Thank you Musk, for a bombshell thrown down to the calm AI circle.
OpenAI, which has been entangled in various copyright lawsuits, has made matters worse, and the plaintiff is the former founder and chief - Musk.
He alone sued OpenAI and its associated 8 companies, including Ultraman and others.
Actually, it's been a long time since the two families were talking about it, and Musk has already bombarded OpenAI many times.
Even in April last year, someone asked Musk on Twitter:Why not sue OpenAI for cheating on him.
Musk also replied:Wait and see.
As it turned out, the old horseWords must be done, deeds must be fruitful.
Of course, OpenAI is not instigated at all, and according to Axios, the accusations are firmly denied internally. The CEO of OpenAI, also mocked:It may be due to its regret that it is not involved in the company today. ”
Ultraman also replied on Twitter:Always at your disposal.
However, Shichao finished reading this 120,000 words of litigation documents, but I want to say that OpenAI this time, it may indeed be a bit troublesome.
Listen to me take this report apart and talk about it.
Musk has sued includingBreach of contract, breach of fiduciary duty, unfair business practicesincluding 5 countsDemand the return of revenue, open source technology, and the restoration of non-profits.
The core of its litigation can be summed up as follows:
OpenAI, the kid, violated the non-profit founding agreement for the benefit of all mankind and became a tool to make money for Microsoft.
"Seek open source technology for the benefit of the public".This goal is not just Altman's verbal promise to Musk. It is also written in OpenAI's company registration certificate.
Don't just focusHuman interests, do not profit, but also to adoptOpen-source model
As a result, you thick-eyed openai didn't do any of it.
It started as a non-profit organization and became a limited profit.
Secondly, what Musk doesn't understand is that he said that he wanted to open sourceWhy is GPT-4 still tucked away after it was released for a year?
Neither willing to open source, nor revealing technical details.
But he showed his own ** to Microsoft. Also in 2019, a deal was signed with MicrosoftAgreed to license pre-AGI technology exclusively to Microsoft.
According to the lawsuit documents, OpenAI's rhetoric of benefiting the public has actually made GPT-4 beMicrosoft's proprietary algorithms, integrated into the Office suite
Microsoft can make huge gains by selling GPT-4 to the public.
In addition, OpenAI has more than just a GPT-4. For example, they're working on a secret project called Q*, so strong that even internal employees have warned about its potential power.
Are these confidential technologies included in this exclusive licensing agreement?
Musk also threw out a bunch of statements to prove that in front of Microsoft, OpenAI has long been not so independent.
Last year's wonderful palace drama was forced by Microsoft to force Ultraman to reinstate.
According to the new interpretation of the ship in the lawsuit report, after the board of directors was reshuffled, it turned out that the internal onesThe structure that guarantees a non-profit mission is completely brokenFinish.
In January of this year, Microsoft's vice president swaggered into OpenAI's board of directors.
If Ultraman's return to the factory is not successful, Microsoft is most likely to get all the results of OpenAI.
Because, after Ultraman left the board of directors, he boldly said: even soOpenAI is gone, and he's not worried。Because they have the talent, the computing power, the data, the dataand all.
It means that if OpenAI doesn't accept it, he will be able to take all the results and move into Microsoft.
This series of operations, no wonder Musk will accuseToday, Oepnai has long since become a closed-source subsidiary of Microsoft, the world's largest technology company.
The development of so-called general artificial intelligence for the benefit of mankind is all nonsense, but I just want to make a few more money.
OpenAI's various heart-changing behaviors broke Musk's heart.
Reading between the lines, they are crying about how they were betrayed by Ultraman. How did Ultraman take the initiative to find himself, how did he promise to join hands to build a non-profit organization to fight Google.
Musk, who was moved back then, was before and after the saddle, and he used his connections to recruit people, and contacted him one by one.
Tens of millions more are being thrown in, and from 2016 to 2020, Musk contributed more than $44 million. Even, he found and rented OpenAI's office, and he bought the first supercomputer.
Why did I pull openai big with a handful of and a handful of pee,In the end, he was tricked.
Musk described the feeling as if you were giving oneProtecting the Amazon rainforestof the non-profit organization, and then he turned around and made a profitAmazon Logging Companyand use the fruits of your donation to cut down trees.
Musk really couldn't figure it out. How did a non-profit organization that spent almost $100 million on its own become a for-profit organization with a valuation of $30 billion?
He also took the issue to a higher level.
If OpenAI can continue to operate in this way, it will not be that all entrepreneurs can cheat donations under the guise of non-profit. After the research is successful, turn around and start a new company to make money.
That's why I said that OpenAI might really be a bit of a hassle this time.
Musk not only used this to threaten to open source the open-source technology and turn it back into a non-profit organization, but also demanded the return of the income during the contract.
Maybe when you see this, you will think that "all human beings are not all human beings, they are all high-sounding".
In business. Musk may just feel that his X AI can't do it, and he is afraid that the two will join forces, resulting in him having nothing to eat in AI.
Let's be honest, not necessarily, but let's rewind the clock to 2012.
That year, Lao Ma met the co-founder of Deepmind, and the other party told him that AI is very dangerous to human society, and that there may be no need for people in the future.
An investor also said,The best thing he can do for humanity is to shoot the founder of Deep Mind on the spot.
As soon as this remark came out, one was planted in Musk's heart"AI threatens humanity".of seeds.
Since then, he has been an AI-safe preacher.
I even went out of his way to contact Obama and said that AI is dangerous, so you should take care of it. Of course, the end result is that people don't bird him at all.
Later, he started to prevent Google from attacking Deepmind.
In '13, Google was buying Deepmind. Lao Ma chatted with the founder for more than an hour and wanted to invest in the acquisition. Unfortunately, he died in the end.
It wasn't until 2015 that he met the destined Ultraman.
The two hit it off with their ideas, and Ultraman took the initiative to sayLao Ma, let's build a non-profit artificial intelligence lab together to catch up with Google.
It was this sentence that triggered a series of subsequent chain reactions.
Therefore, at least from these points, Lao Ma is consistent with his words and deeds.
It's just that whether this is his pure idea or mixed with the idea of disrupting OpenAI, it is unknown.
In addition to the above, he asked:Thoroughly investigate OpenAI's donation programs, as wellGPT-4 is AGI or not
To be honest, compared to OpenAI's profit-making or not, Shichao cares more about these things that are indeed hidden. And what is the true face of the Q* algorithm? What is the core technology of GPT-4?
Don't worry too much, this play won't keep us waiting long.
A case management conference is currently scheduled for July 31, 2024.
I believe that when the time comes, it will be a very exciting debate.