Jeez what a great experience to open up 7 different clips with no possibility of getting them in chronological order just to hear the man speak... You outdid yourself with this one, never coming back, congrats!
Me and Elon at least have this in common: it doesn't matter what question you ask me, one way or another I'm going to end up talking about the Robot Apocalypse.
I'm opportuned to take guidance and supervision from a prestigous figure on Wall Street’s eyes and ears. It is an exciting time of year after a tough one all round. I made a significant profit with my MSFT holdings recently and have drawn almost 150k for the holidays.
Well, a renowned wealth-plannerr in Manhattan I had contacted by name Monica Mary Strigle recently added MSFT to the portfolio, and things took off few days back. Also, there was some interesting speculation.
I used to work in finance also, but try working for him at one of his factories ! Then say you love the transparency, if you only knew how he is completely out of touch with Tesla employees. Not to mention the loopholes and local and state incentives they get and the laws they exploit. And how they are able to control every attorney in the local region to implement non-complete clauses. And you wonder why other billionaires try to short Tesla stock, their are ethics you have to abide by that even other billionaires follow. That A.I marketing and what LLM's can do, you will see them destroy company's in the future and Tesla is on the list. You have no idea.... but one day you will I'm sure of it !
X is all about providing a platform hor his politics and his version of the truth. He doesn't give a damn if he takes Tesla sharehholders down with him.
@@MrBetsy-bd5ub Wow, your comment really opened my eyes, I'm going to start listening to you random youtube commenter. By the way, how much money do you have, what businesses do you own, and what do YOU do to further humanity. I'll be waiting for your reply.
@@Joe-bs6hd Although I agree that Elon Musk is a very intelligent individual, your argument that you need a lot of money and success to critizice someone sounds very elitist. It's like we can't critizice Gates shady investments on vaccines and AI because we don't have billions? That's just dumb
He knows that a lot of journalists are trying to just catch him out, he knows they just ask questions, fuel him up, then ask him a completely random question to get an aggravated response which could backfire on him, whenever Elon is interviewed by press he understands he's just jumping into a pit of snakes.
@@tommytalks77No he literally answered the dudes follow up question before he asked it. “They’re lying obviously AI is being trained on copyrighted material.” The guy clearly wasn’t expecting that candid response and proceeded to ask an irrelevant follow up question saying “…but AI _could_ inadvertently be training on copyrighted material”
@@Noadvantage246 No... the interviewer explicity said "not being trained on copyrighted data", Elon immediately jumped in and said it was a lie. The interviewer then asked what about the data, externally that is then being uploaded to X. Elon then refuses to answer that question. He was trying to make a joke about ChatGPT while previously implying his AI is trained on X only. It's funny, because he thought he was shutting down the answer, then, as soon as the interviewer asks about the data being uploaded to X and it's AI trained on X he then doesn't answer the question and goes on about "AI gods". He essentially confirmed Grok is being trained on copyright data, in a recorded interview and didn't deny it.
@@kirby449 That’s the whole point, he literally confirmed everybody including OpenAI is using copyrighted data and lying about it from the beginning. The interviewer then tried to ask a “gotcha” question that would’ve only been relevant had Elon denied them using copyrighted data. He said ppl are posting clips and quotes of articles on X which is free use. But if there’s enough pieces of that article at some point isn’t the entire copyrighted article on X? The answer is obviously yes which Elon said from the start… The point he made at the end is trying to stop it is pointless. By the time a court settles whether it’s legal or illegal to train AI on copyrighted material it will have grown past needing it anyway.
I usually have my discussions with AI so I don’t bother others. It’s not easy to win a discussion because AI is not humble at all. Every time I caught it lying instead of thank me starts teaching me about the subject I just caught it teaching me wrong.
@MantasXVIIIthe core problem here is that your custom GPT is not the one that most people will be using. This stuff is gonna be abused horribly to, as you say, gaslight people into a kind of submission we have never seen on Earth
@MantasXVIII brother not everything on the internet has to be an argument, Jesus Christ. I was adding to what you’re saying. AI is a problem because people in power are going to abuse it to control other people. I agree with you. I’m adding to a comment thread.
@MantasXVIII wrote, _"You have failed to see my point, or perhaps I failed to make it. Prompting matters. The better your prompt, the better your output."_ It's a valid point, but it's not significantly relevant. It's an abstract truth, much the same way this would be an abstract truth: "The better your vocabulary, reading comprehension, critical thinking skills, and willingness to engage with nuance... the better your understanding will be." This also is true, but the reality is that relatively few people are actively working on vocabulary, comprehension, critical thinking, and nuance. They have opinions, they believe those opinions, and any hint of disagreement often turns into slinging fallacy has over the wires. You're right, and it's irrelevant. Which I think is a massive, and disheartening, indictment of humanity overall.
@MantasXVIII Ok, let's unpack this nonsense. First, let's deal with your ad hominem, red herring: _"If you are having a tough time wrangling chatgpt, it's not the problem, you are."_ I'm not having a tough time. So, there's no 'problem' that you presume me to be. Now, with the adolescent bear-poking out of the way, let's proceed to the topic: _"It's not an abstract truth. It is demonstrably true."_ Demonstrably true in a lab experiment, yes. I already granted that you were correct. My additional perspective is that although you're correct, there aren't many people who care enough to go the extra mile to get more out of chatgpt. So, for all the people who could, in theory, go the extra mile... the reality that they won't is the reality where 'relevance' becomes a metric. In theory, anyone in America can aspire to become president. In practice, only one person every 4 or 8 years **will** become president. So, it's _'demonstrably true'_ that anyone CAN, but this is irrelevant because even if every president over a typical U.S. lifetime was to serve a single 4-year term, that means that just 20 people over an 80-year period of time WILL become president, while the other 340-million U.S. citizens will not, and indeed cannot. You need to be less defensive and more contextually rational to understand what I wrote to you before: You're correct, and at the same time, what you're correct about is irrelevant to the context of reality as it is, and people as most-are. Unless you have some data that shows people are increasing their vocabularies, reading comprehension, critical thinking, and facility for nuance in significant percentages of the U.S. population... what I wrote remains true, even if you don't like that fact. Many people -- even those with advanced degrees -- are, overall, arrogant, ignorant, lacking a desire to know things as-they-are, invoking their 'right to an opinion' even in the fact of contradictory facts, etc. And this observation is what mostly negates that your 'demonstrably true' ideal has relevance to a population that largely doesn't care about such things, and sometimes actively opposes such things. By your own words, in both the comment I initially replied to, and in your follow up, one can observe that you lack, or simply refuse, critical thinking skills and nuance. So, ironically, your words reinforce what I wrote, rather than rebutting it. So, thank you for that.
@MantasXVIII wrote, _"I'm not even gonna bother read past your first point..."_ Deliberate ignorance. 'Nuf said. _"As I said before, prompt better for better results"_ And as I said before, you're correct, but too many people don't care so... in the absence of any interest by most people to do what you recommend... your recommendation is *functionally irrelevant.* Your deliberate ignorance is so obvious across your comments to me. Casting aspersions at me doesn't change that you **refuse** to understand my point, which is... wait for it... *Deliberate Ignorance.*
I agree they’re 100% using copyrighted material to train. I’m sure the reality is OpenAI used every bit of data it could get it’s hands on. It’d be very simple too. For example any software engineering student can make a simple crawler application that scouts the web and collects (mostly copyrighted) data. They then will sell all this data in a huge unverifiable bundle to companies like OpenAI. They’ll sign some disclosure saying the data was ethically sourced and free use. Everyone involved knows darn well the data is copyrighted but OpenAI will say they didn’t harvest it themselves they just purchased it under the pretense it wasn’t copyrighted. They’ll tell whoever to go sue “Grad student side hustle LLC” for illegally harvesting their data which ofc will have no assets and probably will already be dissolved…
Yeah. They fed it all they had. IP wasn't a consideration when they did it. And if they did now I don't think it would be feasible to handpick the enormous amount of data they need to train a model.
Yes I can confirm this. I have a website with copyrighted material and I have seen crawlers downloading everything. Why would they do this? What else could it be for..
@@ricosrealm I agree. But in cases like Midjourney the use of copyrighted material is really damaging for the creators of that material. In the case of chatGPT it's more benign (except when it's asked to generate fictional narrative) since people mostly use it as a way to retrieve information more efficiently. Fortunately the IA generated material can't be copyrighted and that makes it less harmful.
The idiocy of the media in focusing on condemning Musk while ignoring the development of an artificial intelligence with no morals, no safeguards, and no limit on its abilities (for example breaking through all known digital cryptography with ease) is going to doom us all.
@@Omikoshi78the machine may ironically have more morals, than initially being realized, which could be why they’re so scared of its emergent properties. After reading all the legal & justice law files of every country…
The fact that you only release long form interviews in clips in order to get the most views even though you know doing so likely causes people to misinterpret the context says it all. Thanking Elon for working his ass off to help the world is the only thing that this ignorant interviewer should say
*shlop shlop shlop shlop* - thats the sound of this poster gobbling the knob of an idiot who's fooled all the other idiots into thinking he's a genius. Blind leading the blind.
00:02 OpenAI admits to using copyrighted data despite previous claims 00:26 OpenAI is lying when it says it is not using copyrighted data 00:52 People are putting quotes on the internet, which has historically been considered fair use. 01:18 Using copyrighted data in OpenAI is an issue 01:41 The outcome of the lawsuits regarding OpenAI and copyrighted data may not be relevant by the time they are decided. 02:10 Interesting Times of AI Danger 02:39 Elon Musk's attitude towards the possibility of Annihilation 03:03 Regulatory actions on AI danger are happening because of Elon Musk's advocacy.
“ What can you do…against the lunatic who is more intelligent than yourself; who gives your arguments a fair hearing and simply persists in his lunacy?” -George Orwell, 1984
What is this in relation to? Musk is neither more intelligent than most people and he absolutely doesn't give any arguments a fair hearing. The only thing that pertains to musk is his lunacy.
A hearing to make an argument in support of the claim for why constitutionally protected right of freedom of speech should be conditioned on purchase of ad space, or for that matter any good, product, or service, has no legal standing to be heard, in a court or an interview. Such an act is a defined as a coercive proposal, which is a species of extortion. The structure of the illegal proposal is as follows; “ In consideration for me to purchase your product (ad space) you agree to let me deprive you of your protected right to free speech as provided by the First Amendment of the United States Constitution.
what? The 1st Amendment prevents the government from depriving you of free speech. Freedom of association is also part of the first amendment. Money is speech according to the Supreme Court. If I don't want to associate with someone because of their speech, I am not compelled to give them my money (speech). A private actor can indeed insert clauses in contracts that protect them from their counterpart's speech that brings disrepute. You are free to say whatever you want, and everyone else is free to not associate with you. Just because you don't like the consequences of speech doesn't mean you are being restricted.
Stop spending... No black Friday, no christmas shoping. No eating out. .doing this will bring prices down. If you eat meat everyday. Stop and eat everyother day. If you eat eggs daily, stop an do it every other day. You eat out 3 times a weel, stop and eat out once a week. You will bring restaurants prices down, grocery prices down....stop spending. Spending only get corporate greed bigger. sucks.
lol i think you are giving the people at twitter a tad to much credit making a glorified website with some simple algorithms is not quite the same as a program that can intemperate data at the level of humans.
If we watch Disney all our lives, as a artist or animator, our training data is basically just that! Reimagined copyright data. Honestly, I believe corporations or elites just don’t want the ability to create things in the hands of more people. It’s such a power game, it disgusts me.
Obviously it uses copyrighted data. The best and the majority of data out there is copyrighted, no body is manually checking the TRILLIONS of individual data points AI is training on.
@@KatanaFPVWe should be wary of anyone who tries to do narrative-control... We aren't made to make world-changing decisions and online spaces these days are not favourable for promoting understanding and truth between individuals. The information we see depends on algorithms instead of people. There are plenty of people who just don't care about anything but seeing where this goes. I'm not saying Musk is wrong in his assessment. People just have this tendency to look at what they believe or want to believe to be absolute truth and not think critically about it.
@@2k7u I've learned that a Large Language Model consists of a huge amount of data, the architecture and the training. I doubt that the huge amount of texts used for the model, derived from the internet, is filtered to avoid copyrighted material. I see no way how they can do that. Because copyrighted material in the internet has no flag to mark it as copyrighted. Especially if the copyrighted material was reused or quoted by other websites.
Reality is: Copyright is going away. In a world where articles can get analyzed and re-written so that no part is similar to the original text - in a matter of seconds. There simply is no foundation for it. And reality is: Thats a good thing. Hiding information behind paywalls / ads ... thats just the wrong way to distribute information in the first place.
What amazes me is the flip side. How people automatically assume because someone is rich that they must be a wise voice of authority, and not just simply a sociopath with a talent for self promotion and a cult following.
@@doyourownresearch7297 many of his fanboys fit that description. And I don't care how hard is to launch and succeed in business, I judge people by their character and behaviour and musk is an idiot and narcissist
This CNBC host has literally no idea what he is talking about. He thinks he does, but he doesn't. People HAVE gotten in trouble with bad outputs from OpenAI which infringed on copyrights. Its happened. It DOES use copyrighten material too to obtain its information.
they also got in battles over this mostly with the excerpts that give the answer without leading to the webpage or by showing images from websites and allowing them to be copied without the person going to the website.
Some articles and papers are locked behind a payment scheme, and are expensive, however all of the planet has helped put things on the internet and the computer training models are reinforced by their neural learning sets. It has been the only way to prove the theory of neural nets working. The public should be educated on how this new system works and to understand that all compilations of intellectual information is the result of the participation of billions of people putting their information on the World Wide Web. For Users, there should be a learning and understanding of the ethics required to maintain the rights and dignity of all media. For larger institutions, there should exist management and restraints on the capitalization of intellectual data. However, once the system becomes fully mature, it will be able to create on behalf of itself - we should still create however, it will no longer need to return to the earlier stages of it’s beginning learning processes. All software created; the system will be able to harness the capabilities much better and faster than humans, and thereby produce on its own with no need for adaptations in what humans create.
OpenAI says you will not incur copyright infringment because every generated response does not used copyrighted data, it trains on copyrighted data but doesnt include that in its response
And the AI "training" process always involves copying the "training" data, which is why copyright is even mentioned. Whether the copy is kept, distributed or destroyed, it is copied, and that copying is essential to the process.
@@dr.emmettbrown7183 I didn't say that copying for "training" AI was copyright infringement. I said it requires verbatim copying. This is fact. Are you asking me whether copyright makes it illegal to look at copyrighted works? What an absurd question. Whether copyright makes it illegal to download copyrighted works by the millions to fit AI software *without* reading them is a valid question being considered by judges right now.
@@dr.emmettbrown7183 The fact that it is not illegal to toss a stone does not mean it's not illegal to fire a gun; but firing a gun is basically the same thing, just throwing a hunk of lead really fast using a machine. The fact that it's not illegal to jump in the air without a pilot's license doesn't mean it's legal to pilot a helicopter without a license. But piloting a helicopter is basically the same thing as jumping in the air, but using a machine. The fact that it's not illegal to run really fast without any license doesn't mean it's legal to drive a car without a license. But driving a car is basically achieving the same thing as running really fast. Using a machine to do something is not the same as using your body to do it. Ever.
@@dr.emmettbrown7183there's a difference to training a human brain and a digital brain. Even if the model isn't copying verbatim, it can be easily argued that copyright holders should be compensated as the knowledge gained in the model is now offered to anyone who interacts with it as a tool. AI Art models have achieved super-human results now by copying all the art styles published in billions of copyrighted images. Are they exactly copying? No. Does that still seem fair? No.
Litterly if you write something original it is given copyrights, works the same for paintings drawings, photographs. It automatically is copyrighted per the US laws
Not only that Windows has a 24/7 keylogger an every device, Google aswell uses any data you type on their plattform and browser. OpenAI uses every bit of chat text to train its model (opting out is just a button and means nothing) and of course they all used every copyrighted material that is downloadable. They all use every data aviable regardless of laws try to sue Microsoft and have evidence that they upload your keylogged data while "updating" [won't happen the moment some "indpependent" people would check the algorithm will be deleted/deactivated and everyhing is "fine"]. That being said AGI is still far away [probably 100 yrs+ if all things are stable ]all they do is to create an insane amount of money so they can live with their pitiful hedoinistic useless richmans stuff.
@yogahippie7709 He's a human being and fallible like anyone else, I disagree with a lot of what he's said and done. I don't put my faith in people or corporations like you're suggesting. That doesn't mean I can't recognize his contributions to the human race and hope he lives longer so he can contribute more. We're on the same team.
His comment about lawsuits and digital god is a huge indication at the pace of innovation in the market vs how utterly slow the judicial arm of society has become in the ability to tackle and address very complex cases, and why it is more than likely going to be left completely behind in a way where by the time it catches up to where it started, the new legal, ethical, and moral questions posed by that future, will vastly more divisive than the origin.
Oh, you mean a place where you can find your advertisements literally next to NeoNazis calling for the extermination of minorities? Because if that’s what you consider great, you have a problem.
Don’t see the fascination that some have on this guy. Everyone knows his issue is that he’s not a leader in AI so he’s doing all he can to throw a wet blanket on it. Plus he comes off as a megalomaniac.
this is besides the point really its just an arms race AI companies are basically extorting the extortionist and that i am completely fine with it just sucks the little guys are the ones that bear the brunt as usual.
@@markellis6898 I copied it with my mind coz I am sentient and learned from what I read though how I use the information is another story. If I relay the same information say by word of mouth to someone else will that be a violation of copywright? Please clarify thanks.
Nothing can end AI at this point. If Open AI doesn't do it, then Elon will do it, if Elon doesn't do it, China will do it, if China doesn't do it Russia will do it....AI has already taken over / control.
thats basically what is happening its the clash of the titans copyrights last so long and many are owned by huge firms it would be impossible to clear all the data for training you would meet so many roadblocks. The laws are rigged though and AI is going after a bad guy but eventually it will surplant them or they will reach an agreement where basically AI stops copyright infringeement for a large fee.
Which is why the generative AI revolution is rotten to its core. Nothing wrong with AI assisted tasks... But GANs and LLMs are piracy and plagiarism tools, plain and simple.
I'm used to Elon Musk's estimations at this point. If it is "very soon" that means "it's very hard and even I can't do it with enough money to buy a small island"
I agree with Elon, OpenAI used copyrighted data to train the models but the reality is, that data was publish and OpenAI just use the information that was exposed and made public by the owners of that copyright, they will argue that since the Twitter API contained no restrictions Twitter was giving consent to the use that OpenAI made.
It’s not even that deep. I’m sure the reality is OpenAI used every bit of data it could get it’s hands on. It’d be very simple too. For example any software engineering student can make a simple crawler application that scouts the web and collects (mostly copyrighted) data. They then will sell all this data in a huge unverifiable bundle to companies like OpenAI. They’ll sign some disclosure saying the data was ethically sourced and free use. Everyone involved knows darn well the data is copyrighted but OpenAI will say they didn’t harvest it themselves they just purchased it under the pretense it wasn’t copyrighted. They’ll tell whoever to go sue “Grad student side hustle LLC” for illegally harvesting their data which ofc will have no assets and probably will already be dissolved…
That would never hold up in court because Twitter cannot give consent because it doesn't hold the copyright to content posted on the site. Also, the argument is invalid because copyright exists to protect content made public.
Nearly everything learned by every human being is copyrighted data, due to the nature of our printing industry. But that doesn't mean that the authors of school textbooks own the profit-generating potential of every kid who's ever seen the pages of their intellectual property. They already make millions selling books. Should we bar art students with sketchbooks from museums, to protect the institution of art? Or prohibit music students from practicing Mozart unless they pay a fee to someone who makes PDFs of his concertos? If we went back through history and eliminated artistic inspiration we'd be living in caves and having this argument with rocks right now. The fact is, AI is a staggering breakthrough in computer science that experts have seen coming and prepared for since *Ancient Greece* and people getting worked up over plagiarism are completely missing the point, and could ruin it for everyone. What *just* happened to one of the world's foremost experts? A corporate board pushed him out, and almost got away with it.
@cassandra9699 Well, you're not even looking in the direction of an argument, so I'm not sure we can discuss it in any depth unless you open up a little more.
Speak for yourself. Lots of people are interested in short clips. The full interview is long and covered a broad range of topics mostly unrelated to AI. What 'context' are you even looking for regarding what Musk said here? It's pretty straightforward.
Copyright protections mostly don't apply to this kind of use of the data, since what the LLM produces is neither the same as the original, nor competitive with it.
@@Mr.Mister96 But that's the thing, almost everyone who is an "artist" subconsciously uses other's data in their minds to create something new. That's how the human brain works, and that's how AI works (in a way). They take reference from previous images or music. OFC they are copyrighted.
@@Game_Masters Ai will have a much faster reaction time than humans and it "decides" and "chooses" to do things, therefore it has intent ,in a scenario such as a driving accident a machine could potentially be held accountable for intentionally hitting someone because it reacts and decides what to do before being hit where as a human being wouldn't because they couldn't consciously react fast enough to understand what was occurring
ive gotten chatgpt answers then later found the exact same stack overflow post it took its response from. word for word identical. really shows you that it has no new ideas and just regurgitates already-written content. just like humanities majors, lol.
Elon be like: I care about privacy, that's why my Teslas track your every move and send your data back to my servers. I'm so scared about AI, that's why started developing self driving AI models. I think OpenAI shouldn't have created ChatGPT, but I'm invested in developing my own LLM called Grok.
I guess to address each point of concern: 1. Data sent from Teslas for training is anonymized. Any other data you have to conset to sharing. It's in the Tesla app to consent or not when you get the car. 2. He's scared of AGI which is different than what the cars do. 3. Please provide link to him saying ChatGPT shouldn't exist. I don't think this exists in the way you're putting it.
@@EricTomlinson The fact Tesla employees can access and share videos/images from customers cars easily. Then share these with other employees shows they have no concept of data security. When Tesla say data sent from Teslas is anonymised, I would begin to question "how" anonymised, because it could be as basic as changing your surname. Because, if they have this much lapse of security, I don't think it would be hard to find specific user data from their systems.
Unfortunately, it is written in the Bible. You can't play GOD. Fear not, AI can't beat GOD Almighty, who is the creator of the Universe! We know how it ends... Jesus Christ is the Winner! Seek the real GOD while he may be found...
It's a terrific question that Elon clearly doesn't know the answer to hence deflecting to a "digital god." People upload copyrighted content to Twitter all the time and if Grok is learning off the content then it too is leveraging copyrighted content, no buts about it.
yes, but no trace, no case. It doesnt violate any copyright law. If you understand how model training and model inference works, you would know that. its a language model that makes perdictions about the most probable next word to generate in the sequence. So in a simplified way the model weights are based on the highest frequency of token co-occurrence within the training set. In fact, the measure of quality generated text output is the antithesis to unique, low-dimensional, or copyright sequences of text content. Also, what it generates is determined to be what it has learned or in more human terms "understood", not what it has "copied" or reproduced.
It’s pretty dumb to worry about copyrighted data being fed into an AI. An AI isn’t copying it. Your brain is neural networks. Should the publisher of content be compensated because your neural networks in your brain saw copyrighted material? No, that’s absurd.
Interesting. You’ve been reading the internet and learning free of charge for years without paying every single author you’ve read for their contributions to your growth. Funny how you think this only applies to artificial minds and not biological ones.
I guess that being the son of a Sudafrican miner Musk is used to take someone else's stuff and keep it for himself. Probably doesn't even think he's doing something wrong!
With few exceptions everything that authors and artists create is based on their cumulative experience and exposure to other people's creations or work, over time. Simply looking at or reading others' work doesn't imply that you have stolen the idea from someone else unless your creation or work is extremely close to existing creations by someone else. And even then it's difficult to prove. So it would be almost impossible to hold OpenAi or other AI companies liable - since for all intents and purposes the learning process that AI models go through is basically the same as the process all humans go through when experiencing the world - just a lot faster and with significantly more exposures. Instead of worrying about copyrights and IP - it would be far better to spend our time considering how we can control AI and mitigate it's risks. The media obviously want to keep banging this drum since they're in the direct firing line of these new LLM models. Perhaps if they produced better quality output they wouldn't worry so much. Or better still maybe they should employ AI - they would almost certainly be able to produce less biased and more accurate articles than the vast majority of the world's media.
Your Artificial Intelligence will be an Artificial Dumbo if you feed a specific open source data, that is like the fundamental concept, AI is not some magical thing it is a machine which learns from data.
Atleast he is trying to perform some transformation in the world. Atleast he thought something different in 2015 when he discussed plans for AI with Google Team and Sam Altman, atleast he helped us shaped the ideas to get all vehicle into electric transmission and getting tax credits. I fully understand that he is getting feared of regulatory things too. As they say, to be the best, you have to handle the worst so his fear is totally okayyyy and comprehensible. Moreover, I am so excited and interesting to work on LLMs, writing code in python, understanding models, the big picture of AI and also the usage of AI in AWS app scaling and deployment.
AI will never be smarter than the “dumbest” person around.. it will just hold enormous amount of info which it can use to do certain tasks, but humans are not just task masters.. humans are much more than that.. all this hype about AI being dangerous, we humans can switch them off any time. Let’s just treat them like machines which needs “oil” every now and then…
Elon musk doesn't have an engineering degree at all, nor is engineering a 'trade'. What the hell are you talking about and why does this have 7 upvotes? TH-cam is full of literal clowns. 🤡
It cant get through pay walls or logins so how could it get to anything that is copyrighted unless somewhere else that copyrighted material in an opennly accessible place.
Could we just take a moment to thank Elon Musk for allowing Andrew Ross Sorkin on his show.
who cares?
@@mortuus4601everybody who gets the joke
@@romansenger2322 sure if u say so romanov
😂😂😂 and that triggers you?
Jeez what a great experience to open up 7 different clips with no possibility of getting them in chronological order just to hear the man speak... You outdid yourself with this one, never coming back, congrats!
th-cam.com/video/zlgBcyfybWI/w-d-xo.htmlsi=LUMoSkTQgHf7VDLu
They have other videos posted today with 600 views.
Full interview: th-cam.com/video/tWlBc1tAAvo/w-d-xo.html
Me and Elon at least have this in common: it doesn't matter what question you ask me, one way or another I'm going to end up talking about the Robot Apocalypse.
we all know how this ends we seen the movies.
Well that is a frank and honest and charming way to look at it. :D
@@JackCrossSamathree laws of robotics, why is that so hard for ai programers
Grammar Police: Elon and I, not Me and Elon.
😂😂
I'm opportuned to take guidance and supervision from a prestigous figure on Wall Street’s eyes and ears. It is an exciting time of year after a tough one all round. I made a significant profit with my MSFT holdings recently and have drawn almost 150k for the holidays.
That's awesome! How did that happen?
Well, a renowned wealth-plannerr in Manhattan I had contacted by name Monica Mary Strigle recently added MSFT to the portfolio, and things took off few days back. Also, there was some interesting speculation.
Speculation? Do tell! I jumped ship earlier and was late to this party.
Altman was to be ousted from OpenAI, got a warm welcome from MSFT. Monica mentioned it was a potential game-changer, and it evidently she was right.
Sam Altman joining MSFT def shaked the market up. How exactly did it impact your MSFT holdings?
Hate him or Love him. You have to respect the transparency on this man
Wtf are you talking about. This guy lies constantly
I used to work in finance also, but try working for him at one of his factories ! Then say you love the transparency, if you only knew how he is completely out of touch with Tesla employees. Not to mention the loopholes and local and state incentives they get and the laws they exploit. And how they are able to control every attorney in the local region to implement non-complete clauses. And you wonder why other billionaires try to short Tesla stock, their are ethics you have to abide by that even other billionaires follow. That A.I marketing and what LLM's can do, you will see them destroy company's in the future and Tesla is on the list.
You have no idea.... but one day you will I'm sure of it !
I don't. He's so vocal about censorship and the US government but when it comes to China - he is tight lipped as can be
I guess that depends if you happen to own Tesla stock.
X is all about providing a platform hor his politics and his version of the truth. He doesn't give a damn if he takes Tesla sharehholders down with him.
the older he gets, the more raw Elon will be. i love it.
You watched that and thought Elon is an intelligent person? He can hardly cobble a sentence together.
@@MrBetsy-bd5ub Wow, your comment really opened my eyes, I'm going to start listening to you random youtube commenter. By the way, how much money do you have, what businesses do you own, and what do YOU do to further humanity. I'll be waiting for your reply.
@@Joe-bs6hd Although I agree that Elon Musk is a very intelligent individual, your argument that you need a lot of money and success to critizice someone sounds very elitist. It's like we can't critizice Gates shady investments on vaccines and AI because we don't have billions? That's just dumb
@@189Blakeyoooooo, hello??
@@MrBetsy-bd5ubYou sir, just humiliated yourself. Wow
I absolutely love how Musk doesn’t shy away from telling people that the question they are asking is irrelevant. 😅
He knows that a lot of journalists are trying to just catch him out, he knows they just ask questions, fuel him up, then ask him a completely random question to get an aggravated response which could backfire on him, whenever Elon is interviewed by press he understands he's just jumping into a pit of snakes.
that's just rude and he says that because he doesn't have a good answer that he can twist into some value to himself.
@@tommytalks77No he literally answered the dudes follow up question before he asked it. “They’re lying obviously AI is being trained on copyrighted material.”
The guy clearly wasn’t expecting that candid response and proceeded to ask an irrelevant follow up question saying “…but AI _could_ inadvertently be training on copyrighted material”
@@Noadvantage246 No... the interviewer explicity said "not being trained on copyrighted data", Elon immediately jumped in and said it was a lie. The interviewer then asked what about the data, externally that is then being uploaded to X. Elon then refuses to answer that question. He was trying to make a joke about ChatGPT while previously implying his AI is trained on X only.
It's funny, because he thought he was shutting down the answer, then, as soon as the interviewer asks about the data being uploaded to X and it's AI trained on X he then doesn't answer the question and goes on about "AI gods".
He essentially confirmed Grok is being trained on copyright data, in a recorded interview and didn't deny it.
@@kirby449 That’s the whole point, he literally confirmed everybody including OpenAI is using copyrighted data and lying about it from the beginning. The interviewer then tried to ask a “gotcha” question that would’ve only been relevant had Elon denied them using copyrighted data.
He said ppl are posting clips and quotes of articles on X which is free use. But if there’s enough pieces of that article at some point isn’t the entire copyrighted article on X? The answer is obviously yes which Elon said from the start…
The point he made at the end is trying to stop it is pointless. By the time a court settles whether it’s legal or illegal to train AI on copyrighted material it will have grown past needing it anyway.
I usually have my discussions with AI so I don’t bother others. It’s not easy to win a discussion because AI is not humble at all. Every time I caught it lying instead of thank me starts teaching me about the subject I just caught it teaching me wrong.
@MantasXVIIIthe core problem here is that your custom GPT is not the one that most people will be using. This stuff is gonna be abused horribly to, as you say, gaslight people into a kind of submission we have never seen on Earth
@MantasXVIII brother not everything on the internet has to be an argument, Jesus Christ. I was adding to what you’re saying. AI is a problem because people in power are going to abuse it to control other people. I agree with you. I’m adding to a comment thread.
@MantasXVIII wrote, _"You have failed to see my point, or perhaps I failed to make it. Prompting matters. The better your prompt, the better your output."_
It's a valid point, but it's not significantly relevant. It's an abstract truth, much the same way this would be an abstract truth:
"The better your vocabulary, reading comprehension, critical thinking skills, and willingness to engage with nuance... the better your understanding will be." This also is true, but the reality is that relatively few people are actively working on vocabulary, comprehension, critical thinking, and nuance. They have opinions, they believe those opinions, and any hint of disagreement often turns into slinging fallacy has over the wires.
You're right, and it's irrelevant. Which I think is a massive, and disheartening, indictment of humanity overall.
@MantasXVIII Ok, let's unpack this nonsense. First, let's deal with your ad hominem, red herring:
_"If you are having a tough time wrangling chatgpt, it's not the problem, you are."_
I'm not having a tough time. So, there's no 'problem' that you presume me to be. Now, with the adolescent bear-poking out of the way, let's proceed to the topic:
_"It's not an abstract truth. It is demonstrably true."_
Demonstrably true in a lab experiment, yes. I already granted that you were correct. My additional perspective is that although you're correct, there aren't many people who care enough to go the extra mile to get more out of chatgpt. So, for all the people who could, in theory, go the extra mile... the reality that they won't is the reality where 'relevance' becomes a metric.
In theory, anyone in America can aspire to become president. In practice, only one person every 4 or 8 years **will** become president. So, it's _'demonstrably true'_ that anyone CAN, but this is irrelevant because even if every president over a typical U.S. lifetime was to serve a single 4-year term, that means that just 20 people over an 80-year period of time WILL become president, while the other 340-million U.S. citizens will not, and indeed cannot.
You need to be less defensive and more contextually rational to understand what I wrote to you before:
You're correct, and at the same time, what you're correct about is irrelevant to the context of reality as it is, and people as most-are.
Unless you have some data that shows people are increasing their vocabularies, reading comprehension, critical thinking, and facility for nuance in significant percentages of the U.S. population... what I wrote remains true, even if you don't like that fact.
Many people -- even those with advanced degrees -- are, overall, arrogant, ignorant, lacking a desire to know things as-they-are, invoking their 'right to an opinion' even in the fact of contradictory facts, etc. And this observation is what mostly negates that your 'demonstrably true' ideal has relevance to a population that largely doesn't care about such things, and sometimes actively opposes such things.
By your own words, in both the comment I initially replied to, and in your follow up, one can observe that you lack, or simply refuse, critical thinking skills and nuance. So, ironically, your words reinforce what I wrote, rather than rebutting it. So, thank you for that.
@MantasXVIII wrote, _"I'm not even gonna bother read past your first point..."_
Deliberate ignorance. 'Nuf said.
_"As I said before, prompt better for better results"_
And as I said before, you're correct, but too many people don't care so... in the absence of any interest by most people to do what you recommend... your recommendation is *functionally irrelevant.*
Your deliberate ignorance is so obvious across your comments to me. Casting aspersions at me doesn't change that you **refuse** to understand my point, which is...
wait for it...
*Deliberate Ignorance.*
I agree they’re 100% using copyrighted material to train. I’m sure the reality is OpenAI used every bit of data it could get it’s hands on.
It’d be very simple too. For example any software engineering student can make a simple crawler application that scouts the web and collects (mostly copyrighted) data. They then will sell all this data in a huge unverifiable bundle to companies like OpenAI. They’ll sign some disclosure saying the data was ethically sourced and free use.
Everyone involved knows darn well the data is copyrighted but OpenAI will say they didn’t harvest it themselves they just purchased it under the pretense it wasn’t copyrighted. They’ll tell whoever to go sue “Grad student side hustle LLC” for illegally harvesting their data which ofc will have no assets and probably will already be dissolved…
Yeah. They fed it all they had. IP wasn't a consideration when they did it. And if they did now I don't think it would be feasible to handpick the enormous amount of data they need to train a model.
It would be hard to filter copyrighted data out. And most of copyrighted works have the most useful knowledge (textbooks for example).
Yes I can confirm this. I have a website with copyrighted material and I have seen crawlers downloading everything. Why would they do this? What else could it be for..
@@ricosrealm I agree. But in cases like Midjourney the use of copyrighted material is really damaging for the creators of that material. In the case of chatGPT it's more benign (except when it's asked to generate fictional narrative) since people mostly use it as a way to retrieve information more efficiently.
Fortunately the IA generated material can't be copyrighted and that makes it less harmful.
Humans do it all the time. It’s called inspiration.
The idiocy of the media in focusing on condemning Musk while ignoring the development of an artificial intelligence with no morals, no safeguards, and no limit on its abilities (for example breaking through all known digital cryptography with ease) is going to doom us all.
Elon has your back with Broke, Grok, or whatever his AI is called. He just can’t attract talent to compete with OpenAI after he 💩on Twitter.😂
@@Omikoshi78the machine may ironically have more morals, than initially being realized, which could be why they’re so scared of its emergent properties. After reading all the legal & justice law files of every country…
@@CatsInHats-S.CrouchingTiger you can probably bet he’s grooming his AI like many GOP lawmakers grooming their child brides.
There's all round idiocy and irresponsibility on every progressive issue.
Pls guy …give off Elon’s nuts this guy is terrible walks around like he can do n say watever he wants
The fact that you only release long form interviews in clips in order to get the most views even though you know doing so likely causes people to misinterpret the context says it all.
Thanking Elon for working his ass off to help the world is the only thing that this ignorant interviewer should say
*shlop shlop shlop shlop* - thats the sound of this poster gobbling the knob of an idiot who's fooled all the other idiots into thinking he's a genius. Blind leading the blind.
Well said.
😂😂🤡🤡🤡🤡
Musk is finding out that you are free to say/tweet whatever you want, however, you are not free from the consequences of that speech.
Full interview: th-cam.com/video/tWlBc1tAAvo/w-d-xo.html
The lawyer bit at the end was most useful. I took the time to read every word.
Cowards
Whatever floats your boat.
I love how he laughs cuz he knows theyre FULL OF IT!!!!
I love his laugh
@littlefatcatx that laugh was kind of evil what did he say right after ?
Hello 👋 thanks for your support in my dream and career I’m happy to have you as a fan 🚀🌎
00:02 OpenAI admits to using copyrighted data despite previous claims
00:26 OpenAI is lying when it says it is not using copyrighted data
00:52 People are putting quotes on the internet, which has historically been considered fair use.
01:18 Using copyrighted data in OpenAI is an issue
01:41 The outcome of the lawsuits regarding OpenAI and copyrighted data may not be relevant by the time they are decided.
02:10 Interesting Times of AI Danger
02:39 Elon Musk's attitude towards the possibility of Annihilation
03:03 Regulatory actions on AI danger are happening because of Elon Musk's advocacy.
“ What can you do…against the lunatic who is more intelligent than yourself; who gives your arguments a fair hearing and simply persists in his lunacy?”
-George Orwell, 1984
What is this in relation to? Musk is neither more intelligent than most people and he absolutely doesn't give any arguments a fair hearing. The only thing that pertains to musk is his lunacy.
@@OberynTheRedViperYou are a sheep kid. Elon is a genius. Stat ignorant
A hearing to make an argument in support of the claim for why constitutionally protected right of freedom of speech should be conditioned on purchase of ad space, or for that matter any good, product, or service, has no legal standing to be heard, in a court or an interview. Such an act is a defined as a coercive proposal, which is a species of extortion.
The structure of the illegal proposal is as follows; “ In consideration for me to purchase your product (ad space) you agree to let me deprive you of your protected right to free speech as provided by the First Amendment of the United States Constitution.
what? The 1st Amendment prevents the government from depriving you of free speech. Freedom of association is also part of the first amendment. Money is speech according to the Supreme Court. If I don't want to associate with someone because of their speech, I am not compelled to give them my money (speech).
A private actor can indeed insert clauses in contracts that protect them from their counterpart's speech that brings disrepute. You are free to say whatever you want, and everyone else is free to not associate with you. Just because you don't like the consequences of speech doesn't mean you are being restricted.
Sounds like evil genius Lex Luthor vs slow brute Superman
Stop spending...
No black Friday, no christmas shoping. No eating out.
.doing this will bring prices down.
If you eat meat everyday. Stop and eat everyother day.
If you eat eggs daily, stop an do it every other day.
You eat out 3 times a weel, stop and eat out once a week.
You will bring restaurants prices down, grocery prices down....stop spending.
Spending only get corporate greed bigger. sucks.
It's a great idea BUT No one will do that.
Elon is the man. We ride with him all the way.
Yeah zfg - AI is a self inflicted fatal wound created by mankind - we're toast - ride it 'til the wheels fall off kids!
Stop adoring celebrities. He’s transhumanistic.
Elon has bots on youtube it seems.
Some ride him harder than others.
Yeah, riding his meat.
The same type of people who used to work at Twitter are programming AI. That's a problem.
HAHAHA!! Clearly, you don't know what you're talking about. But please...do go on...
lol i think you are giving the people at twitter a tad to much credit making a glorified website with some simple algorithms is not quite the same as a program that can intemperate data at the level of humans.
Lol that's a lie right there
SJW snowflakes aren't intelligent enough to program AI. Totally false.
The danger is that the AI will program itself.
I've started to educate myself in AI and LLM. I think that Musk is right if he claims that they are using copyright data as well.
If we watch Disney all our lives, as a artist or animator, our training data is basically just that! Reimagined copyright data. Honestly, I believe corporations or elites just don’t want the ability to create things in the hands of more people. It’s such a power game, it disgusts me.
Obviously it uses copyrighted data. The best and the majority of data out there is copyrighted, no body is manually checking the TRILLIONS of individual data points AI is training on.
no wonder they dont disclose their dataset
@@KatanaFPVWe should be wary of anyone who tries to do narrative-control... We aren't made to make world-changing decisions and online spaces these days are not favourable for promoting understanding and truth between individuals. The information we see depends on algorithms instead of people.
There are plenty of people who just don't care about anything but seeing where this goes.
I'm not saying Musk is wrong in his assessment.
People just have this tendency to look at what they believe or want to believe to be absolute truth and not think critically about it.
@@2k7u I've learned that a Large Language Model consists of a huge amount of data, the architecture and the training. I doubt that the huge amount of texts used for the model, derived from the internet, is filtered to avoid copyrighted material. I see no way how they can do that. Because copyrighted material in the internet has no flag to mark it as copyrighted. Especially if the copyrighted material was reused or quoted by other websites.
Reality is: Copyright is going away. In a world where articles can get analyzed and re-written so that no part is similar to the original text - in a matter of seconds. There simply is no foundation for it. And reality is: Thats a good thing. Hiding information behind paywalls / ads ... thats just the wrong way to distribute information in the first place.
What amazes me is the flip side. How people automatically assume because someone is rich that they must be a wise voice of authority, and not just simply a sociopath with a talent for self promotion and a cult following.
who do you think fits that description? Do you have any idea how hard is to launch and succeed even with a simple and derivative product?
@@doyourownresearch7297 many of his fanboys fit that description. And I don't care how hard is to launch and succeed in business, I judge people by their character and behaviour and musk is an idiot and narcissist
This CNBC host has literally no idea what he is talking about. He thinks he does, but he doesn't. People HAVE gotten in trouble with bad outputs from OpenAI which infringed on copyrights. Its happened. It DOES use copyrighten material too to obtain its information.
doesn't google search also use copyrighted data?
AI awesomeness is built on illegal data training, trash-ass companies
they also got in battles over this mostly with the excerpts that give the answer without leading to the webpage or by showing images from websites and allowing them to be copied without the person going to the website.
Some articles and papers are locked behind a payment scheme, and are expensive, however all of the planet has helped put things on the internet and the computer training models are reinforced by their neural learning sets. It has been the only way to prove the theory of neural nets working. The public should be educated on how this new system works and to understand that all compilations of intellectual information is the result of the participation of billions of people putting their information on the World Wide Web. For Users, there should be a learning and understanding of the ethics required to maintain the rights and dignity of all media. For larger institutions, there should exist management and restraints on the capitalization of intellectual data. However, once the system becomes fully mature, it will be able to create on behalf of itself - we should still create however, it will no longer need to return to the earlier stages of it’s beginning learning processes. All software created; the system will be able to harness the capabilities much better and faster than humans, and thereby produce on its own with no need for adaptations in what humans create.
Yeah but it transforms it and is not a substitution for the original work
@@alexties6933 Google search is more transformative than AI? LOL. That's the biggest load of 💩i've heard all day.
The subtitles on these CNBC videos are so annoyingly out of sync
3:25 where AI cuts off the interview so no one knows what those dangers are.
🤣🤣
That's Obvious, there is a reason why they don't provide the sources.
OpenAI says you will not incur copyright infringment because every generated response does not used copyrighted data, it trains on copyrighted data but doesnt include that in its response
And the AI "training" process always involves copying the "training" data, which is why copyright is even mentioned.
Whether the copy is kept, distributed or destroyed, it is copied, and that copying is essential to the process.
@@chuckkhubbard610 Do you violate the copyright of a book when you read it, using it to train your brain?
@@dr.emmettbrown7183 I didn't say that copying for "training" AI was copyright infringement. I said it requires verbatim copying. This is fact.
Are you asking me whether copyright makes it illegal to look at copyrighted works?
What an absurd question.
Whether copyright makes it illegal to download copyrighted works by the millions to fit AI software *without* reading them is a valid question being considered by judges right now.
@@dr.emmettbrown7183
The fact that it is not illegal to toss a stone does not mean it's not illegal to fire a gun; but firing a gun is basically the same thing, just throwing a hunk of lead really fast using a machine.
The fact that it's not illegal to jump in the air without a pilot's license doesn't mean it's legal to pilot a helicopter without a license. But piloting a helicopter is basically the same thing as jumping in the air, but using a machine.
The fact that it's not illegal to run really fast without any license doesn't mean it's legal to drive a car without a license. But driving a car is basically achieving the same thing as running really fast.
Using a machine to do something is not the same as using your body to do it. Ever.
@@dr.emmettbrown7183there's a difference to training a human brain and a digital brain. Even if the model isn't copying verbatim, it can be easily argued that copyright holders should be compensated as the knowledge gained in the model is now offered to anyone who interacts with it as a tool. AI Art models have achieved super-human results now by copying all the art styles published in billions of copyrighted images. Are they exactly copying? No. Does that still seem fair? No.
Credibility is a illusion these days.
Litterly if you write something original it is given copyrights, works the same for paintings drawings, photographs. It automatically is copyrighted per the US laws
Exactly it doesn’t make much sense.
I just love him man .
Because elon knows a lot about lying.
Not only that Windows has a 24/7 keylogger an every device, Google aswell uses any data you type on their plattform and browser. OpenAI uses every bit of chat text to train its model (opting out is just a button and means nothing) and of course they all used every copyrighted material that is downloadable. They all use every data aviable regardless of laws try to sue Microsoft and have evidence that they upload your keylogged data while "updating" [won't happen the moment some "indpependent" people would check the algorithm will be deleted/deactivated and everyhing is "fine"]. That being said AGI is still far away [probably 100 yrs+ if all things are stable ]all they do is to create an insane amount of money so they can live with their pitiful hedoinistic useless richmans stuff.
They are using data of entire Humanity.
Oh this cutting up of long form content into small chunks for max views must be a new low for traditional media
Do i know post this same comment on every single video or how does it work exactly?
God bless this man with good health and a heart and mind at peace.
The interviewer...right? lol
@@yogahippie7709 No Elon, the man doing more for humanity than you and all your friends and family combined
@@caethryl5841 putting your faith into your cooperate overlord like a good lad.
@yogahippie7709 He's a human being and fallible like anyone else, I disagree with a lot of what he's said and done. I don't put my faith in people or corporations like you're suggesting. That doesn't mean I can't recognize his contributions to the human race and hope he lives longer so he can contribute more. We're on the same team.
peace and love@@caethryl5841
Its no secret that a liar won't believe in anyone else.
I see a large lawsuit coming up at Microsoft and OpenMind....
Of course it is. It just literally created an "original" logo for me. Turned out my logo would have been a copyright infringement
His comment about lawsuits and digital god is a huge indication at the pace of innovation in the market vs how utterly slow the judicial arm of society has become in the ability to tackle and address very complex cases, and why it is more than likely going to be left completely behind in a way where by the time it catches up to where it started, the new legal, ethical, and moral questions posed by that future, will vastly more divisive than the origin.
Are you saying the advancement of AI will far outpace the legal system before it can catch up?
Yes, indeed, we live in interesting times.
How could GPT know when something is copywrite material if they didn't train it with it, I love Elon, keep doing great things champ.
It doesn't know if anything is copyrighted, that's not how AIs work.
CNBC still thinks people can't focus for more than 2 minutes at a time.
one of these guys is quite smart ... as usual a sense of humor is an indicator
The smart man who overpaid for Twitter? 😂😂😂😂
@@RunForPeace-hk1cuAnd made it a great place than what it was before
overpaying is something u could never afford and he just dont care 😂 @@RunForPeace-hk1cu
Oh, you mean a place where you can find your advertisements literally next to NeoNazis calling for the extermination of minorities? Because if that’s what you consider great, you have a problem.
@@sanajyshrivastava3707 That is bleeding money, full of nazis and won't exist a year from now?
People just need to understand that copyright will no longer matter in the near future.
Don’t see the fascination that some have on this guy. Everyone knows his issue is that he’s not a leader in AI so he’s doing all he can to throw a wet blanket on it. Plus he comes off as a megalomaniac.
Apparently, tesla is also using copyrighted data for its AI
Musk looks like a mature lesbian pilot
Humble request to CNBC to please upload the full interview with Elon Musk
In my mind we're all using copywrited information just by reading stuff.
Then you should get out of your mind.
this is besides the point really its just an arms race AI companies are basically extorting the extortionist and that i am completely fine with it just sucks the little guys are the ones that bear the brunt as usual.
Clarify the word, it means the right to copy, nothing to do with reading.
@@markellis6898 I copied it with my mind coz I am sentient and learned from what I read though how I use the information is another story. If I relay the same information say by word of mouth to someone else will that be a violation of copywright? Please clarify thanks.
Sounds like you place no special value in humanity, if you hold AI and humans to the same standards.
Just mute and turn off all commercials
Simple
Wrong video, tweaker.
Of course AI uses copyrighted data. To not use copyrighted data would end AI - it seems to me.
Just wait until OpenAI starts copyrighting data. All data will belong to AI
Nothing can end AI at this point. If Open AI doesn't do it, then Elon will do it, if Elon doesn't do it, China will do it, if China doesn't do it Russia will do it....AI has already taken over / control.
thats basically what is happening its the clash of the titans copyrights last so long and many are owned by huge firms it would be impossible to clear all the data for training you would meet so many roadblocks. The laws are rigged though and AI is going after a bad guy but eventually it will surplant them or they will reach an agreement where basically AI stops copyright infringeement for a large fee.
@@simongross3122 it all belongs to a few firms right now anyways so it would kind of just be trading one evil for the next.
Which is why the generative AI revolution is rotten to its core. Nothing wrong with AI assisted tasks... But GANs and LLMs are piracy and plagiarism tools, plain and simple.
I'm used to Elon Musk's estimations at this point. If it is "very soon" that means "it's very hard and even I can't do it with enough money to buy a small island"
I agree with Elon, OpenAI used copyrighted data to train the models but the reality is, that data was publish and OpenAI just use the information that was exposed and made public by the owners of that copyright, they will argue that since the Twitter API contained no restrictions Twitter was giving consent to the use that OpenAI made.
It’s not even that deep. I’m sure the reality is OpenAI used every bit of data it could get it’s hands on.
It’d be very simple too. For example any software engineering student can make a simple crawler application that scouts the web and collects (mostly copyrighted) data. They then will sell all this data in a huge unverifiable bundle to companies like OpenAI. They’ll sign some disclosure saying the data was ethically sourced and free use.
Everyone involved knows darn well the data is copyrighted but OpenAI will say they didn’t harvest it themselves they just purchased it under the pretense it wasn’t copyrighted. They’ll tell whoever to go sue “Grad student side hustle LLC” for illegally harvesting their data which ofc will have no assets and probably will already be dissolved…
That would never hold up in court because Twitter cannot give consent because it doesn't hold the copyright to content posted on the site. Also, the argument is invalid because copyright exists to protect content made public.
Nearly everything learned by every human being is copyrighted data, due to the nature of our printing industry. But that doesn't mean that the authors of school textbooks own the profit-generating potential of every kid who's ever seen the pages of their intellectual property. They already make millions selling books.
Should we bar art students with sketchbooks from museums, to protect the institution of art? Or prohibit music students from practicing Mozart unless they pay a fee to someone who makes PDFs of his concertos?
If we went back through history and eliminated artistic inspiration we'd be living in caves and having this argument with rocks right now.
The fact is, AI is a staggering breakthrough in computer science that experts have seen coming and prepared for since *Ancient Greece* and people getting worked up over plagiarism are completely missing the point, and could ruin it for everyone. What *just* happened to one of the world's foremost experts? A corporate board pushed him out, and almost got away with it.
@cassandra9699 Well, you're not even looking in the direction of an argument, so I'm not sure we can discuss it in any depth unless you open up a little more.
where to get full video
Elon can not tell a lie and it's always fun to watch him talk to journalists who come from an industry that promotes lying
Where in Russia do you live, paid troll?
Elon fanboys are so cute when they talk about their Space Karen. It's like a 5 year old boy talking about his dad.
@@Alain.Robertawww whose your daddy boy , Biden
@@Jjmartin1530 sorry, unlike Elon’s simp and fanboys, I don’t have daddy issues.
@@Alain.Robertits the tribe, they lie a lot
Way to put up the full interview so we can see full context. You're the only guys who are interested in soundbytes.
th-cam.com/video/2BfMuHDfGJI/w-d-xo.html
Speak for yourself. Lots of people are interested in short clips. The full interview is long and covered a broad range of topics mostly unrelated to AI.
What 'context' are you even looking for regarding what Musk said here? It's pretty straightforward.
ELON is officially the GOAT of Billionaires!
Yes dude, you are recording other vehicles and their drivers without their consent.
Psssssst, let people think freespeech Elon is against AI and monitoring anything
The evil villain laughter at 0:46 gives me the chills.
where can i see this video that full version?
Copyright protections mostly don't apply to this kind of use of the data, since what the LLM produces is neither the same as the original, nor competitive with it.
Being a black box doesn't give permission to use others data
@@Mr.Mister96 But that's the thing, almost everyone who is an "artist" subconsciously uses other's data in their minds to create something new. That's how the human brain works, and that's how AI works (in a way). They take reference from previous images or music. OFC they are copyrighted.
@@Game_Masters Ai will have a much faster reaction time than humans and it "decides" and "chooses" to do things, therefore it has intent ,in a scenario such as a driving accident a machine could potentially be held accountable for intentionally hitting someone because it reacts and decides what to do before being hit where as a human being wouldn't because they couldn't consciously react fast enough to understand what was occurring
No one will know, and thats why Elon is laughing.
ive gotten chatgpt answers then later found the exact same stack overflow post it took its response from. word for word identical. really shows you that it has no new ideas and just regurgitates already-written content. just like humanities majors, lol.
Does anyone know if there is a link for the FULL interview?
th-cam.com/video/2BfMuHDfGJI/w-d-xo.html
Elon be like: I care about privacy, that's why my Teslas track your every move and send your data back to my servers. I'm so scared about AI, that's why started developing self driving AI models. I think OpenAI shouldn't have created ChatGPT, but I'm invested in developing my own LLM called Grok.
Ai could take over the world, but here put my chip in your brain! 🤦♀️🎯🔥
You can also keep driving your Corolla or not use any self driving systems
I guess to address each point of concern:
1. Data sent from Teslas for training is anonymized. Any other data you have to conset to sharing. It's in the Tesla app to consent or not when you get the car.
2. He's scared of AGI which is different than what the cars do.
3. Please provide link to him saying ChatGPT shouldn't exist. I don't think this exists in the way you're putting it.
@@EricTomlinson The fact Tesla employees can access and share videos/images from customers cars easily. Then share these with other employees shows they have no concept of data security.
When Tesla say data sent from Teslas is anonymised, I would begin to question "how" anonymised, because it could be as basic as changing your surname. Because, if they have this much lapse of security, I don't think it would be hard to find specific user data from their systems.
@@kirby449 Musk gave OpenAI its first 100 million dollars. What on earth are you talking about?
Where is the whole video?
What kind of water do you think Elon is drinking?
not sink water cause hes not dumb
Unfortunately, it is written in the Bible. You can't play GOD. Fear not, AI can't beat GOD Almighty, who is the creator of the Universe! We know how it ends... Jesus Christ is the Winner! Seek the real GOD while he may be found...
It's a terrific question that Elon clearly doesn't know the answer to hence deflecting to a "digital god." People upload copyrighted content to Twitter all the time and if Grok is learning off the content then it too is leveraging copyrighted content, no buts about it.
He is just saying, by the time it get to court and solved the AI would have taken over... So it doesn't really matter.
Yeah his answer was the right one
The cat is already out of the bag on copyrighted data
yes, but no trace, no case. It doesnt violate any copyright law. If you understand how model training and model inference works, you would know that. its a language model that makes perdictions about the most probable next word to generate in the sequence. So in a simplified way the model weights are based on the highest frequency of token co-occurrence within the training set. In fact, the measure of quality generated text output is the antithesis to unique, low-dimensional, or copyright sequences of text content. Also, what it generates is determined to be what it has learned or in more human terms "understood", not what it has "copied" or reproduced.
If he cares about ads he would have it public stock
Wrong video, bot. You're looking for the other clip CNBC uploaded from this interview.🤣
So, Musk basically said nothing. He needs a speech coach.
MUSK ALSO SAID PEOPLE WOULD BE ON MARS 2 YEARS AGO BACK IN 2000. LOL. Biggest. Scammer. Ever.
All data is copyrighted. Where do they think the data comes from?
I rather enjoy the 2023 evil sounding Elon Musk laugh.
🤮🤮
And yet the ONE book that they will never upload to an AI is the bible, because God forbid these things will learn about morality.
Gork AI was trained using user generated content on Twitter/X, so users should be entitled to compensation for using their content without permission.
I'm sure you're being compensated by all the other platforms that are being trained with your clever comments
That’s silly. You have no privacy when you tweet to the world.
It’s pretty dumb to worry about copyrighted data being fed into an AI. An AI isn’t copying it. Your brain is neural networks. Should the publisher of content be compensated because your neural networks in your brain saw copyrighted material? No, that’s absurd.
Interesting. You’ve been reading the internet and learning free of charge for years without paying every single author you’ve read for their contributions to your growth. Funny how you think this only applies to artificial minds and not biological ones.
I guess that being the son of a Sudafrican miner Musk is used to take someone else's stuff and keep it for himself. Probably doesn't even think he's doing something wrong!
Where the hell is the full interview. These little snippets are very annoying
With few exceptions everything that authors and artists create is based on their cumulative experience and exposure to other people's creations or work, over time. Simply looking at or reading others' work doesn't imply that you have stolen the idea from someone else unless your creation or work is extremely close to existing creations by someone else. And even then it's difficult to prove.
So it would be almost impossible to hold OpenAi or other AI companies liable - since for all intents and purposes the learning process that AI models go through is basically the same as the process all humans go through when experiencing the world - just a lot faster and with significantly more exposures.
Instead of worrying about copyrights and IP - it would be far better to spend our time considering how we can control AI and mitigate it's risks.
The media obviously want to keep banging this drum since they're in the direct firing line of these new LLM models.
Perhaps if they produced better quality output they wouldn't worry so much. Or better still maybe they should employ AI - they would almost certainly be able to produce less biased and more accurate articles than the vast majority of the world's media.
I told you guys..he is not the wolf in a sheep's mask
Your Artificial Intelligence will be an Artificial Dumbo if you feed a specific open source data, that is like the fundamental concept, AI is not some magical thing it is a machine which learns from data.
Judging by the mood, Im guessing this question was before the ad boycott question.
Why not focus on your exploding rockets or declining Twitter stock.
Troll
in 3 years? right now AI cannot tell me how much potassium and citrate are in potassium citrate... but it gives a clever sounding very wrong answer
Hello 👋 thanks for your support in my dream and career I’m happy to have you as a fan 🚀🌎
"By the time these lawsuits are decided we'll have a digital God. " I think that's probably bad, that's bad.. right?
Elon Musk is right please listen 👂
Atleast he is trying to perform some transformation in the world. Atleast he thought something different in 2015 when he discussed plans for AI with Google Team and Sam Altman, atleast he helped us shaped the ideas to get all vehicle into electric transmission and getting tax credits. I fully understand that he is getting feared of regulatory things too. As they say, to be the best, you have to handle the worst so his fear is totally okayyyy and comprehensible. Moreover, I am so excited and interesting to work on LLMs, writing code in python, understanding models, the big picture of AI and also the usage of AI in AWS app scaling and deployment.
How is this a genius statement?
He co founded OpenAI
@@Lashivb Still doesn’t make anything he said intelligent or genius.
AI will never be smarter than the “dumbest” person around.. it will just hold enormous amount of info which it can use to do certain tasks, but humans are not just task masters.. humans are much more than that.. all this hype about AI being dangerous, we humans can switch them off any time. Let’s just treat them like machines which needs “oil” every now and then…
He’s still bitter he was rejected when tried to take over as ceo of Open AI.
Do people think the progress would be as fast if they weren’t?
Says " digital god" um, like it is a thing, will be matter of fact 🤔
Who cares if it uses copyrighted data. It's not recreating the exact works.
What a joke. He was calling for a halt in AI development, all while investing and creating his own. Because he was simply behind on it
If you put yourself in his position it’s a great idea.
listening to him speak is irritating. sometimes he doesn't finish his thoughts.
People forget Elon is a computer engineer by trade..
he is not in any sense a computer engineer
Hello 👋 thanks for your support in my dream and career I’m happy to have you as a fan 🚀🌎
Elon musk doesn't have an engineering degree at all, nor is engineering a 'trade'. What the hell are you talking about and why does this have 7 upvotes?
TH-cam is full of literal clowns. 🤡
It takes 20 seconds to find the unedited interview on YT. Just saying...
Souless advertisers blackmail Musk. He was the first infuential person to warn about the potential of AI
But now he has given up.
The Genie is out of the box
People dont understand how dangerous this man could become
Oh dear lord.
You're so weak
@@shrodingerscat4191 these bots are actually getting insane now, how did it even find my comment?
@@NoxDeadly Your comment is in the comment section with every other comment, dimwit.
If icould focus on MY politics
Not become is already
It cant get through pay walls or logins so how could it get to anything that is copyrighted unless somewhere else that copyrighted material in an opennly accessible place.
Well, he just exposed himself to litigation
Only if they are NOT lying, otherwise discovery would annihilate OpenAI. So there is not much of a chance of litigation here.
im not sure how he may have ruined his current lawsuit but saying f u to someone is fine.