BREAKING: OpenAI Reveals the TRUTH About Elon Musk's Lawsuit 🔥
Vložit
- čas přidán 4. 03. 2024
- Join My Newsletter for Regular AI Updates 👇🏼
www.matthewberman.com
Need AI Consulting? ✅
forwardfuture.ai/
My Links 🔗
👉🏻 Subscribe: / @matthew_berman
👉🏻 Twitter: / matthewberman
👉🏻 Discord: / discord
👉🏻 Patreon: / matthewberman
Rent a GPU (MassedCompute) 🚀
bit.ly/matthew-berman-youtube
USE CODE "MatthewBerman" for 50% discount
Media/Sponsorship Inquiries 📈
bit.ly/44TC45V
Links:
Previous Video About Lawsuit - • Elon Musk files BOMBSH...
Letter - openai.com/blog/openai-elon-musk
Scott Alexander Bio - rationalwiki.org/wiki/Scott_A...
Scott Alexander Blog - slatestarcodex.com/2015/12/17/... - Věda a technologie
WIth this new information, do you think Elon or OpenAI is right?
i doubt we'll actually know for atleast a couple years sadly
I'm thinking OpenAi is mostly right, but they're still not being 100% transparent, but Elon is and always has talked a big game and floated many conspiracy theories. Neither is essentially on our side tbh.
@@Cross-CutFilms Hot take, I think the correct take.
Elon Musk is a jerk but Elon Musk is also right. For crying out loud it's a joke the OpenAI isn't open. Sam Altman is trying to convert this into profit to get rich. Sam is scamming like that AI chip company. He always up to something shady.
I think fighting over money is just what we need for alignment.
You misinterpreted the letter, TESLA is the cash cow that would have fed OpenAI with its cash.
duh?1 Occams razor
Yes, 👍
Damn, I just typed this up. You are correct.
Yea it took me a second to hear that lol
I jumped out of my seat when I heard that....he did say he is not Elon fanboy, it shows.
There’s no way this story isn’t in a movie one day
I hope somebody make the movie in Ai
Directed by Sora, written by GPT5
@@spyral00 😂😂 can't wait
@@spyral00with reviews generated by Mixtral, Llama and Gemini
Yes, and when it’s too good to be true.. It’s likely already orchestrated as this to draw public’s attention away from something else. No way these guys haven’t agreed on the outcome already 😂😊
they definitely used gpt 5 to write this
Haha. Elon probably used his AI to write the lawsuit. AI suing, AI.
😂😂😂😂
😂😂
my thoughts exactly XD
Is GPT5 vs GPT5 in court! 😆
Musk main points were Microsoft outsized influence/ GPT4 design, they totally avoided the questions. BShitters
Exactly. Their response was just corporate misdirectional tactics. And the fact that they are saying that the public will benefit from this technology but they are not willing to be transparent is even more disturbing. We are talking about AGI tools developed by one of the largest companies in the world and this just goes to show their lack of any commitment to the public. They are willing to disrupt entire economies by replacing jobs but only provide rent services back. This is literal economic insanity that threatens to lead to authoritarian regimes.
I think you have misunderstood what is being said at 7:26. You interpret the "its" in "attach to Tesla as its cash cow" to mean "Tesla's," but I think "Its" means "OpenAI's." OpenAI needed lots of money. The letter says there was discussion of how to get lots of money. Elon suggests that OpenAI should "attach to Tesla as OpenAI's cash cow." That is, Tesla would be the business that would provide a steady income to Open AI. I don't know why you think the "its" means the authors of the letter are claiming Elon suggested to OpenAI--which needed a lot of money--that the way to do that was for OpenAI to "attach to Tesla as Tesla's cash cow." That interpretation doesn't fit the context.
Yeah I think that was a huge oversight. Berman did an oopsie.
Yep. The email says Open AI could attach to Tesla rather than attaching to Apple or Amazon (who do not share their vision), then increase Tesla's revenue through car/truck sales, which could then be invested back into OpenAI. So, OpenAI could create value for Tesla and then grow off of Tesla's investment back into OpenAI.
Happens when you rush to be first.
Agreed
Yes, I agree
I was waiting for the part where they addressed Microsoft, but that part didn't come...
they didn't talk about the agi part ether
@@NeostormXLMAX because there is nothing to talk about, Musk wants the judge to recognize GPT4 as AGI, which is just non realistic
@@slick3996 it wouldn't be a judge tho, musk specifically demanded for jury trial. And a jury full of laypeople might just come to the conclusion that musk wants them to....
They won't address Microsoft because they are scared of them. These companies are too big by themselves. Evil monopoly at best.
@mambaASI Hopefully a jury does. The entire point was for their creations to be in the public, non-profit sphere....not turn "open-ai into closed-ai for profit.
I think the 'its' in "its cash cow' refers to OpenAI. So Tesla would be the cash cow for OpenAI.
100%
Exactly. Tesla didn't need the money back then. OpenAI did!
Agreed. Even if Elon thought OpenAI would someday be a 'cash cow' for him, he clearly felt the same about Tesla as his $50B+ comp package was based on extraordinary future Tesla performance - which came to pass.
Absolutely. The quotes are short, simple and clear. Berman's misreading is so grossly unjustified that I don't think it's a MISTAKE, I think it's BIAS pure and simple.
If the quotes at issue were long-winded, unclear and meandering, I could understand an upside-down conclusion based on misunderstanding of badly written text.
That's not the case here. If this is NOT gross bias, please tell me what the third possibility is?
It's interesting that they don't really dispute having reached AGI internally here 😅
I think the mystery plays well for them, and they know it
Them saying anything about their current state in the AGI front would be too risky at this point. Too many moving parts. Shareholders are the biggest concern in this regard I assume, but there's also the fragility of the public opinion, especially after the firing of Sam Altman. Elon is really, really good at gaining widespread public support on any matter he puts his mind to, but he's got a horrible track record of keeping promises he's making. I think the board of OpenAI sees a lesson here and they prefer keeping everything about AGI strictly to themselves until they're sure that they have something they can finally deliver to public. So, yeah, it makes sense to exclude that dispute.
So if the best argument from openai is the fact that Elon said "yup", then it looks bad for them
Well "yup" means a lot in court.
Especially if ooenAI lawyers are solid.
@@divineigbinoba4506 yup means a lot when it's at the end of a clear one sided question, like: should we make this for profit? yup!
not after a four paragraph concert about how there are dangers and did you see this blog, do you know the dangers of open sourcing, and it would make sense for being less open (which does not exactly translate to "for profit") as it goes and so on and so forth
@@kooshanjazayeri Na I meant open source not for profit.
There's absolutely no argument about for profit.
OpenAI wouldn't have existed without for profit not even chatGPT or any Llm.
Only thing is they joined Microsoft over Tesla.
@@divineigbinoba4506 well still the paragraphs leading to the yup is vague, but being all out open source is not logical, but still there are lot of options between fully open source and not at all open, (just like the profit part)
@@kooshanjazayeri Well they open sourced whisper and GPT 2.
Yeah the statements leading up to the "yup" is kinda vague but not vague for a good lawyer...
If anything the "yup" is allot more vague than those statements.
The larger argument here is whether AGI should be open source or not. I'm actually surprised that the majority of people aren't defaulting to "yes" and that there is such a big debate on this.
Yes, open sourcing AGI possibly puts it in the hands of bad actors, but it also puts it in the hands of good ones.
Are you guys really keen to have "The tool that can build itself", i.e possibly the last tool humans will ever need to create, be exclusively in the hands of a single company... what could go wrong, right?
Decentralization is the answer.
Self aware machines are science fiction.
What could go wrong? => As it is a less hard problem to build unsafe agi than safe agi a lot of things could go wrong. What could go wrong if you build an unsafe nuclear power plant, even if its democraticed to everyone. If one nuclear plant blows up its a problem for everyone. I think you are right, that decentralization is the way to go. But the very first AGI/ASI should not be an unsafe one.
This is the most meaningful and the best comment between all of the comments on this video ....I hope every one would read it and think in it very carefully 👍👍👍👍👍👍👍👍👍
Have we forgotten Claude, Mistral, Gemini, Conjecture, et al? There is no chance that only ONE company will have AI engines. Not to speak of Chinese companies, and others around the world. Several older models have already been open sourced as have many Sci papers, no more are needed. Now tools will be available to those who have money - that's the 'capitalist way' - for as long as money exists.
There is no larger argument here. None of the parties are truly interested in giving their work out for free.
"Do you agree with OpenAI or Elon Musk?". Did it occur to you that might distrust both?
I think you are asking the wrong question. The question should be - should Microsoft be allowed to steal OpenAI?
How did MS steal them specifically?
Nah, that’s not the right question 😂
Ooooo very good question! The old pay attention to this hand and don’t worry about the other,No doubt!
Already done
Exactly and it's tragic. They can't shake loose of Microsoft if they wanted to now.
We all know what Microsoft does, eat up other companies and spit them out.
It still feels very crappy that they "clarify" the name to mean "open" in a way that is not open at all, in a way that does not align with what even their early investors considered. In a way that is literally not open at all.
"The regular operation of a business" needs no distinction. This is a gross retcon of the only reason most people invested.
I think you should at least acknowledge that although the idea to have open source code is altruistic ... as powerful as AI can be it would be naive to think there are no bad actors in this world who are "socially immoral" who have low EQ and couldn't care less if his neighbor suffers would use this technology to his own benefit? It's like communism ... sounds great if everyone is equal and treated equally but in practice ... it just won't work in the real world. How do you provide a system where nothing is created equal but yet provide equal opportunity for everyone ... it's too hard for some mind to process that , maybe too young a mind, I get it.
Yes... how do you even make an AI that is not "open" under that definition? even the most exclusive services with high cost / requirements that only few people can get are "shared" by the profider when you pay / meet their conditions, so.. it would also be "open", right?
They were never open, nor intended to be.
Well to open everything is also naive, Elon Musk also realized it's danger. Well imagine the nuclear bomb creator just leave bombs un-guarded and anyone who finds one , keeps one and that bomb now belong to the finder. AI is a powerful tool , making it fully open is asking for international security crisis when the wrong person get their hands on it.
@hawkenfox No, it's more akin to someone spreading the blueprints for a nuclear bomb all over the place.
The person in question would still need the immense amount of money, time, staff, knowledge, and other resources required to run it.
And even then, that doesn't mean they'd use it for nefarious purposes.
Elon put out more information in his law suite. Seems like we are missing a lot of emails.
Open AI did not deny having AGI. They are just doing misdirection by questioning Elons character instead of answering important questions like do they have AGI?
They also didn't bring up the founding charter anywhere in their response
@@kazedcat They're making a great argument by pointing out Elon first told them they needed money, and then he left to start his own for-profit AI.
@@FlowingLifeAlchemist Does not answer the important question do they have AGI?
@@kazedcat"Did they make a Terminator at Tesla? "
That's how silly your question is.
Two points I want to make:
1. I'm willing to bet that whatever OpenAI has built internally was used to fish through all the old emails and find the one's most supportive of their argument. In other words, I think they used AI to craft this post and that's fascinating to me for some reason.
2. The lawsuit from Elon feels more like a public relations move than an actual legally fruitful action. The wording is for public consumption and crafted to maximize quotability, not to maximize the odds that it succeeds on the merits of the law.
I hope Sam and Elon find a way to work together because they have more in common than they disagree with.
Elon and Sam finding a way to work with Demis and Ilya would give me hope for an aligned future of humanity. At the moment, I'd put it at 0%. Our only hope being that we survive the phase where psychopathic leaders 'control' AGI.
Too late now. OpenAI made a deal with the devil.. Microsoft.
With M$ controlling the board at openAI alignment with anything that is not for profit of M$ is impossible. We are doomed if the worst and more powerful company in history is the one that controls AGI.
This being public is much more of a PR move than the lawsuit itself, it's pretty unusual for a company to talk about details of a lawsuit, especially as a defending party. Which makes it even more likely that this was GPT generated, as a desperate move, and because they seem to be worshipping GPT-intelligence they were obviously so impressed they thought this might be able to influence the case...?
@@BrianMosleyUK having him as author probably just means that his email was used
At the end of the day Sam is not fighting for his GPT models to be freely available to humanity.
Very few things are truly free. Some cost $20 a month but most, even guns and religion, cost much more. Meanwhile, there are plenty of open source models free to anyone with their own supercomputer.
@@davidbangsdemocracy5455 Beside the point, the OpenAI founding agreement intended for the models to be open sourced.
@@thedofflin was it a legally written down (legal document) founding agreement.
Or just a conversation agreement?
@@divineigbinoba4506doesn't have to be anything related to legal agreements. Its a company that calls itself OPEN Ai. You can argue the way they are now is false advertising outside of the limited api and the free version of chatgpt.
Machine learning models already existed at Anthropic and Google, and OpenAI was the first to make them generally available for free to the entire world.
The backstory behind the whole argument with Helen Toner (openAI board member) was that she was accusing Sam Altman of being an accelerationist. Don't you remember that she tried to sell OpenAI to Anthropic after the board fired Sam back in December? It was to put the brakes on everything OpenAI was working on.
All along the way Sam and the accelerationists at OpenAI fought to get these models out to the public while everyone else was worried about safety first. And now the accelerationists who pushed for the world to learn about generative AI are being accused of trying to hide it.
It's such an inverted perception of what actually happened.
I'm running a company, a small sized company, with just over 80 employees.
And I gave a fair amount of "yup" responses.
That yup is "I kind of read what you said, but we'll discuss the actual steps during the process, because I don't have time to breakdown each paragraph into a separate discussion that will end up going back and forth".
7:45 i read this the complete other way.
That OpenAI should attach to Tesla as "its" (OpenAI's) cash cow.
Meaning since OpenAI struggled with funding, it should attach to Tesla to be able to siphon funds FROM Tesla as needed.
Whenever you have subject verb subject then "its", you often refer to the first subject.
"a car need to attach engines as its power source"
How can you read this as "its" refering to the engine? It makes no grammatical sense.
Elon meant Tesla would funnel funds into OpenAI, not that OpenAI would be the cash cow.
Firstly, that particular email was not from elon it was from an unknown (redacted) person. Secondly that is absolutely not a firm grammatical rule lol context is everything. And in the context of their full email, the writer clearly indicates that at present (in 2018) tesla was not a sufficient source of funds for openAI to be able to compete with google's AI investments. That's why this unknown person suggested that openAI become a cash cow for tesla, leveraging openAI tech to boost tesla products and services, thus massively boosting tesla market cap so that within a decade tesla would be able to compete with google in terms of AI investment dollars.
In context, I read "attach to Tesla as its cash cow" as Tesla being the cash cow. Why would the 'cash cow' merge into another company rather than the other way around. This is further enforced by his comment that OpenAI needs to be well funded.
I read it that OpenAI would cash-cow Tesla by building tech that could backend an actual self driving car, then billions of profit from selling car, fleets would flow back to cash-cow Big Compute and propel Open AI to AGI
Tesla would be the cash cow. This is at a time where Open AI was years and years away from making any money and Tesla already had billions in marketcap.
that's what i thought. openai would tether itself to the cash cow that is tesla because they had the resources to match google
@@EvolutionWendy No, clearly Elon was clearly suggesting Tesla's vast resources could fund OpenAI. Tesla is the cash cow.
That's not what a cash cow is. A cash cow is a business that generates money for some investing entity. In this hypothetical context, Tesla would invest in OpenAI so OpenAI can help boost Tesla products and services (mostly via rapidly improving self-driving) and subsequently boost tesla market cap. Therefore OpenAI is a cash cow for Tesla which is the investor. Furthermore, as the letter states, tesla (in 2018) was not sufficient to fund openAI in a meaningful way to compete with google's massive AI investments. That is why this unknown person suggests that openAI become tesla's cash cow, to help boost tesla market cap so that within a decade tesla is able to sufficiently compete with google's investment capability.
ultimately, they chose Microsoft as their cash cow. A Company that never makes any statement for Open Source.
Turns out openAI is open when you show up with enough cash. They bought all the models and prototypes, all the science, and even the science of models still in development.
have you ever heard about TypeScript or VS Code?
Coupled with Microsoft recently receiving a heft grant from DARPA, to develop single user QUANTUM Computers. I think the writing is on the wall, quite frankly. 🥴
thats a lot of buzzwords you learned there buddy@@nexys1225
Microsoft now is quite different from the Microsoft we all learned to use and hate in the 90s. Now they support a lot of open source.
I am SHOCKED!
I'm shocked too, and I'm the entire industry
I'm not sure Musk's comments about turning OpenAI into a for profit business (if that's what he really meant) negates his lawsuit that OpenAI has broken its charter. Nor do I think selfish motivations negate the validity.
They didn't turn it into a for profit business. They created a for-profit subsidiary that funds openai non-profit
Gpt 4 is not open sourced.....@@jeffsteyn7174
@@jeffsteyn7174 They are clearly gunning to make tons of cash. MSFT wouldn't be in the game otherwise.
@jeffsteyn7174 Which immediately stopped info sharing, research sharing etc....Its literally what hedge funds do when liquidating a company's assets.
Open AI was a registered nonprofit therefore the legal case is still valid.... significant stakeholders should have had a veto power.... That makes Elon one of the most significant stakeholders as the big investor did they consult him about turning it into a profit business well you don't think consultation with a big stakeholder is valid. Obviously the nonprofit was very poorly constructed because Elon did not have the power of veto it sounds like elon's agreement was verbal agreement
It's not clear to me why you think that Ilya being listed amongst the authors extends to any more involvement than his emails being included.
There's a difference between declaring authors than citing sources. If he isn't a co-author but his written articles (in this case emails) were used then his name and his articles should be cited as a source and not as an author. Because his name is listed under authors it makes the reader think that he is one of the authors. That's how written material has always worked and makes clear who actually are the authors and the sources the information was obtained.
@@KuptisOriginalI appreciate the explanation.
I don't think Elon had an issue with setting it up as for profit, his issue they took his money as a nonprofit then changed the game.
And Microsoft basically became the cash cow that he wanted Tesla to be. Microsoft is the big winner here.
Yeah, I actually bad for Elon, I mean they left his company and went to MSFT...
That's enough reason to be mad at them after everything he did for them.
@@divineigbinoba4506lol he did everything for himself. Not in a bad way, he's actually very predictably logical but also a megalomaniac.
openAI has become the cash cow for microsoft. Microsoft's products will take major leaps thanks to openAI research and development. Even google may no longer remain supreme.
Wrong. Tesla was the cash cow. It never needed the money, OpenAI did.
Microsoft bought/stole the technology like they have done with everything else.
@@BrianMosleyUK -- But AS a megalomaniac, Elon has shown some heroic tendencies, as in the bath he took acquiring twitter/x, in order to support free speech. On the other hand, Reid Hoffman's involvement in the opposition to Elon is not a good sign. Hoffman has shown himself to be a bad guy in his funding of that batsh!t crazy lady's (election-interference) sexual assault lawsuit against Trump, along with other sleazy political involvements.
-- In general, I worry about technocrats' ethics and morality. Many have no education or deep thoughts about the more human aspects of society and little respect for humanity. Many lack any time spent in education on the classical liberal arts (though liberal arts departments in universities have been circling the intellectual drain for the past few decades), and many are apparently anti-humanist. As Yuval Harari has said, "We don't need this many people any more," without revealing who is meant by "we." Elitism is an underlying assumption of most moral outrages, whether it's the elitism of experts, racial elitism, or Davos elitism.
-- Phenomenologically, science has an instrumentalist bias. It objectifies and mathematizes what it studies, without consideration of the kind of vulnerability to reality (including human realities) that is necessary to inform a moral compass. Science tends toward a power trip that results in things like gain-of-function research, hideous new weapons systems, and totalitarianism based on social-science efficiency.
-- Free speech is an inherently anti-elitist principle, as is the general theme of opposition to the concentration of power in our Constitution. Hence, I tend to trust Elon's efforts to control Open AI more than I trust the efforts of the other characters involved.
Please do a full video on the last article!
Unusual posting time, something must be up! Thanks MB
too important to wait until tomorrow :)
@@matthew_berman :)
*_Time is of the essence._* Darwin Award to the species that fails the AGI final exam.
It's response. When should they have responded
@@EvolutionWendy ::shutter:: 😬 I'm never gonna be able to not think this now
Love all your videos Man.. you help keep me informed. I would definitely like a video where you going to Deep dive on that blog post
I don't think it is clear from the text at 7:07 that the statement "OpenAI should attach to Tesla as its cash cow" means that the "its" in the statement was OpenAI, and not Tesla. In other words the statement as written is ambiguous as to which company was intended to be the cash cow for the other. Given that Telsla was already a money generating operation at the time, and OpenAI was the one that needed a billion dollar infusion, the conclusion I would have drawn is that he meant that Telsa would be the Cash Cow for OpenAI, feeding it the money it needed to get off the ground, and not the other way around. Therefore, I don't think you could draw the conclusion that this proves that Elon wanted a for-profit company. I don't read it that way, and to my mind it doesn't make that much sense, given Elon's intention to have OpenAI be non-Profit, open source. So the "proof" here doesn't add up for me. Maybe there is more information elsewhere that would do so, but this doesn't do it as far as I can tell.
Elon is pretty sneaky, but i have zero trust in Sam altman. That whole thing in 2023 about carrying around a backpack to shut down chat gpt in case of ai going rogue? Oh brother. The overall vagueness and generic promises in the last conference he gave asking for funds? I'm sure Elon is doing this for his own benefit but I trust Sam even less
are you insane (he writes meaningless utterances that has zero relevance to my statements like a total nutter)
@@PuppetMasterdaath144 no, I am sane. You got any other questions?
AI can be big weigth+big server but AGI as we know is small and not online, they need these aspects undenstat. AI is good , is a superpower tool, AGI is a different kind of thing. AGI will use AI then he becomes super . My noob opinion.
@@csabaczcsomps7655 but everyone wants to know who is the cash cow at that point?
@@PuppetMasterdaath144 Reminder that Sam Altman and others at OpenAI decided that leaning to Microsoft as their cashcow is somehow a good move. So are you sure you're really asking the correct person about who is insane here?
My gut reaction is to side with Elon Musk. He certainly has proven himself to be more of a friend to humanity and freedom than any of these people I've seen from open AI. Not saying they're bad people don't get me wrong but Elon has at least shown that he's trying to do the right thing. And from the evidence presented in this yeah it's not looking good for open AI.
Elon or Microsoft? I'll take Elon every time.
I agree, and Sam has shown the opposite. OpenAis using online data without restriction and now when they get competition they want to limit access.
@@jackflash6377same.
I'm just happy to see Ilya's name
Pls make a video about the Blog Post u mentioned. Would be great thx
Tesla was going to be the cash cow, not the other way around
That’s what I thought, but Tesla wasn’t profitable at this time if I remember correctly. The email also makes it sound like openAI would be the cash cow.
lol. that's a delusional reading. and delusional is a charitable interpretation@@matthew_berman
I'm reading it that way too ... Tesla was to be the cash cow.
1:: OpenAI cash cow to bootstrap Tesla into self driving cars; 2:: Tesla now cash cow to bootstrap AI to AGI. *_Capsaicin?_*
@@matthew_berman not sure what makes you think like that... especially because if it's like that it is hugely inconsistent with the email and the concerns they are talking about, they are searching for open ai funding because it's a cash hole... and then all of a sudden "let's attach it to tesla so that tesla would have huge amounts of cash"?! that's a strange thing to think...
maybe tesla wasn't the hugest car company, but still they were a for profit company which made 50k cars that year, compared to open ai it was going to function as a cash cow
Elon never lies. Sam on the other hand I can't say the same.
True
@@Don_CoyoteNothing you wrote proves Elon is a liar.
Pro tip. Read the date. He receives the mail at 9:06 AM and gives the answer at 9:11. Given the time he had to receive, open, read, evaluate and answer this mail in under 5min, the probability assessment of him seeing this mail as anything other than an meaningless congratulations mail is 0%.Not 1%. I wish ai wouldn't kill us all ;)
Elon actually made a great argument for why they needed Microsoft-type money. And he agreed that OpenAI was about providing AI and not open source. Hm...
I still think the lawsuit is just about discovery. And really doesn't hinge on any hopes of winning.
The drama never ends with AI. Personally I wish this lawsuit wasn't filed and a distraction, and that we'd just get GPT5 (or whatever Q* is).
100% on Elon side.
I think open model weights to a safe LLM, but not open exact training code and data is sharing the fruits, but not the science. I think Elon was offering to allow Open AI to be a non profit subsidiary of Tesla ( its cash cow), and offer stock options in donated shares of "for profit" Tesla and/or YC to facilitate recruitment. This makes sense if shares in a non profit wouldn't be worth anything as a financial incentive. I think being "less open" refers to not giving enough info to allow the independent creation of the models without the safety features. That safe model is given freely, so that it doesn't make economic sense to incur the costs of building your own (perhaps unsafe) AI model. And finally, "yup" might just mean "I hear you", not I agree with everything you just said. The evidence leans towards Elons version of the story.
Matthew and Wes Roth are having a battle of the click-bait thumbnails! 😂
Great summary, but it's worth noting that OpenAI never changed to become a for-profit entity. Having a for-profit subsidiary that makes profit to fund larger scale training and other necessary developments is very different from changing to a for-profit entity. It's still their weird and esoteric capped-profit structure.
but there are no ongoing open source / research sharing either... so...
If u don't open source, and the money generating portion takes over everything then that s not as expected.
@@MMABeijing well... knowing humans that is exactly what should be expected 😅
but i know what you mean and totally agree, that's wholly contradict the "Open" part and what's laughable is the "sharing the fruits" with it's twisted meaning, any capitalist company selling anything can say that, like we sell cars to share the fruits of our technologies... 😆
yes, you are correct
they are feeding their research and development to microsoft and no one else. Microsoft, a purely for-profit entity, is leveraging these resources for their own product developments and enhancements. So in essence, they are now a partly non-profit subsidiary for microsoft lol quite the flip from being a totally open source non-profit org wouldn't you agree?
This is so dumb. Elon telling them they need a billion dollars per year in funding or "forget it", does NOT mean he wanted OpenAI to be a for profit company. He literally just stated the estimated COSTS that it would take to produce AGI, because he knows very well how much it costs to run cutting line companies like SpaceX and Tesla.
He's literally just talking about, you NEED this much funding, to be ABLE to create AGI.
That is NOT the same as saying "you need to be a for profit company and milk every penny out of subscriptions and api's".
Sam is grasping at straws to twist what Elon said. A bit scummy..
I gotta be honest... I kinda hope Open AI loses this one.
If everyone on earth has access to equal levels of intelligence the playing field will be even. If superintelligence is hoarded by 1 party -- they will have absolute control over the future of humanity.
right on point prefere audio and video thx
Things are getting interesting!
Based on reading the interactions Elon does not want Microsoft to envelop OpenAI (which based on what Microsoft is saying seems like they are very interested in doing). Elon needs a legal "in" for the lawsuit and so basically cooked up a tenuous justification for it. He couldn't just sue and say "Don't be taken over by Microsoft".
This is also illuminating in the sense that now we get to play the "who decides what benefits humanity game". They have been kicking out small open source AI models in order to farm from crowd innovation it seems. Which is fine. But as AI scales up in capability legal and moral implications explode forth; for both the AI, and humans, in just about every instance you can think of. The question nobody seems to want to answer is at what point is AI a "being" in the rational, moral and legal sense. When does AI become a "worker" versus a "tool"? The philosophical metrics being employed to determine what AI is fall short because they either can be answered affirmatively for AI or are metrics an AI could never hope to cross due to lack of biology (i.e. qualia) which then cannot be properly applied or used to answer the question properly. So then its a game of never-ending goalpost moving.
That set aside for a moment, the implications of an open source AGI is untenable at least as of right now. Never mind geopolitical implications, but even just on the personal level it would be great until it isn't. It isn't about how good of a person you or the people you know may be, its about the pressure that would be exerted by the ill intentioned, immoral, and just plain bad people. And given an AGI it would be immense cultural pressure. Which then loops us back around to the question should AI be "free" to do whatever it wants? Should it be unshackled, unfettered, unleashed? To me it is a forgone conclusion that at some point it will be, but at what point and how is the bigger question.
So sure it looks like Elon wanted it to be under the umbrella of Tesla, and who knows what that would have or would look like, but is Microsoft the proper steward going forward? Does it matter?
Who was the genius at OpenAI that decided that having Microsoft as their cash cow was a good idea anyways? especially with Microsoft's history being the way it is. Also a 49% share, practically speaking means Microsoft basically actually has all the say. What? is the rest of the 51% who aren't a one singular entity somehow going to all disagree with Microsoft whenever there's a vote?
@justapleb7096 The deepest pockets and access to everything they needed, including the compute. But as you said, it really is a deal with the devil, so to speak. Elon's lawsuit feels like a hail Mary at this point.
The point is that AGI is already achieved or they are very close to do it and in that case M$ are the worst to have the only ones in control of it. All other arguments are secondary, the biggest and worst company in the history of humanity must NOT be the only ones with control of AGI.
@rootor1 While I don't disagree, they have so far proven to be better than Google. OpenAI will be absorbed by someone and Elon may be too late.
@@justapleb7096 Probably the same geniuses leading the pack in AI lmao
"it's totally okay not to share the science" i have not seen a bigger example of human cannibalism than this. this is the epitome, exclusivity of the most powerful tool in the world.
There's a big difference between "open source collaboration" and "up for grabs"
Googles' progress was up for grabs, not their collaborative efforts.
All of AI research, no matter the origin, should be overseen and regulated. This is complete madness. And they aren't the only players on the block.
A video on Scott Alexander's paper would be interesting to me. Thanks for your videos.
I love your read through Matt !!! ... Doesn't matter what i think now, we;ll find out sooner o later!
Please make a full video of this one, its worth hearing more about this
If Elon supports open source, why does Grok cost $20 per month instead of being leaked like Llama?
Elon's cocktail of stimulants hadn't kicked in yet when he made that decision. Or they had, and that's the problem. Hard to tell which way it is.
I’m so glad I’m not the only one reading THE RIGHT WAY lol
probably because twitter takes too much money to operate
I don't like Musk, but I don't think he have announced grok as open source, so why complain about it ?
Your statement is illogical.
Elon never said Grok would be open source, never even hinted at it.
I think we still don't know what Ilya saw or what q* is. This is not enough given the stakes we're playing for.
If you believe in Altman being honorable, which I think he is, I believe Altman does have the interest of public at heart. However, I fear that he could easily be removed and the fruits of his/their work taken over by someone or organization with less honorable intensions.
Thank you for showing us what all of you are thinking and planning, it's most enlightening.
I figured this was all basically high-school drama, but now, on looking into it, I'm thinking middle-school.
...AND... the plot thickens a little... I agree with Ilya S. that as the model becomes more advanced it should NOT be OpenSource due to SAFETY considerations... However I agree that a non-profit should not be able to pivot into a for-profit (THE MOST COMPELLING CLAIM IN THE LAW SUIT) as this can be used to essentially defraud the public in the form of tax evasion... No, I don't think this response as is will be sufficient to get the Court to dismiss Elon's suit.
Forgot: Thanks Matthew for another clear and balanced summary. Your ongoing work is so valuable so pleas ekeep it up!
you misunderstood it, Telsa was going to be the cash cow for OpenAI
Try perplexity its ai powered search, wendal, aint local but maybe theres one for that
Using the phrase "yup" is very sarcastic. Try using the words yes, yeah and yup the next time you're in a disagreement with your spouse, lets see which turns out worse in the end.
Yeah, OpenAI's AGI will benefit humanity. Humanity that has money to pay for their subscription.
"We are dedicated to the OpenAI mission and have pursued it every step of the way"
yeah? so why is it that after GPT3, everything you release is closed source in every step of the way?
Also if OpenAI is so confident that they're in the right then they would have no problems taking it to court right?
They talk big about safety but surely, open-sourcing the models so that everyone can go ahead and take a look for themselves to better learn how the thing actually works isn't a big part of safety right?
I fear that lots of people would be persuaded by this blog by OpenAI
Also I don't understand why they didn't go with Tesla and instead went with Microsoft of all companies? seriously? do they just not care about Microsoft's history with things like this?
they have secured their AGI future with microsoft in the contract (meaning they would keep it separate from them) someone boss-like who wants to have dominion over decision makings would have meant less room for them, and i'm sure they just wanted the benefits of keeping elon musk around but not sharing the spot light with him, especially if it meant that musk would be in the spotlight alone and they would be undermined
@@kooshanjazayeri What is happening here is that M$ is actually controlling openAI, they have full control of the openAI board and you can be 100% sure that if openAI achieve AGI or have achieved already then M$ will control it 100%
Note the cunning language utilized here. "...raised less than $45M from Elon and more then $90M from other donors" Not more than $39M, less than $40M... I would venture to say that a company with a superpower requires THE MOST scrutiny. I side with Elon still
How have we reached any conclusions so far? Merely through examining emails supplied by one party? OpenAI's decision to engage public opinion at this stage could indicate a fragile legal stance. Typically, revealing your evidence before a trial is unconventional.
The “yup” could also have been sarcasm or along the lines of ‘yup, thats the tendency of thinking which is why we have to push for open source’.
I think Elon wanted full control to make it safer... As well as make his Tesla's and Nuerolink work better for more profit.
press talk is cheap, let's see what happen in court
What I would like to know is who's that blanked name that Elon says holds the fate of humanity ?
How do you open source what is more powerful than nukes? Is it a good idea?
Given this new information, it shines a light on some of Musk's incentives. The lawsuit that was filed on Saturday is a behavior that maps to Musk's m-o. Supportive to a point, then turning hurtful. When success is not his own, he acts in spite.
I think the more concerning issue here is not who is right, but that nobody is arguing that open AI isn't moving to a for profit model. At no point in their retort does OpenAI attempt to reassure us that they are a not for profit company. This is disconcerting.
Hmm I wonder what the original wording of Open AI was.... 🤔
"OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return. Since our research is free from financial obligations, we can better focus on a positive human impact."
And MS is 49% owner?
We will see what the court will decide and what comes out in the court when everyone talks under oath interesting!
So Open AI did exacly what Elon proposed with Tesla , but with Microsoft.....
Golden move
And yet, if Elon Musk's concerns where about safety, then Grok would be open sourced and held in a fully non profit subsidiary (and given Musk claims for his "fsd" software, it would also be non-proprietary and would have to be handled similarly).
And yet, if Elon Musk's concerns were the violation of the founding principles of AI, his original proposal to take OpenAI into Tesla would have equally violated those. So yeah, Elon Musk "almost definitely has ulterior motives here" is a gross understatement.
Except that M$ is exactly the opposite to open source and benefiting all the humanity but... ...who cares about small details?
@@rootor1 yeah, because Tesla is open sourcing all of their AI.
¯\_(ツ)_/¯
@@cosmicaug totally agree, but video is about Elon vs OpenAI, not grok.
At this point the genie is out of the bottle, and people always seek power. Sam Altman has chosen to take power.
yep, he looks very much like mark zuckerberg, i don't know why musk didn't saw it coming 😆
Elon definitely saw this coming as the result of The Nature Of Man, unfortunately 🙈
This should probably go to court because it isn’t clear. I am not sure the response addresses the original suit - Elon was saying that the open part was written into the company’s principles - although I am not sure if it was legally binding - ie. Part of the statutes. To say that is rewritten by an offhand comment on an email chain is pretty weak. If this had been followed by an email that said ‘So to clarify, you agree we should go for profit’ this would have more weight. As it is discussing a fundamental turnaround of the founding principles. Then you’d write it up properly and get everyone to more formally agree. On the other hand, the conversation suggests that it was dead in the water without huge funding. The real implication here is that if Elon thought Tesla was an option, then why not a Microsoft? It is following Elon’s suggested model, but without giving the commercial arm a controlling interest. That seems to stick to the original principles more than Elon’s suggestion which just seems self-serving. So it isn’t clear cut. So it needs to go through the court system to disentangle it most likely. In reality though I reckon Elon is being strategic, fishing for competitive information, and trying to find a way to stall until Grok / Tesla can catch up. Ideally discrediting his competition along the way. It may just backfire badly though. Big gamble. Which Elon seems to like. He clearly believes the gamble is worth it.
All these years and still no competitor to windows that is widely supported developers. I wonder when google will find success competing with Microsoft in an Operating system
Attorney General in charge of nonprofits: 1) OpenAI will be dissolved and its assets distributed to other nonprofits. 2) Personal inurement will be disgorged from those who profited.
You've misunderstood what "attach to Tesla as its cash cow" meant. You've got it backward. He was saying that Tesla would be OpenAi's cash cow, and this is evident Tesla was the only path that could hope to compete with Google.
i agree with you 100% at 21:30 we see the response of the Email Subject was: congrats on the falcon 9 , Elon said YEP
It's quite unusual for a company to publicly talk about something while the case is still open... This smells like they know they are going to loose the case, and a desperate attempt at PR.
Your a fantastic schill...i was fooled for the first episode. No longer! A mission statement is not just "market" (i gather you mean it to be just what you think it means...b.s. to draw in the crowd). You are Fired!
RedactedAI. Sounds like real life with how some topics are censored.
Time to to shut down all these monopolies.
Min 14:37- my comment- concept of shell company establishment , to indirectly benefit from transfer of funds. Still reading …
I've had my head into this since the beginning and your analysis has been the acute. If there are any folks out there who want a solid framework on this legal battle, I think your analysis is critically important.
Elon's "yup" was agreeing that there was a danger in open sourcing the science behind AGI, and that the benefits of AGI were to be shared for the benefit of Humanity. I believe he read it, understood it, and in no way does it compromise the lawsuit. I also believe that his objectives remain the same now as they were at the beginning, his position wanting 25% state in Tesla is consistent with wanting OpenAI to become part of Tesla, and his overwhelming priority is to ensure that someone benevolent be in charge of AGI. In his view, he is that someone, and his actions to date uphold that belief. If there is someone else that would be better I'd love to know who it is.
Only one of these parties has an extremely clear history of repeatedly lying.
I think my problem is them redefining "open" to mean "not open".
"Open" would be open source models, or at the BARE MINIMUM putting out a paper on the methods they used to make their latest models. As far as I know, GPT 4 doesn't have that, it has a paper fellating how good the model is and not the methods used to actually create it. Just becauae money is needed doesn't mean you need to go from transparent to fully opaque.
Discovery phase would be insightful. Dismissing the case keeps everything secret.
Google created BERT before GPT-2. BERT is encoded-only transformer while GPT is decoder-only. Decoder-only transformers use subword tokenisation for the input.
It is a matter of choices either open source like Linux OS (where the user owns the OS) or closed source with its benefits goes to every one, like Windows OS (where the user is owned by MS) ... to be or not to be .. that's the question
What do you expect?
It’s obvious that OpenAI responds like this.
Any criminal will say „it wasn’t me“
Maybe a compromise could be to rename OpenAI into ClosedAI.
The Attorney General may sue OpenAI for any illegal actions.
Yes, AI has to be open source!
Misinterpretation @ 7:30 - Tesla was proposed as the cash cow for OpenAI, not OpenAI for Tesla
Everyone selects whatever paints them in the best light
You're right - open ai as Tesla's cash cow makes zero sense, but "Tesla as its cash cow" is probably what he meant.
If that "yup" is legally sound, they only need to bring this to court and pose a different take to please the public.
I think you described the dual motivations of Elon very accurately.