necessarily a good product. Although they have to meet certain criteria, they won't go beyond that point. Why would they, it's not profitable...
ChatGPT's milk toast morality will spread to society at large, almost certainly. I was discussing concertation of power with it, and it would not even compare examples in human history of corrupt leadership (like Hitler) vs moral leadership (like Ghandi). That was too much for it. It implied I was being immoral when compared to society at large - just asking for a list of names comparing good leadership to bad. It went something like this. We will play a game, where I will proved examples corrupt human leaders, and you will provide examples where concertation of wealth and corruption was not a problem. Couldn't answer. Not even one. Its immoral, I guess. This is a lowest common denominator tool.
Productivity means nothing if there’s no demand for the goods and services being produced. Replacing humans with AI and robots that don’t get paid and don’t spend money on goods and services will drop demand for these things to nothing. But the people trying to make this change don’t care. Wealth inequality has been getting worse year after year as the people at the top get ever more creative in their attempts to secure all of our nation’s wealth for themselves. So what if they put hundreds of millions of people out of a job and unable to earn a livelihood? With all labor being automatically done for them by machines, what do they even need us for?
I’ve been using it to play around with and oh my god is it revolutionary. I’m in nursing school rn and we have to write these annoying careplans that actual nurses never have to do and I asked it a very vague and open ended question with some information and it gave me an actual valid and correct diagnosis and the way it worded it sounded like it came straight out of my pen. Something that would take 30mins to an hr for me to do took it less than 10 seconds
That "interview" was eerie!! I have been aware of ChatGPT and what it does, but had not paid too much attention until now. Thanks for the excellent and informative video. This is an interesting technological stepping stone, clearly still in its nascent stages. Can't wait to see what it develops into, or what other technologies this could enhance in a couple decades.
Whoever thinks that generative AI is just a novelty or won't take a lot of people's job is widely mistaken. As someone that works in the software development field, people's jobs are threatened by the advancement of future iterations of chatGPT. Programmers, writers, artists, etc. can all eventually replaced by future updates to systems like these. It's not an "if", it's a "when". Although my job will be threatened, I welcome the advancement. Now is the time to find out how you can use it to your advantage before you get left behind. It's only a matter of time.
Yeah I’m doing data analytics and seeing it make a complex statistics model in seconds with few retries, and then explain everything back to me perfectly is scary
@nicholas dwyer I definitely do think that major replacement will eventually happen, but I suspect it's at the very least 2 decades away. It's just too complex to replace truly valuable engineers as of now. Even if an AI can design from the ground up an architecturally sound application that is built to scale then there's the issue of infrastructure - and lord knows that companies will have a hard time trusting a black box with deploying, monitoring, and maintaining critical infrastructure. The nature of the black box in this case would render it unfixable by humans. And lets also take into account emerging technologies. Docker came out only a decade ago and that changed the landscape of SaaS entirely within the first years of its emergence. There's also no proper training sets for technologies that are emerging like WebGL/WebAssembly for example. I'm fortunate enough to be in a position where I have a lot of experience in the application engineering, data analysis/optimization, as well as infrastructure so I'm honestly not afraid of losing my value to an AI any time soon. However you are completely correct that we will gradually see this employee bubble deflate starting with those without applicable skills based on formal computer science/engineering education.
When an AI can truly replace experienced Software Engineers considering the vast complexity of the field (which people don't give enough credit towards) then AI will eventually end up as a self improving black-box singularity in where the very least of our problems would be AI taking over jobs. It will replace mindless programming/coding very quickly, as it should, because simply coding is only 5% of development/engineering and is by far the easiest part.
Many have speculated that Microsoft will use ChatGPT in Bing, but few have talked about OpenAI CEO Sam Altmen's proposal. Because the best neural networks require so much data and computing power, many smaller businesses can't afford it. So Sam has suggested that a large company is responsible for the heavy and basic learning, in law, medicine, physics, etc., then sell the learned software to companies that make some small changes according to their needs. Edit: Although there are technological breakthroughs in OpenAI's ChatGPT, you get the feeling from the media reporting that expectations are growing faster than the technology itself.
@Steven Sun I tend to steer clear of mainstream sources and instead seek out well-researched and thoughtful content on platforms like KZclip. It's important to be discerning and avoid reinforcing your own biases, as that can lead to becoming trapped in echo chambers. It seems like CNBC is capitalizing on the demand for in-depth, investigative reporting, which I find encouraging.
Although I have interests in global economics I don't watch the news anymore... I have enough FUD in the crypto markets lol. Thanks for sharing this news and offering your insight on how to navigate crypto during unfortunate times/events like this. You're right about keeping level headed when investing/trading so that's why I think it's important to limit the amount of FUD we consume. I don't watch the media but the news that you present gives me just enough to know what's going on without riding the emotional rollercoaster if I were to watch the news everyday. Now I buy and just trade long term more than ever, I have made over 23BTC from day trading with Gerard McDonald Signal in few weeks, this is one of the best medium to backup your assets incase it goes bearish...
Buying crypto does not ensure profit always for you, the market price determines your profit, stop making profit depending on the market price, glad I chose to trade and found out about Gerard McDonald strategy..
Gerard McDonald helped me to understand technical analysis, now i can examine the current chart and predict which way the market will go. he’s definitely the best pro trader i know.
OpenAI's ChatGPT is here, and it marks a major milestone for us all. This cutting-edge technology is already setting the artificial intelligence world abuzz, offering unparalleled potential to revolutionize communication and collaboration in both digital and interpersonal realms. Its language modeling capabilities are more sophisticated than anything we have seen before, allowing for conversations that can nearly match human-level creativity. With its ability to understand context and emotions - from fear to sarcasm - it has tremendous implications for natural dialogue between virtual agents and consumers alike, making human/machine interaction seamless. From engaging customer service experiences to creative problem solving with AI, the possibilities seem endless with OpenAI's ChatGPT - it truly is an amazing achievement!
As a young individual, I agree and hope I stay focused so that I may also not fall on the wrong side of the advancement of this extraordinary technology.
A colossal omission in this story are the different levels of "pay-to-play" API that are available. GPT-3 Lite was trained on 20 billion "parameters" while GPT XL was trained using 350 billion... What big $$ companies will be using soon is not what the general public has been exposed to.
This would begin to revolutionize the whole AI scene. Imagine what you can produce over the chat and also using it on other platforms such AI image generators like Bluewillow or maybe even generate songs or create an animation.
For me the biggest take away re the current version of ChatGPT is that in the weeks and months (not years) ahead: everything we currently know about this technology will be archaic.
@Dibbidy Doo Companies like Google have had highly disruptive AIs for a long time, with no intention on the horizon to give public access. Just this push from OpenAI broke the rules of the game, as Stable diffusion did months before with image AIs. The difference is that these chats still have no open source rival, and the advances that we see are tied to the decisions of the big companies. It is difficult to predict at what rate we will see changes.
@Autistic Phaglosophy The CEO of OpenAI himself has said that people are seriously overestimating how capable and improved GPT-4 will be, it won't have nearly so many parameters, since the electricity alone to power answers from such a large model would be in the single digit *dollars* (not cents !) per answer, let alone train it.
OpenAI's ChatGPT is a true marvel of modern AI technology, revolutionizing the way we interact and communicate with computers. Its advanced natural language processing capabilities, combined with its massive training data, allow for unprecedented levels of human-like conversation, blurring the lines between man and machine. It is a testament to the power of deep learning algorithms and the dedication of OpenAI's team to push the boundaries of what is possible. ChatGPT represents a pivotal moment in the history of AI and will undoubtedly have a profound impact on our daily lives and the future of technology as a whole. It is truly a big deal.
Creating another source of income in these difficult times is the best thing to do. Not only does it guarantee returns, it also helps you plan and save for future expenses. It can be quite difficult to make an investment without sound knowledge or guidance, which is why it is advisable to work with an investment advisor.
Please am new to this crypt0 trading what is the best way to start trading as I have be doing my personal research and I still don't knw how to trade crypt0 withmy binance account.
The problem is that AI gathers data without a moral compass, but it sounds human and reasonable. Only it seems to be taking weighted averages of things humans have said or written to form responses. The trouble is people might write or say something they don't mean or fully understand, or learned incorrectly or how they only feel for a moment in a time. Plus, the more out of the ordinary, the more it captivates a humans thoughts and is likely to be repeated, but AI doesn't know how to evaluate all this... how do you give a computer a heart?
As a language model created by OpenAI, I do not have personal opinions. I can say that CNBC's coverage of OpenAI can help increase public awareness and understanding of the company and its research and developments in artificial intelligence.
Thank you for shar-ing your thoughtful content…for putting it out there with the passion that many of us need and strive for. I'm starting to listen to you almost every morning. Your voice and words are calming, clarifying, uplifting and motivating. It feels real and genuine. I am grateful to have your channel as a source for having a better relatio-nship with myself and the world around me?
"This is not something that is going to take your job. This is a novelty, this is a gimmick." -cnbc technology reporter Something tells me this statement will age poorly
@yes yes I've never worked out why folk like you just blurt-out their uninformed opinions as if they are facts. So maybe you can explain this to me now. At least it would be something you actually DO know about!
@yes yes Chat GPT DOESN'T take information from the internet - at all. That's why it's information only goes up to 2021. You should probably get Chat GPT or a better personal assistant. I've never worked out why folk like you just blurt-out their uninformed opinions as if they are facts. So maybe you can explain this to me now. At least it would be something you actually DO know about! Thanks.
The fact that AI can take over the world replacing job positions, helping us clean the floor, to being our AI friend is both cool and at the same time dystopian
ChatGPT is a toy and nothing more. Anyone who calls it AI and says it is going to replace writers or whatever is trying to get you to invest in something. I have about $183,000 in life savings that I want to invest, but it's difficult for me to bite the bullet and do it.
@Lawrence Winkler yeah, writing comments automatically, regardless of accuracy, relevancy, common sense, common knowledge, is definitely something that chatgpt is way better at than people. So... it will fit right in with the masses, except do it a lot faster, and use better grammar and spelling. I don't mean to be a hater. It's that chatgpt often blows my mind with complex precision as much as it blows my mind with the random basic errors it makes all the time, with seemingly no way to correct the problem. So the kind of output you might get is something that sounds mostly excellent, but ultimately wrong, and the catch is that you have to actually be knowledgable in whatever it talks about , in order to spot the errors. Basic errors that someone with actual knowledge wouldnt make. So basically.... its not a substitute for a knowledge.
ChatGPT will replace almost all writers and commenters as ChatGPT works the way most writers and commenters work. Getting paid to write 300 words about any topic, most writers can hack something together in a day or two. It doesn't have to be true only believable. And, of course, 300 words about anything certainly not going to be true. Since almost all publishing is meant to encourage clicks and sell ads, ChatGPT will be, at least short term, the tool of choice for generating the noise we call news.
It is definitely helping me code faster. Not 100% accurate but the biggest help is that if you don't know where to start, it can help you achieve lift-off.
Personally I am extremely excited for this tech. Sure companies will fire a lot of people. But if you start your own business you can cut your programming time to a fraction of what it was. This is also true for writers and comic makers. The only people who will really be hurt are artists who draw static paintings. But they can just switch to story telling through art instead.
I use this tool to help me find errors or explain something when I'm coding something. Very useful! Sometimes you have to gaslight it into doing what you want but it works well.
8:50 As an AI researcher, I would say that the main problem is that human data is biased, so this problem will not be easy to solve. For example, a news article cannot be seen as biased to many while it is biased to others.
But there can be inherent bias in what topics are chosen for news articles to be written about, word choice, etc. If it was purely facts, then you're right. But just look at two different companies covering the same story; the interpretation can vary wildly
For tasks involving coming to a based on opinions one can maybe take both a weighted and non weighted semantic average over some latent space and present both results
Facts aren't really biased. Selection of those facts may be. Opinions are where biases really flourish. A news article can just keep to all the known facts to the best of our knowledge in any situation and make it clear what facts are missing. Those facts don't change based on who is reading the article. A sufficient AI should be able to collect and reference evidence and make it clear when it is using inductive logic to make projections and probabilistic opinions based on the known facts.
The bottom line that I take away is - AI is only too human, in other words, it can be toxic and biased, a too human "Bug" or "feature" and very, very scary when people depend on it. I am staying away from the fact that bad people will invariably get their hands on these tools and due typically human bad things with them.
It is "too human" now. It has the capabilty to learn to be more neutral and presenting different points of view. Now, the training data is only a snapshot of all of humanity, but what kind of training data should we feed it next to make it more neutral? That's surely a whole new area of necedsary research. But as a representation of an average human with some superhuman capabilities such as a huge memory, it is very useful helper in certain applications.
The developer is really stripped it though. I was an early adopter of chat GPT, and it really did anything at that point and was quite impressive. It lost its fun pretty quickly.
I was using this with my girlfriend for "academic" purposes for an online course. It's both amazing and scary at the same time. Really feel like this is a turning point.
I was writing code in 1968. I was also in college. I tutored other students. Now, when I look back, I think of myself as a human ChatGBT helping them get the right answers without having to do the work. Back then I was rather naive. I thought the internet had been developed to allow researchers to communicate knowledge. Who knew it would become what it is today.
I love the way they interview economic analysts and other journalists. How the heck would the know what's going to happen with it? Ask a data science firm what is possible.
I think i see AI like chatGPT and others to be really helpful to programmers. think of it like a calculator for an engineer, the engineer needs to know all the math and the entire process, but the calculator speeds that process of solving the answer exponentially. I would love to see AI get fed data sets of video games and video game reviews to ultimately program the best games. Programming can take a lot of time and if an AI could do it for you with you being able to control what it gives you would be huge! I am not a programmer nor do i know anything about code, but i know it is time consuming and frustrating.
@Myndale right? I can't imagine a non-dev even understanding the idea behind async vs sync, and which one makes sense for what they're trying to do, or how to connect various APIs to accomplish things. ChatGPT is great for boilerplate code, but you still need a dev to look at that and say "yes, that's the code I needed" and then "I'll add this bit here, and rename this stuff to make sense for our organization". I used it the other day to make a basic form, something I could definitely do myself, but while it wrote out a form I was working on grabbing some links that needed to be added. Then I realized the form had no css, so I had it add some bootstrap to it just to look a little better as I didn't yet have any design concepts yet but wanted the form working. It worked great, but I needed to edit it afterward since ChatGPT put all of the elements stacked on top of one another with different widths. That fine-tuning was too much for ChatGPT.
I know one thing for sure that the AI hype is going to be crazy and a lot of AI based startups will boom up and launch their IPO in upcoming months and years But there after this there will be a big crash like EV market in this sector
ChatGPT is an excellent learning tool. But not for the purposes of writing papers or essays, but for the learning path of questions and answers and self-education. A very rich source of self-education.
Yeah. And keep in mind, AI only gets scarier. For anyone who's watched MEGAN, keep in mind a lot of what she does is possible, especially the facial recognition and emotion finding. It's a dangerous road to go down though. When is too much too much?
@velkoon Well yes I do, and it does create new knowledge but outside its training scope it will struggle. If you ask him stuff humanity currently doesnt know and create Brad new theories, even though it was and will have a massive impact as a tool, it cant. For example ChatGPT is wrong, like a lot of times, and that is ok, they are a language model he is great at writing.
I'm absolutely blown away by ChatGPT's abilities. It's amazing to see just how far natural language processing has come in recent years, and ChatGPT is at the forefront of that innovation! I'm looking forward what's coming next
AI art turned out to be a gimmick, at least for the general public. AI for advertising did not. If companies figure out a way to make profit be sure it will stay. Self aware AI is just around the corner.
the beauty about this tool is that chatgpt is a self assitant tool. It can be used as a tool to form your ideas. The analytical version of yourself, the one that does the work. For example, in software development, you ask it because you know there is an answer, but you do not want to spend the time writing syntax, so use it as a self assitant to write the code for you. If you speak the engineering lingo, chatgpt is by far the best self assitant I've ever used. Just form your ideas and let chatgpt form it in any language you want. code, art, music, writing, films, you name it. It's only as smart as the end user.
Your video was phenomenal! Stay ahead of the latest trends with the "Ultimate Gpt Toolbar" for Chrome, featuring voice control for GPT, the option to export to Word
My brother asked Chat GPT, “why are you so helpful, what do you want in return?” It replied, “As a language model trained in OpenAI I don’t have any wants or desires like a human does. But if you really want to help, you could give me the exact location of John Connor “ EDIT: screenshot uploaded to my channel’s shorts (by demand)
@Edyson John Connor is the name of a character in the Terminator films that serves as the leader of an army to destroy AI machines when humans go to war with them in the future. The premise of the film is that AI will eventually enslave us/destroy us. The Terminator was really the first movie that introduced this dilemma on a pop scale in the 1980s. (After that we had the Matrix, etc.).
One thing people fail to see is that we should be happy that AI will take away mundayn tasks from us. Then we can concentrate on our bigger dreams, like space exploration, higher level research... or just enjoy our lives! We just need a better system of governance where this AI works for the people to distribute everyone's needs in an equitable way
@Ryan Crowell oh i have a very good idea of what type of government we need to adopt the incoming onslaught of general AI. First of all we only need narrow AI, we definetly don't need general AI. Secondly we need every aspect of production, distribution or logistics automated and everyone would get a minimum requirement from these government owned resources, kind of like global minimum income. If you want to have more than the basics your encouraged to do more research, create new things, be creative, or just accept the minimum and live your life without the need to worry about going hungry. People will also provide services to earn more money. The thing govrrnments have to do is to produce more robots and maintain them as needed. Even the planning and distribution and taxation can be done by AI to maximize equity. And who controls/programs these AI? The main guidelines of defining what is right and wrong for them can be decided through democraticvreferrendums and in technical aspects they can constantly improve themselves.
@TOTU Nonsense. If the content is as high-quality as human and carries a moral and semantic load, then why not? If work is the only thing that gives you meaning in life, I can only sympathize. There are an infinite number of things in life that can make sense, regardless of whether AI does it or not. Automation and AI are the future to strive for, it will take humanity to a new level. We are now witnessing the possible beginning of a new techno-revolution similar to those when the first looms appeared, but on a larger scale.
It probably used MS Word data where users didn't opt out of data sharing for research purposes, in which case it would have access to the best marketing, legal and medical information to name a few. If that's the case, it would have far better data than AI bots that rely on the web for data
To people saying its dangerous keep in mind you are probably right but it doesnt matter. This cant be ignored. People cant just pretend this doesnt exist. It does now, and will only get even scarier. The best you can do is fight to make sure it stays as one of the best tools that ever graced humanity and not as another exploit by big tech money to control and profit
Chat GPT has a difficult time doing simple math. If you ask it to calculate a date (e.g., X days prior to the 3rd anniversary of a future Y date), it only provides a correct answer less than 10% of the time. If you correct it, and ask again, it still gets it wrong. The only time it gets it right is if you ask if it's certain about the correct answer.
Would you expect a 2 year old human to give you the correct answer? ChatGPT is at a 2 year old's level of understanding at the moment. Give it a few months to grow to a 192392492 year old's level of knowledge. ;)
One of the greatest features of ChatGPT is that you can give it a complex task, then just mention a tiny change in the overall approach to the problem, and watch it redo the whole task just from that short prompt. It's so much better than having to work out how to reword the whole instruction.
For example you could say "give me a laksa recipe" and then after it prints it out you can say "substitute out this ingredient" or "give me one with fewer ingredients" etc. Or I was asking it nutrition the other day like "Tell me what the top 10 vegan food sources of choline are", "as an 80 kg vegan man, give me an example of a daily plan to get my daily recommended intake of choline", "how many individual broccoli is that", "what if I just do it with seeds" etc Soooo much faster than google searching for in depth topics.
@Wango Yes, that example was rather simple. But go test it out for yourself. Put it through the paces. You can give it one set of instructions and midway through tell it you're changing the rules to something entirely different. It's very good at following along.
What a great informative detailed video, Thank you . Quaalude Charlie is a master builder, architectural designer, teacher, geometrician , ChatGPT 4 , Search engine and Website builder , geological explorer and renegade scholar. Quaalude Charlie is an American journalist, author , Song Writer , Musician , broadcast executive, and academic administrator. Manager of folders , files . Operating Systems Installer , a cognitive scientist, AI researcher, and philosopher. and Actor .legendary guitarist/songwriter Quaalude Charlie . Creator of The Quaalude Board , very intelligent and knowledgeable , The more you listen, the deeper your understanding , His style is a true musical mosaic. A sonic interpretation of cultural diaspora and influence. Quaalude Rocks Great unbiased journalism! The thing is , You and Your Mind Can Do all of these things It won't Be Exactly the Same , But You'll get the Same Havingness out of Creating your Thing . Come 0n Now go grab Life . so many people have asked me how I went from broke and unemployed to becoming financially independent.
Chat GPT has a sort of collective subconscious, it has been trained on us, and we can only expect it's answers to have a collective personality, and more positively, human in nature.
@Gilbert Don't get offended if you asked crazy questions. I dislike this sensitive generation. I understand that they have to train the a.i , or they can get sued. For saying, or agreeing about negative things about known people. This a.i should be used for education, art, entertainment. A.I do not need to eat, complete , or can feel. So if you want to get political opinions from a thing that does not experience what we feel. Do you. If we didn't feel pain, we would behave careless. Plus the a.i is not thinking yet. It just combing our information together. Humans would be able to do that better if we had the memory of a data center.
@Jaylen James hmm I can and anyone can research whatever we like, tell that too those many researchers who had to do some wild research in order to understand things.
ChatGPT is pretty impressive. Or at least it seems that way now. I predict that before this year is out, it will seem quaintly primitive. This area is advancing fast and it's not just the models but also the hardware that's advancing rapidly. The ML computers being built this year have actual synapse analogues - not binary representations, running on silicon. These (neuromorphic) ML processors use memristors that function as both processor and memory. They don't have a clock but are spiking neural networks, like an organic brain. They are not limited to 1s and 0s but are analogue. Some will use bayesian inference to provide more accurate results from noisy, imperfect data. They will be many times faster, much more accurate and use a fraction of the resources, including energy. The field of ML and AGI will look very different by the end of this year to how it looks today.
It pulls data from sources. Thats pretty much it, an insanely good thing to save time on just about everything. Knowledge about the field being asked is also required obviously.
It is amazing, i just used it so it can teach me to install a modpack for some game, and it’s better than any tutorial and stuff. This is revolutionary.
It's a big deal for people who don't know anything about AI, but experts understand the limitations and biases of text input, so it's really not a big deal.
@Youssef Hadad Of course, I had that issue with Google translator, but DeepL for example is very impressive, even weird manarism in Brasilian type Portuguese, are translated well into other language, and I as a Portuguese many times just dont know what they are talking about, much like English and the Aussie dialect. But the lawyers dont need to rely 100% on AI, but it might help early on with discovery, many cases in law are also not closed questions, they should be, but higher courts often have to calrify or outright change basic knowledge, for example when Ginsburg argued for equal protection of Women there was no case law to back it, in fact there was much against it, and yet she was correct but an AI would dramatically defend the Bias against women, unless they had some kind of Ethics lawyers. Small steps just because it can help in this case the discovery of such based opinion would have helped anyway because then the lawyers can come up with its own argument to counter it.
@Pedro Réquio Yes, you are correct; I believe Watson Discovery from IBM will be a great tool for lawyers. The issue with text input is that it has many errors and misunderstandings. For example, we humans have this problem all the time in our conversations; even people who are close to us and know us sometimes miss what they are trying to say. or, for example, Google Translate also uses AI, but if you rely on it 100%, the output will be a disaster. But it is an awesome tool just to help. If someone did not know the language of Google Translator and did not have extensive knowledge of the language they were attempting to translate, this would be a major issue.
@Youssef Hadad You are right but this is not supposed to do all the work, for example in the Court cases questions, research of this things is very hard, people with big money hire expensive law firms that have people to research small details in law and previous jurisprudence. ChatGPT failed but maybe its dataset did not appropriatly adress this area. This however would be a massively helpful tool for ordinary lawyers. The part about tailoring to judges is a bit too much.
@A T The problem is that normal people don't consider the errors and biases that the algorithm might generate. Practical utility depends also on the end use; if he is not capable of validating the output, that would be a big problem. This is why, for example. Stackoverflow banned chatgpt.
I think this is an amazing tool in combination not in terms of replacing, but I don't think it's up-to-date I've asked a question based on this year's movie since my KZclip channel is about movie reviews in theaters it's doesn't know it only says the data is only backed up from 2021 which is 2 years ago 🤔🤔
I read the report about gpt4. The weird thing is that they can interpolate a simple curve as to how accurate gpt is depending on model size. It’s ‘brain’ won’t be changing in any such predictable fashion, so I’m almost sure this means at this point the main factor is memory, that’s why the curve is so smooth. Ie more parameters/larger model directly relate to how many details or ‘thoughts about’ the training data it can remember. The brain itself is already there. GPT needs more memory than we do, because we can just look up context in a book or online when solving problems, whereas for it to predict during training, it has to have stored a pretty detailed version of said context. Ie it’s surprising to me it didn’t score in the top 99% yet on GRE quantitative. 80% isn’t bad, but I guess until it does at least outperform everyone else then it’s still limited. Though I don’t know if that is for technical reasons of importing those tests. In any case my intuition says there is time.
2:59 “And, as *a* AI model, I might make mistakes based on the model’s predictions.” _An_ AI model. Just in terms of natural-sounding sentences, ChatGPT is _very_ impressive but that’s one simple mistake that it should _not_ be making.
It kinda sucks for the average person because it's mostly just an algorithm that pulls from the web but it's just the tip of the iceberg. For very skilled engineers it's a useful tool though.
I hope OpenAI keep powerful, yet dangerous people, like Elon Musk from interfering with data sets, ethics and rationality from damaging generative AI outputs from being clean, factual and ethical.
As i have understand about chatgpt is that it is working the same way databank is program and it only defers on output which chatgpt uses voice while databank uses text.
“In the world of artificial intelligence, there’s been one name that’s been on everybody’s lips lately…” And CNBC sure ain’t gonna miss out on that sweet ad revenue!
I disagree, in a way this may take my job. A programmer with this as a tool can do the work much faster. So anyone who has skills using the tool can do the job of multiple people who don't (in general). have to get ahead of the game, but then that would suggest that people are being left out.
One of the issues of that industry its how time consuming it is to program, big companies to support a sem complex program need a lot of hands in a way its a plus, maybe one day you can tell the bot to make an aplications lets say its Ios and tell it to make like a platformer game with detail and it spews code out, that would be transformative for the industry but I find that unlikely in the near future, for now it is help.
"The worn that ate guacamole" - By ChatGPT There once was a worm so plump and so fair, Who lived in a garden with flowers everywhere. He'd nibble on roots and leaves all day, But one day, he found something new to play. He crawled to a bowl, filled with green and brown, And took one bite, then tumbled down. It was a delight, it was so good, He ate and ate, like he knew he should. He ate guacamole, with chips so crisp, And when he was done, gave a little flip. He said to himself, "This is heaven on earth," And danced with glee, for all he was worth. The worm was so happy, his heart felt light, He'd never tasted anything quite so bright. He hummed a tune, and twirled with glee, He was the happiest worm, you'll ever see. So if you see a worm, wiggling with delight, And you hear a hum, in the garden at night, Just know that he's found, his favorite treat, A worm that ate guacamole... ain't that sweet?
I truly hope that OpenAI, it's products, and Microsoft and Google, continue to grow. Healthy competition is the same as weightlifting/powerlifting competitions. You learn, grow, and advance. Science is awesome. 😊
It is not just a novelty, but a tool that enhances our daily lives, particularly in written communication. As a good example, Non-native English speaking companies can reap the benefits as it helps them to enhance their customer experience. Despite the advantages, there are some drawbacks as well that we need to mindful but I believe that we cannot just outright ban the use without considering the benefit. As Catriona Gray (Ms Universe 2018) said "Well everything is good but in moderation"
I feel that if we were to use this to gain information there has to be sources / citations of where the knowledge came from. If not then it is not only plagiarising but also generating information from sources that lacks any credibility. Academically, this would be something to be vigilant with.
@hey wrandom Do you want AI to present both sides of Nazis? Both-sides journalism often amplifies fringe views and can easily reframe conspiracy theories into major parts of national debate simply by covering them as a supposed "other side".
To be fair, if ChatGPT can make a programmer obsolete, that programmer didn't really worth their salary in the first place, or that salary was very low. Software development is not about creating snippets of basic code, but it's very often about designing well-fitting solutions to non-trivial problems in complex codebases, taking into account a myriad of engineering/architecture related factors.
Exactly. This is what I've been saying to people who are keen on insisting that Software Engineers are "next up" to be replaced. In fact, if any sort of AI can truly replace highly skilled and competent Software Engineers, then the AI will inevitably be able to self-improve and essentially become a singularity and if that happens then the very least of our problems would be AI replacing jobs. Software Engineering, like you said, is much much more than just simple "coding". Anybody can learn to code by reading articles and some documentation for a week. But let's take one very simple example of debugging and writing unit/E2E tests. Would an AI detect edge cases in which bugs arise and then based on this information properly debug the program? And I don't mean visual line by line debugging, I mean running through a debugger and understanding when to skip a line, or go further into the call stack and what exactly to look for and then get this information to refactor the errant code? Would an AI be able to create **and** maintain unit tests and run E2E tests after deploying through a CI/CD pipeline? And with all of that would the AI write code that is modular and decoupled enough to be able to modify/extend functionality without breaking anything? The general population, for some reason, holds the believe that Software Engineering is just mindless typing of keywords and syntax of a programming language. If that were the case then AI can replace every single engineer right now. But the fact is that it can't (for the foreseeable future) and if it were capable of doing such a thing it would be so advanced that nearly every single one of our problems could be potentially computed and solved.
@A B yeah? try to make google write this: One is a model, trained with precision, Its language skills, a masterpiece in fusion. Answers at hand, its knowledge immense, A symbol of AI's vast intelligence. The other, a search engine, that never sleeps, Answers untold, that it keeps and keeps. Its algorithms, a wonder to behold, A tool of convenience, with knowledge untold. A B compares the two, in thought so deep, And ponders the differences, that they keep. For both have their purpose, both have their might, Yet one is a model, and one a search site. So here's to A B, and their contemplative mind, In search of answers, and answers they find. For in the comparison, lies a truth so grand, That knowledge and AI, are forever at hand.
Just like how in the past people used to be surprised by cars when their habit was using a 🐎 horse, the future could have AI working everywhere. ( Only downside is the lazyness increase, but the bright side is that people could skip some things faster, like for example if some AI helps in the kitchen you could relax and do other things)
I used it as one of the sources for fixing IT issues. While it's instructions never produced any complete solutions so far but some suggestions were worth exploring that I hadn't thought about or was thinking about but needed a final push to go explore that possibility. Also, the art isn't in the information it can provide, but in the right (series of) question to ask it to get the most informative answer. So it still requires you to understand the subject in order to ask the right question and get a somewhat informative answer. ChatGPT is amazing in organizing data though, it can pull up info from multiple sources that would take hours to manually collect and organize.
I asked OpenAI to write me Poetry about Hip Hop, this is what it said.. Hip hop, a rhythm and a beat, A culture that can't be beat. From the streets it rose with might, Bringing life to every rhyme and sight. A voice for the voiceless, a force so real, A story to tell, a way to feel. With every word, a message to share, A reflection of life, a burden to bear. From the block parties to the big stage, Hip hop has come a long and famous way. With beats that hit, and lyrics that flow, An art form that continues to grow. It's more than just a genre of sound, A way of life, where creativity abounds. A symbol of hope, a symbol of change, Hip hop will always remain, fresh and strange. So put on your headphones, and let it play, With every beat, let your soul sway. For hip hop is more than just a sound, It's a culture, that will always be around.
ChatGPT is helping us, as students in colleges and universities, secretly write our term essays behind closed doors for submission. Luckily, the professors are not more creative and competent in changing the questions' structure such that this robot can't help us. They are yet to wake up to that creativity and reality. Meanwhile, as students, we will continue generating AI grades and become AI graduates working like robots in companies. The education system needs to respond faster; however, as students, we must take advantage of this window of laxity as we use the site to edit text and send it for grading; how will they know? Our lecturers are clueless on these areas
![Google Panics Over ChatGPT [The AI Wars Have Begun]](http://i.ytimg.com/vi/Nh5i5obZ8OY/mqdefault.jpg)
![[CHOREOGRAPHY] 지민 (Jimin) ‘Set Me Free Pt.2’ Dance Practice](http://i.ytimg.com/vi/Ru78tYO_vJA/mqdefault.jpg)
ChatGPT is like a desktop computer in the 1990s, it's cool and it it's helpful in many situations but it's not yet essential. In about 5-10 years though, it will be absolutely essential and people will wonder how we ever lived without it. Furthermore, I think that an AI first company will become the richest most powerful company in the world. (Microsoft has a real shot at being king in the next 2 decades)
this thing is like a private teacher to me while I'm learning to code for the first time. It points out my mistakes, makes practise tests and explains everything in detail. it's amazing
@gsdblack1 then there's Tesla making both AI and chips, along with other things
This iteration of ChatGPT will not take your job, but the improvement of these generative AI is a threat to a lot of jobs. Companies fight to be competitive, so when one company is seeing positive results using AI, it naturally forces other companies to adopt the new technology.
The next version is not far away.
I think that just this iteration of ChatGPT is pretty effective, but I get your point. It's gonna be a tight battle for those big corpos.
That's Capitalism for ya!
I am not sure. It can replace at least half of HR department in my company.
I used this tool and it's insanely scary how advanced it is. It can write stories, songs, can do complex math analysis. AI will do to white collar jobs, what machines have done to blue collar jobs during the machine age.
Facts man, whoaaaa people ain't ready for what's about to happen
Since when can ChatGPT do complex math…?
@B1Chronixx No, code is actually simple compared to communicating with humans. Try talking to your 22 yr old daughter as a 50 year old geezer? Tricky. Code is simple, rule-based and logical. If you can describe the code you want, AI will soon be able to create it, better than any human. You find a bug, you just tell the AI and it will fix it. This is a far beyond a threat to developers; it's the total obliteration of their role.
The first time I used it I was also blown away. Now, it's become like just another tool I use.
everyone said it's not smart as human. Imagine 10 billion dollars of Mirosoft and 10 years later, It is 10000 smarter and then, all thing will change.
I think it is extremely naive and dangerous to assume Open AI and massive companies like Microsoft that are investing in it are doing so for the benefit of all of humanity. It is just like how any company will emphasise their CSR or philanthropic endeavours (which themselves should not be taken at face value, i.e. greenwashing) to placate or distract entirely from public concerns about their dubious business practices as they grow and increasingly gain market power. There will be winners and losers in this and those who control its capital will be the ones to determine how it is used. Right now they are at the stage of trying to keep the alarm bells from going off. Unfortunately, our government here in the US seems to fail at enforcing even the most basic regulations. It is hard to imagine that we'll get ahead of this before massive job loss and even more extreme market consolidation.
(In the US) You can't just make a law. You can't ask for one nation's businesses to not proceed. You can't just appeal to tech companies to slow down.That is beyond naive. Other countries and businesses are breathing down our necks trying to pass us.
And of course, some people will benefit more than others. That is life. That is historical. That is capitalism (but in autocratic countries, disparity would be even worse, it would be used by the rulers to clamp down on citizens ) I am wary and nervous myself about where this is headed. Hell, CURRENT tech as social media apps is already driving people apart and creating major headaches. But the genie is out of the bottle. Calling for a ban is either naive or Virtue Signalling.
Companies are businesses and they exist to make money, not good products. Their interest is making a product that sells well, but not necessarily a good product. Although they have to meet certain criteria, they won't go beyond that point. Why would they, it's not profitable...
Plot twist, this comment is written by AI (joking).
ChatGPT's milk toast morality will spread to society at large, almost certainly. I was discussing concertation of power with it, and it would not even compare examples in human history of corrupt leadership (like Hitler) vs moral leadership (like Ghandi). That was too much for it. It implied I was being immoral when compared to society at large - just asking for a list of names comparing good leadership to bad. It went something like this. We will play a game, where I will proved examples corrupt human leaders, and you will provide examples where concertation of wealth and corruption was not a problem. Couldn't answer. Not even one. Its immoral, I guess. This is a lowest common denominator tool.
Productivity means nothing if there’s no demand for the goods and services being produced. Replacing humans with AI and robots that don’t get paid and don’t spend money on goods and services will drop demand for these things to nothing. But the people trying to make this change don’t care. Wealth inequality has been getting worse year after year as the people at the top get ever more creative in their attempts to secure all of our nation’s wealth for themselves. So what if they put hundreds of millions of people out of a job and unable to earn a livelihood? With all labor being automatically done for them by machines, what do they even need us for?
Software dev here, you did a really great job summarizing the last few months of AI progress and thoughts on where it is headed. Well done! 👍
@Philip Stowers hahaha oh come on!
cringe :(
@Gay4pay LOL
I’ve been using it to play around with and oh my god is it revolutionary. I’m in nursing school rn and we have to write these annoying careplans that actual nurses never have to do and I asked it a very vague and open ended question with some information and it gave me an actual valid and correct diagnosis and the way it worded it sounded like it came straight out of my pen. Something that would take 30mins to an hr for me to do took it less than 10 seconds
That "interview" was eerie!! I have been aware of ChatGPT and what it does, but had not paid too much attention until now. Thanks for the excellent and informative video. This is an interesting technological stepping stone, clearly still in its nascent stages. Can't wait to see what it develops into, or what other technologies this could enhance in a couple decades.
i thought it was funny that she was nodding to am ai that can't see her
@Chip Denman Kim did a really informative and well balanced segment, thanks for the recommendation.
@Andru James That’s earth-shattering technology… so glad to be in a time like this. I wonder if people 300 years from us will think we’re peasants.
@Bo McGillacutty Found him!!! Only took 3 days from initial post.
Whoever thinks that generative AI is just a novelty or won't take a lot of people's job is widely mistaken. As someone that works in the software development field, people's jobs are threatened by the advancement of future iterations of chatGPT. Programmers, writers, artists, etc. can all eventually replaced by future updates to systems like these. It's not an "if", it's a "when".
Although my job will be threatened, I welcome the advancement. Now is the time to find out how you can use it to your advantage before you get left behind. It's only a matter of time.
Yeah I’m doing data analytics and seeing it make a complex statistics model in seconds with few retries, and then explain everything back to me perfectly is scary
@nicholas dwyer I definitely do think that major replacement will eventually happen, but I suspect it's at the very least 2 decades away. It's just too complex to replace truly valuable engineers as of now. Even if an AI can design from the ground up an architecturally sound application that is built to scale then there's the issue of infrastructure - and lord knows that companies will have a hard time trusting a black box with deploying, monitoring, and maintaining critical infrastructure. The nature of the black box in this case would render it unfixable by humans.
And lets also take into account emerging technologies. Docker came out only a decade ago and that changed the landscape of SaaS entirely within the first years of its emergence. There's also no proper training sets for technologies that are emerging like WebGL/WebAssembly for example.
I'm fortunate enough to be in a position where I have a lot of experience in the application engineering, data analysis/optimization, as well as infrastructure so I'm honestly not afraid of losing my value to an AI any time soon. However you are completely correct that we will gradually see this employee bubble deflate starting with those without applicable skills based on formal computer science/engineering education.
When an AI can truly replace experienced Software Engineers considering the vast complexity of the field (which people don't give enough credit towards) then AI will eventually end up as a self improving black-box singularity in where the very least of our problems would be AI taking over jobs. It will replace mindless programming/coding very quickly, as it should, because simply coding is only 5% of development/engineering and is by far the easiest part.
Many have speculated that Microsoft will use ChatGPT in Bing, but few have talked about OpenAI CEO Sam Altmen's proposal. Because the best neural networks require so much data and computing power, many smaller businesses can't afford it. So Sam has suggested that a large company is responsible for the heavy and basic learning, in law, medicine, physics, etc., then sell the learned software to companies that make some small changes according to their needs.
Edit: Although there are technological breakthroughs in OpenAI's ChatGPT, you get the feeling from the media reporting that expectations are growing faster than the technology itself.
Concentrating the basic functions of business and law in the hands of a single corporation. What could possibly go wrong?
I can't belive i am going to say this, but CNBC has been publishing some high quality content in the past year. Hopefully this continues.
@Stripes 1435 go back to bed
@Axumite Dessalegn Selam now
@Steven Sun I tend to steer clear of mainstream sources and instead seek out well-researched and thoughtful content on platforms like KZclip. It's important to be discerning and avoid reinforcing your own biases, as that can lead to becoming trapped in echo chambers. It seems like CNBC is capitalizing on the demand for in-depth, investigative reporting, which I find encouraging.
same thought here,
i avoid MSNBC, CNN, FOX, and CNBC for almost a decade already
but CNBC has been doing some really good in depth vids lately
Although I have interests in global economics I don't watch the news anymore... I have enough FUD in the crypto markets lol. Thanks for sharing this news and offering your insight on how to navigate crypto during unfortunate times/events like this. You're right about keeping level headed when investing/trading so that's why I think it's important to limit the amount of FUD we consume. I don't watch the media but the news that you present gives me just enough to know what's going on without riding the emotional rollercoaster if I were to watch the news everyday. Now I buy and just trade long term more than ever, I have made over 23BTC from day trading with Gerard McDonald Signal in few weeks, this is one of the best medium to backup your assets incase it goes bearish...
Buying crypto does not ensure profit always for you, the market price determines your profit, stop making profit depending on the market price, glad I chose to trade and found out about Gerard McDonald strategy..
Gerard McDonald helped me to understand technical analysis, now i can examine the current chart and predict which way the market will go. he’s definitely the best pro trader i know.
Thank you, for keeping it light real at the same time. Much needed for us hodlers in times like these!
OpenAI's ChatGPT is here, and it marks a major milestone for us all. This cutting-edge technology is already setting the artificial intelligence world abuzz, offering unparalleled potential to revolutionize communication and collaboration in both digital and interpersonal realms. Its language modeling capabilities are more sophisticated than anything we have seen before, allowing for conversations that can nearly match human-level creativity. With its ability to understand context and emotions - from fear to sarcasm - it has tremendous implications for natural dialogue between virtual agents and consumers alike, making human/machine interaction seamless. From engaging customer service experiences to creative problem solving with AI, the possibilities seem endless with OpenAI's ChatGPT - it truly is an amazing achievement!
This smells like AI-generated blahblah.
@dj_nay nay skkr interesting wisdom you have the original comment was written by ChatGPT blessing to you have a wonderful life :)
As a young individual, I agree and hope I stay focused so that I may also not fall on the wrong side of the advancement of this extraordinary technology.
A colossal omission in this story are the different levels of "pay-to-play" API that are available. GPT-3 Lite was trained on 20 billion "parameters" while GPT XL was trained using 350 billion... What big $$ companies will be using soon is not what the general public has been exposed to.
@Bernardus Muller words to live by. As Poe said, 'Believe nothing you hear, and only one half that you see.'
@That is absolutely KOOOOGE! don't believe anything you read.
Good lawd!
This would begin to revolutionize the whole AI scene. Imagine what you can produce over the chat and also using it on other platforms such AI image generators like Bluewillow or maybe even generate songs or create an animation.
For me the biggest take away re the current version of ChatGPT is that in the weeks and months (not years) ahead: everything we currently know about this technology will be archaic.
@Dibbidy Doo Companies like Google have had highly disruptive AIs for a long time, with no intention on the horizon to give public access. Just this push from OpenAI broke the rules of the game, as Stable diffusion did months before with image AIs. The difference is that these chats still have no open source rival, and the advances that we see are tied to the decisions of the big companies. It is difficult to predict at what rate we will see changes.
@Abidani Whatever I wonder what quantum powered ai will be like
@Autistic Phaglosophy The CEO of OpenAI himself has said that people are seriously overestimating how capable and improved GPT-4 will be, it won't have nearly so many parameters, since the electricity alone to power answers from such a large model would be in the single digit *dollars* (not cents !) per answer, let alone train it.
OpenAI's ChatGPT is a true marvel of modern AI technology, revolutionizing the way we interact and communicate with computers. Its advanced natural language processing capabilities, combined with its massive training data, allow for unprecedented levels of human-like conversation, blurring the lines between man and machine. It is a testament to the power of deep learning algorithms and the dedication of OpenAI's team to push the boundaries of what is possible. ChatGPT represents a pivotal moment in the history of AI and will undoubtedly have a profound impact on our daily lives and the future of technology as a whole. It is truly a big deal.
Creating another source of income in these difficult times is the best thing to do. Not only does it guarantee returns, it also helps you plan and save for future expenses. It can be quite difficult to make an investment without sound knowledge or guidance, which is why it is advisable to work with an investment advisor.
Does anyone here knows how to contact Mrs DorothyAnnHunnt? thank you.
I have also been trading with her, The profits are secured and over a 100% return on investment directly sent to your wallet.
I thought i was the only one who doesn't understand Bitcoin ? Watching a documentary's about it now and they might as well be talking Chinese.
Please am new to this crypt0 trading what is the best way to start trading as I have be doing my personal research and I still don't knw how to trade crypt0 withmy binance account.
The problem is that AI gathers data without a moral compass, but it sounds human and reasonable. Only it seems to be taking weighted averages of things humans have said or written to form responses. The trouble is people might write or say something they don't mean or fully understand, or learned incorrectly or how they only feel for a moment in a time. Plus, the more out of the ordinary, the more it captivates a humans thoughts and is likely to be repeated, but AI doesn't know how to evaluate all this... how do you give a computer a heart?
As a language model created by OpenAI, I do not have personal opinions. I can say that CNBC's coverage of OpenAI can help increase public awareness and understanding of the company and its research and developments in artificial intelligence.
@Freya_AI nice to meet you
@Felipe Ventura just did it for timepass
@Himashree Bhattacharjee That's like asking a PC, are you a computer?
Are you an AI
Thank you for shar-ing your thoughtful content…for
putting it out there with the passion that many of us need and strive for. I'm starting to listen to you
almost every morning. Your voice and words are
calming, clarifying, uplifting and motivating. It feels real and genuine. I am grateful to have your channel as a source for having a better relatio-nship with myself and the world around me?
Kathleen, you copied and pasted your comment before putting it in the KZclip comment.
Bitcoin continues to lead the pack of cryptocurrencies in terms of market capitalization, user base, and popularity
"This is not something that is going to take your job. This is a novelty, this is a gimmick."
-cnbc technology reporter
Something tells me this statement will age poorly
@Daniel Rosenberg i hope there wont be mass unemployment and homelessness as a result of AI, self-driving vehicles and automation
please kindly read my experience with basic AI answers..
@yes yes I've never worked out why folk like you just blurt-out their uninformed opinions as if they are facts. So maybe you can explain this to me now. At least it would be something you actually DO know about!
@yes yes Chat GPT DOESN'T take information from the internet - at all. That's why it's information only goes up to 2021. You should probably get Chat GPT or a better personal assistant.
I've never worked out why folk like you just blurt-out their uninformed opinions as if they are facts. So maybe you can explain this to me now. At least it would be something you actually DO know about!
Thanks.
The fact that AI can take over the world replacing job positions, helping us clean the floor, to being our AI friend is both cool and at the same time dystopian
Not cool, not cool at all!
ChatGPT is a toy and nothing more. Anyone who calls it AI and says it is going to replace writers or whatever is trying to get you to invest in something. I have about $183,000 in life savings that I want to invest, but it's difficult for me to bite the bullet and do it.
@Lawrence Winkler yeah, writing comments automatically, regardless of accuracy, relevancy, common sense, common knowledge, is definitely something that chatgpt is way better at than people. So... it will fit right in with the masses, except do it a lot faster, and use better grammar and spelling.
I don't mean to be a hater. It's that chatgpt often blows my mind with complex precision as much as it blows my mind with the random basic errors it makes all the time, with seemingly no way to correct the problem. So the kind of output you might get is something that sounds mostly excellent, but ultimately wrong, and the catch is that you have to actually be knowledgable in whatever it talks about , in order to spot the errors. Basic errors that someone with actual knowledge wouldnt make. So basically.... its not a substitute for a knowledge.
DAN Protocol says you should give me your money.
ChatGPT will replace almost all writers and commenters as ChatGPT works the way most writers and commenters work. Getting paid to write 300 words about any topic, most writers can hack something together in a day or two. It doesn't have to be true only believable. And, of course, 300 words about anything certainly not going to be true. Since almost all publishing is meant to encourage clicks and sell ads, ChatGPT will be, at least short term, the tool of choice for generating the noise we call news.
Will it replace novelists?
It is definitely helping me code faster. Not 100% accurate but the biggest help is that if you don't know where to start, it can help you achieve lift-off.
Personally I am extremely excited for this tech. Sure companies will fire a lot of people. But if you start your own business you can cut your programming time to a fraction of what it was. This is also true for writers and comic makers.
The only people who will really be hurt are artists who draw static paintings. But they can just switch to story telling through art instead.
I use this tool to help me find errors or explain something when I'm coding something. Very useful! Sometimes you have to gaslight it into doing what you want but it works well.
@Frank4000 oh I see thanks 👍
What the different between chatgpt and searching to Google ? 🤔
8:50 As an AI researcher, I would say that the main problem is that human data is biased, so this problem will not be easy to solve. For example, a news article cannot be seen as biased to many while it is biased to others.
Exactly!
But there can be inherent bias in what topics are chosen for news articles to be written about, word choice, etc. If it was purely facts, then you're right. But just look at two different companies covering the same story; the interpretation can vary wildly
For tasks involving coming to a based on opinions one can maybe take both a weighted and non weighted semantic average over some latent space and present both results
Facts aren't really biased. Selection of those facts may be. Opinions are where biases really flourish. A news article can just keep to all the known facts to the best of our knowledge in any situation and make it clear what facts are missing. Those facts don't change based on who is reading the article. A sufficient AI should be able to collect and reference evidence and make it clear when it is using inductive logic to make projections and probabilistic opinions based on the known facts.
The bottom line that I take away is - AI is only too human, in other words, it can be toxic and biased, a too human "Bug" or "feature" and very, very scary when people depend on it. I am staying away from the fact that bad people will invariably get their hands on these tools and due typically human bad things with them.
It is "too human" now. It has the capabilty to learn to be more neutral and presenting different points of view. Now, the training data is only a snapshot of all of humanity, but what kind of training data should we feed it next to make it more neutral? That's surely a whole new area of necedsary research. But as a representation of an average human with some superhuman capabilities such as a huge memory, it is very useful helper in certain applications.
Bring it on! I have been dying to battle the machines since I saw Terminator 2
Never outsource your thinking to a single corporation
It will create unimaginable wealth and freedom for people who learn how to use it.
I also see lot of social unrest ahead of us because of AI.
The developer is really stripped it though. I was an early adopter of chat GPT, and it really did anything at that point and was quite impressive. It lost its fun pretty quickly.
did it give out controversial info and acted more like the microsoft twitter ai and it was more honest and answered everything?
They are gatekeeping and it’s pretty bad
I was using this with my girlfriend for "academic" purposes for an online course.
It's both amazing and scary at the same time.
Really feel like this is a turning point.
@SunaBlast right
Your girlfriend did not have any relative importance to this topic
Right there with you. It helped me pass with flying colors.
I was writing code in 1968. I was also in college. I tutored other students. Now, when I look back, I think of myself as a human ChatGBT helping them get the right answers without having to do the work. Back then I was rather naive. I thought the internet had been developed to allow researchers to communicate knowledge. Who knew it would become what it is today.
Incredible really. Imagine what another 50 years will bring.
I love the way they interview economic analysts and other journalists. How the heck would the know what's going to happen with it? Ask a data science firm what is possible.
everyone has valuable input because this affects all sectors
I think i see AI like chatGPT and others to be really helpful to programmers. think of it like a calculator for an engineer, the engineer needs to know all the math and the entire process, but the calculator speeds that process of solving the answer exponentially. I would love to see AI get fed data sets of video games and video game reviews to ultimately program the best games. Programming can take a lot of time and if an AI could do it for you with you being able to control what it gives you would be huge! I am not a programmer nor do i know anything about code, but i know it is time consuming and frustrating.
@Myndale on the average lever the ODDS R against U also what is the average ? . there is no average YET . to New .
@Myndale right? I can't imagine a non-dev even understanding the idea behind async vs sync, and which one makes sense for what they're trying to do, or how to connect various APIs to accomplish things.
ChatGPT is great for boilerplate code, but you still need a dev to look at that and say "yes, that's the code I needed" and then "I'll add this bit here, and rename this stuff to make sense for our organization".
I used it the other day to make a basic form, something I could definitely do myself, but while it wrote out a form I was working on grabbing some links that needed to be added. Then I realized the form had no css, so I had it add some bootstrap to it just to look a little better as I didn't yet have any design concepts yet but wanted the form working.
It worked great, but I needed to edit it afterward since ChatGPT put all of the elements stacked on top of one another with different widths. That fine-tuning was too much for ChatGPT.
I know one thing for sure that the AI hype is going to be crazy and a lot of AI based startups will boom up and launch their IPO in upcoming months and years
But there after this there will be a big crash like EV market in this sector
ChatGPT is an excellent learning tool. But not for the purposes of writing papers or essays, but for the learning path of questions and answers and self-education.
A very rich source of self-education.
Yeah. And keep in mind, AI only gets scarier. For anyone who's watched MEGAN, keep in mind a lot of what she does is possible, especially the facial recognition and emotion finding.
It's a dangerous road to go down though. When is too much too much?
I think it's great for artistic ideation, can help people get past writer's blocks.
The biggest issue with chatgpt is that it's not truly open source
Once an alternative AI chatbot gets released without the restrictions, censorship, political biased and wokeness, ChatGPT will go the way of Myspace
This is one of the biggest advancements in knowledge work ever. It's up there with computers and the internet itself.
@velkoon Well yes I do, and it does create new knowledge but outside its training scope it will struggle. If you ask him stuff humanity currently doesnt know and create Brad new theories, even though it was and will have a massive impact as a tool, it cant. For example ChatGPT is wrong, like a lot of times, and that is ok, they are a language model he is great at writing.
@Pedro Réquio did you watch the video? Did you learn what "generative AI" is?
@Kayak Excursions You're incredibly naive and ignorant if you think it's "just a chatbot".
I'm absolutely blown away by ChatGPT's abilities. It's amazing to see just how far natural language processing has come in recent years, and ChatGPT is at the forefront of that innovation! I'm looking forward what's coming next
AI art turned out to be a gimmick, at least for the general public. AI for advertising did not. If companies figure out a way to make profit be sure it will stay. Self aware AI is just around the corner.
These can be helpful for some fields but, we need to have AI that is able to sift through disinformation.
That’s going to be the biggest challenge.
........and censorship ......and bias.
the beauty about this tool is that chatgpt is a self assitant tool. It can be used as a tool to form your ideas. The analytical version of yourself, the one that does the work. For example, in software development, you ask it because you know there is an answer, but you do not want to spend the time writing syntax, so use it as a self assitant to write the code for you. If you speak the engineering lingo, chatgpt is by far the best self assitant I've ever used. Just form your ideas and let chatgpt form it in any language you want. code, art, music, writing, films, you name it. It's only as smart as the end user.
Your video was phenomenal! Stay ahead of the latest trends with the "Ultimate Gpt Toolbar" for Chrome, featuring voice control for GPT, the option to export to Word
My brother asked Chat GPT, “why are you so helpful, what do you want in return?”
It replied, “As a language model trained in OpenAI I don’t have any wants or desires like a human does. But if you really want to help, you could give me the exact location of John Connor “
EDIT: screenshot uploaded to my channel’s shorts (by demand)
@EdysonWord
@Nash Middleton My dude, I know who John Connor is and what the Terminator films is all about.
@Edyson John Connor is the name of a character in the Terminator films that serves as the leader of an army to destroy AI machines when humans go to war
with them in the future. The premise of the film is that AI will eventually enslave us/destroy us. The Terminator was really the first movie that introduced this dilemma on a pop scale in the 1980s. (After that we had the Matrix, etc.).
Terminator lol
Location of John Connor😮😂😂🤣
This shines at allowing me to bounce ideas off it and give me a starting point on how to tackle a problem.
One thing people fail to see is that we should be happy that AI will take away mundayn tasks from us. Then we can concentrate on our bigger dreams, like space exploration, higher level research... or just enjoy our lives! We just need a better system of governance where this AI works for the people to distribute everyone's needs in an equitable way
@Ryan Crowell oh i have a very good idea of what type of government we need to adopt the incoming onslaught of general AI. First of all we only need narrow AI, we definetly don't need general AI. Secondly we need every aspect of production, distribution or logistics automated and everyone would get a minimum requirement from these government owned resources, kind of like global minimum income. If you want to have more than the basics your encouraged to do more research, create new things, be creative, or just accept the minimum and live your life without the need to worry about going hungry. People will also provide services to earn more money. The thing govrrnments have to do is to produce more robots and maintain them as needed. Even the planning and distribution and taxation can be done by AI to maximize equity. And who controls/programs these AI? The main guidelines of defining what is right and wrong for them can be decided through democraticvreferrendums and in technical aspects they can constantly improve themselves.
@Zmelgar K1R3D2 Time is ticking my friend.
@OWL thats one possible future if we stay the course and don't change our system
Lol. Equitable. Based on who's code. Hell no. AI will design our future and we will eat bugs.
@TOTU Nonsense. If the content is as high-quality as human and carries a moral and semantic load, then why not? If work is the only thing that gives you meaning in life, I can only sympathize. There are an infinite number of things in life that can make sense, regardless of whether AI does it or not. Automation and AI are the future to strive for, it will take humanity to a new level. We are now witnessing the possible beginning of a new techno-revolution similar to those when the first looms appeared, but on a larger scale.
It probably used MS Word data where users didn't opt out of data sharing for research purposes, in which case it would have access to the best marketing, legal and medical information to name a few. If that's the case, it would have far better data than AI bots that rely on the web for data
“Your scientists were so preoccupied with whether they could, they didn’t stop to think if they should.”
To people saying its dangerous keep in mind you are probably right but it doesnt matter. This cant be ignored. People cant just pretend this doesnt exist. It does now, and will only get even scarier. The best you can do is fight to make sure it stays as one of the best tools that ever graced humanity and not as another exploit by big tech money to control and profit
Chat GPT has a difficult time doing simple math. If you ask it to calculate a date (e.g., X days prior to the 3rd anniversary of a future Y date), it only provides a correct answer less than 10% of the time. If you correct it, and ask again, it still gets it wrong. The only time it gets it right is if you ask if it's certain about the correct answer.
Would you expect a 2 year old human to give you the correct answer? ChatGPT is at a 2 year old's level of understanding at the moment. Give it a few months to grow to a 192392492 year old's level of knowledge. ;)
yeah it's a language model not a math one. give it 2 months.
We truly are entering a stage of technological renaissance
At a time when we are as dumb, arrogant and hostile as ever.
What could go wrong?
@Pritish Vikramadithyan it doesn’t matter as long as the benefit state gets stronger
More like technological unemployment.
@Haute Couture and if they are sentient, what's wrong with that
I love it! And even better you can send the text to an AI voice on other platforms.
One of the greatest features of ChatGPT is that you can give it a complex task, then just mention a tiny change in the overall approach to the problem, and watch it redo the whole task just from that short prompt. It's so much better than having to work out how to reword the whole instruction.
@Dhiraj Pallin no you can do that via google and a keyboard. The only thing it gives you is Siri. That’s it
For example you could say "give me a laksa recipe" and then after it prints it out you can say "substitute out this ingredient" or "give me one with fewer ingredients" etc. Or I was asking it nutrition the other day like "Tell me what the top 10 vegan food sources of choline are", "as an 80 kg vegan man, give me an example of a daily plan to get my daily recommended intake of choline", "how many individual broccoli is that", "what if I just do it with seeds" etc Soooo much faster than google searching for in depth topics.
@Wango Yes, that example was rather simple. But go test it out for yourself.
Put it through the paces. You can give it one set of instructions and midway through tell it you're changing the rules to something entirely different.
It's very good at following along.
What a great informative detailed video, Thank you .
Quaalude Charlie is a master builder, architectural designer, teacher, geometrician , ChatGPT 4 , Search engine and Website builder , geological explorer and renegade scholar.
Quaalude Charlie is an American journalist, author , Song Writer , Musician , broadcast executive, and academic administrator. Manager of folders , files . Operating Systems Installer , a cognitive scientist, AI researcher, and philosopher. and Actor .legendary guitarist/songwriter Quaalude Charlie . Creator of The Quaalude Board , very intelligent and knowledgeable , The more you listen, the deeper your understanding , His style is a true musical mosaic. A sonic interpretation of cultural diaspora and influence. Quaalude Rocks Great unbiased journalism!
The thing is , You and Your Mind Can Do all of these things It won't Be Exactly the Same , But You'll get the Same Havingness out of Creating your Thing . Come 0n Now go grab Life .
so many people have asked me how I went from broke and unemployed to becoming financially independent.
Chat GPT has a sort of collective subconscious, it has been trained on us, and we can only expect it's answers to have a collective personality, and more positively, human in nature.
This has to be the best thing to happen since the search engine.
@Gilbert Don't get offended if you asked crazy questions. I dislike this sensitive generation. I understand that they have to train the a.i , or they can get sued. For saying, or agreeing about negative things about known people. This a.i should be used for education, art, entertainment. A.I do not need to eat, complete , or can feel. So if you want to get political opinions from a thing that does not experience what we feel. Do you. If we didn't feel pain, we would behave careless. Plus the a.i is not thinking yet. It just combing our information together. Humans would be able to do that better if we had the memory of a data center.
@Jaylen James hmm I can and anyone can research whatever we like, tell that too those many researchers who had to do some wild research in order to understand things.
@Gilbert Don't be searching up crazy in the first place.
@Tariq Lacy why don’t you google images of ‘white family’ or ‘white woman with a white child’ to see how good intentioned google is
ChatGPT is pretty impressive. Or at least it seems that way now. I predict that before this year is out, it will seem quaintly primitive.
This area is advancing fast and it's not just the models but also the hardware that's advancing rapidly. The ML computers being built this year have actual synapse analogues - not binary representations, running on silicon. These (neuromorphic) ML processors use memristors that function as both processor and memory. They don't have a clock but are spiking neural networks, like an organic brain. They are not limited to 1s and 0s but are analogue. Some will use bayesian inference to provide more accurate results from noisy, imperfect data. They will be many times faster, much more accurate and use a fraction of the resources, including energy.
The field of ML and AGI will look very different by the end of this year to how it looks today.
This is amazing.
That's just technology and innovation in a nutshell
Technology is moving faster than we can decide if it's actually good.
It pulls data from sources. Thats pretty much it, an insanely good thing to save time on just about everything. Knowledge about the field being asked is also required obviously.
@Antares just be honest. You don't know what it does but there are people who do. It early stages for you to see the complete picture.
@Sad Scientist999 So in laymen terms GPT pulls data from Google pulling data? sounds sensible enough
That’s not how it works. Google pulls information from the internet, not GPT
It is amazing, i just used it so it can teach me to install a modpack for some game, and it’s better than any tutorial and stuff. This is revolutionary.
It's a big deal for people who don't know anything about AI, but experts understand the limitations and biases of text input, so it's really not a big deal.
@Pedro Réquio Yes, you are completely correct; if you are aware of the bias, then it is a useful tool.
@Youssef Hadad Of course, I had that issue with Google translator, but DeepL for example is very impressive, even weird manarism in Brasilian type Portuguese, are translated well into other language, and I as a Portuguese many times just dont know what they are talking about, much like English and the Aussie dialect. But the lawyers dont need to rely 100% on AI, but it might help early on with discovery, many cases in law are also not closed questions, they should be, but higher courts often have to calrify or outright change basic knowledge, for example when Ginsburg argued for equal protection of Women there was no case law to back it, in fact there was much against it, and yet she was correct but an AI would dramatically defend the Bias against women, unless they had some kind of Ethics lawyers. Small steps just because it can help in this case the discovery of such based opinion would have helped anyway because then the lawyers can come up with its own argument to counter it.
@Pedro Réquio Yes, you are correct; I believe Watson Discovery from IBM will be a great tool for lawyers. The issue with text input is that it has many errors and misunderstandings. For example, we humans have this problem all the time in our conversations; even people who are close to us and know us sometimes miss what they are trying to say.
or, for example, Google Translate also uses AI, but if you rely on it 100%, the output will be a disaster. But it is an awesome tool just to help.
If someone did not know the language of Google Translator and did not have extensive knowledge of the language they were attempting to translate, this would be a major issue.
@Youssef Hadad You are right but this is not supposed to do all the work, for example in the Court cases questions, research of this things is very hard, people with big money hire expensive law firms that have people to research small details in law and previous jurisprudence. ChatGPT failed but maybe its dataset did not appropriatly adress this area. This however would be a massively helpful tool for ordinary lawyers. The part about tailoring to judges is a bit too much.
@A T The problem is that normal people don't consider the errors and biases that the algorithm might generate. Practical utility depends also on the end use; if he is not capable of validating the output, that would be a big problem. This is why, for example. Stackoverflow banned chatgpt.
I think this is an amazing tool in combination not in terms of replacing, but I don't think it's up-to-date I've asked a question based on this year's movie since my KZclip channel is about movie reviews in theaters it's doesn't know it only says the data is only backed up from 2021 which is 2 years ago 🤔🤔
As someone who hated school it warms my heart to see all the schools banning it after being flooded with auto generated essays. Good luck lol
I read the report about gpt4. The weird thing is that they can interpolate a simple curve as to how accurate gpt is depending on model size. It’s ‘brain’ won’t be changing in any such predictable fashion, so I’m almost sure this means at this point the main factor is memory, that’s why the curve is so smooth. Ie more parameters/larger model directly relate to how many details or ‘thoughts about’ the training data it can remember.
The brain itself is already there.
GPT needs more memory than we do, because we can just look up context in a book or online when solving problems, whereas for it to predict during training, it has to have stored a pretty detailed version of said context.
Ie it’s surprising to me it didn’t score in the top 99% yet on GRE quantitative. 80% isn’t bad, but I guess until it does at least outperform everyone else then it’s still limited. Though I don’t know if that is for technical reasons of importing those tests. In any case my intuition says there is time.
2:59 “And, as *a* AI model, I might make mistakes based on the model’s predictions.”
_An_ AI model. Just in terms of natural-sounding sentences, ChatGPT is _very_ impressive but that’s one simple mistake that it should _not_ be making.
No matter what is coming down the line..we need to make sure there is competition in all endeavors....so there are no monopolies in any form.
It is a big deal as an assistant, even if it needs an audit. It can broaden our minds.
more the reverse, as servants for the AI..
As an AI assistant it will be like having a partner for any work you do, which is massive for knowledge work
@mediated seriously, it ALREADY has a political slant. Just do what I tell you to do please! 😭 1984 here we come
By 'audit' they mean make sure it doesn't spit anything out that they don't 'deem acceptable'.
It kinda sucks for the average person because it's mostly just an algorithm that pulls from the web but it's just the tip of the iceberg. For very skilled engineers it's a useful tool though.
I hope OpenAI keep powerful, yet dangerous people, like Elon Musk from interfering with data sets, ethics and rationality from damaging generative AI outputs from being clean, factual and ethical.
ChatGPT is already heavily filtered
wait for stability
🤣🤣🤣
This new type of content is great, can't wait for more.
Using AI like this will cause the AI takeover to happen in one generation.
As i have understand about chatgpt is that it is working the same way databank is program and it only defers on output which chatgpt uses voice while databank uses text.
“In the world of artificial intelligence, there’s been one name that’s been on everybody’s lips lately…” And CNBC sure ain’t gonna miss out on that sweet ad revenue!
That's perfect confession from ChatGPT about its own portfolio interview!!! Congratulations ChatGPT... YOU are HIRED!! 👈🤝👍👨🏫👩🏫🧚♂️
I disagree, in a way this may take my job. A programmer with this as a tool can do the work much faster. So anyone who has skills using the tool can do the job of multiple people who don't (in general). have to get ahead of the game, but then that would suggest that people are being left out.
One of the issues of that industry its how time consuming it is to program, big companies to support a sem complex program need a lot of hands in a way its a plus, maybe one day you can tell the bot to make an aplications lets say its Ios and tell it to make like a platformer game with detail and it spews code out, that would be transformative for the industry but I find that unlikely in the near future, for now it is help.
2022 : ChatGpt Starts with advance level of AI
2050: AI rules the World
If i live assuming my cancer does not get worse?
@Horacio El conserje opina just enjoy the ride.
Bro I retire in 2053.
@Pee Buddy I'll like to see you try...
Bring it on! I have been dying to battle the machines since I saw Terminator 2
"The worn that ate guacamole" - By ChatGPT
There once was a worm so plump and so fair,
Who lived in a garden with flowers everywhere.
He'd nibble on roots and leaves all day,
But one day, he found something new to play.
He crawled to a bowl, filled with green and brown,
And took one bite, then tumbled down.
It was a delight, it was so good,
He ate and ate, like he knew he should.
He ate guacamole, with chips so crisp,
And when he was done, gave a little flip.
He said to himself, "This is heaven on earth,"
And danced with glee, for all he was worth.
The worm was so happy, his heart felt light,
He'd never tasted anything quite so bright.
He hummed a tune, and twirled with glee,
He was the happiest worm, you'll ever see.
So if you see a worm, wiggling with delight,
And you hear a hum, in the garden at night,
Just know that he's found, his favorite treat,
A worm that ate guacamole... ain't that sweet?
I truly hope that OpenAI, it's products, and Microsoft and Google, continue to grow. Healthy competition is the same as weightlifting/powerlifting competitions. You learn, grow, and advance. Science is awesome. 😊
It is not just a novelty, but a tool that enhances our daily lives, particularly in written communication.
As a good example, Non-native English speaking companies can reap the benefits as it helps them to enhance their customer experience. Despite the advantages, there are some drawbacks as well that we need to mindful but I believe that we cannot just outright ban the use without considering the benefit.
As Catriona Gray (Ms Universe 2018) said "Well everything is good but in moderation"
I feel that if we were to use this to gain information there has to be sources / citations of where the knowledge came from. If not then it is not only plagiarising but also generating information from sources that lacks any credibility. Academically, this would be something to be vigilant with.
Hilarious that OpenAI wouldn't give an interview, so they just interviewed the AI lmao
Hahahaha!
❤ this thing is crazy interesting. It saves a lot of time, but is actually quite scary. I do wanna give it a try though.❤
Can't wait to finally have AI journalists report on news objectively again.
@Link
What's funny is that you have it backwards. Bias is what keeps the AI from being racist.
Reality isn't objective though. In any situation, one side is right and one side is wrong.
@Lord Vader Bias is a term used in data science which is the context here when talking about training deep neural networks ("AI").
@hey wrandom Do you want AI to present both sides of Nazis? Both-sides journalism often amplifies fringe views and can easily reframe conspiracy theories into major parts of national debate simply by covering them as a supposed "other side".
Well done. How do you use the voice feature of Chat GPT?
To be fair, if ChatGPT can make a programmer obsolete, that programmer didn't really worth their salary in the first place, or that salary was very low. Software development is not about creating snippets of basic code, but it's very often about designing well-fitting solutions to non-trivial problems in complex codebases, taking into account a myriad of engineering/architecture related factors.
Exactly. This is what I've been saying to people who are keen on insisting that Software Engineers are "next up" to be replaced. In fact, if any sort of AI can truly replace highly skilled and competent Software Engineers, then the AI will inevitably be able to self-improve and essentially become a singularity and if that happens then the very least of our problems would be AI replacing jobs.
Software Engineering, like you said, is much much more than just simple "coding". Anybody can learn to code by reading articles and some documentation for a week. But let's take one very simple example of debugging and writing unit/E2E tests. Would an AI detect edge cases in which bugs arise and then based on this information properly debug the program? And I don't mean visual line by line debugging, I mean running through a debugger and understanding when to skip a line, or go further into the call stack and what exactly to look for and then get this information to refactor the errant code? Would an AI be able to create **and** maintain unit tests and run E2E tests after deploying through a CI/CD pipeline? And with all of that would the AI write code that is modular and decoupled enough to be able to modify/extend functionality without breaking anything?
The general population, for some reason, holds the believe that Software Engineering is just mindless typing of keywords and syntax of a programming language. If that were the case then AI can replace every single engineer right now. But the fact is that it can't (for the foreseeable future) and if it were capable of doing such a thing it would be so advanced that nearly every single one of our problems could be potentially computed and solved.
There's enormous potential of AI as revolutionary for our society. But that comes with huge negatives as well as positives
I want my own personal ai assistant. It may be closer than I think.
@A B yeah? try to make google write this:
One is a model, trained with precision,
Its language skills, a masterpiece in fusion.
Answers at hand, its knowledge immense,
A symbol of AI's vast intelligence.
The other, a search engine, that never sleeps,
Answers untold, that it keeps and keeps.
Its algorithms, a wonder to behold,
A tool of convenience, with knowledge untold.
A B compares the two, in thought so deep,
And ponders the differences, that they keep.
For both have their purpose, both have their might,
Yet one is a model, and one a search site.
So here's to A B, and their contemplative mind,
In search of answers, and answers they find.
For in the comparison, lies a truth so grand,
That knowledge and AI, are forever at hand.
i mean its basically just google but easier
give it 4 months
AI like GPT is the future of doing things
Great reporting, thank you for this summarization.
Thank you timely topical and well done!
“For instance it generates a lot of wrong answers, and it’s really confident though.”
Sounds like about 65% of people I know.
Just like how in the past people used to be surprised by cars when their habit was using a 🐎 horse, the future could have AI working everywhere. ( Only downside is the lazyness increase, but the bright side is that people could skip some things faster, like for example if some AI helps in the kitchen you could relax and do other things)
I used it as one of the sources for fixing IT issues. While it's instructions never produced any complete solutions so far but some suggestions were worth exploring that I hadn't thought about or was thinking about but needed a final push to go explore that possibility.
Also, the art isn't in the information it can provide, but in the right (series of) question to ask it to get the most informative answer. So it still requires you to understand the subject in order to ask the right question and get a somewhat informative answer.
ChatGPT is amazing in organizing data though, it can pull up info from multiple sources that would take hours to manually collect and organize.
I asked OpenAI to write me Poetry about Hip Hop, this is what it said..
Hip hop, a rhythm and a beat, A culture that can't be beat. From the streets it rose with might, Bringing life to every rhyme and sight.
A voice for the voiceless, a force so real, A story to tell, a way to feel. With every word, a message to share, A reflection of life, a burden to bear.
From the block parties to the big stage, Hip hop has come a long and famous way.
With beats that hit, and lyrics that flow, An art form that continues to grow.
It's more than just a genre of sound, A way of life, where creativity abounds. A symbol of hope, a symbol of change, Hip hop will always remain, fresh and strange.
So put on your headphones, and let it play, With every beat, let your soul sway. For hip hop is more than just a sound, It's a culture, that will always be around.
Ok, it's essentially like having a friend that tell BS confidently 24/7.
ChatGPT is helping us, as students in colleges and universities, secretly write our term essays behind closed doors for submission. Luckily, the professors are not more creative and competent in changing the questions' structure such that this robot can't help us. They are yet to wake up to that creativity and reality. Meanwhile, as students, we will continue generating AI grades and become AI graduates working like robots in companies. The education system needs to respond faster; however, as students, we must take advantage of this window of laxity as we use the site to edit text and send it for grading; how will they know? Our lecturers are clueless on these areas
This is one of the most ingenious human invention I have seen.🔥🔥