In just one week, the situation changed suddenly. Anthropic, which was once the "only darling" of the US DoW (Department of War, also known as the Department of Defense) in the confidential field, was out of the game and was even blacklisted - any company that has a contract with DoW cannot cooperate with Anthropic. Then, the position vacated by Anthropic quickly attracted competition. xAI is on the table quickly. While OpenAI publicly supports Anthropic, it also suddenly comes to the table.
A decade ago, Silicon Valley was generally ashamed of cooperating with the U.S. government (especially the military), and there were even incidents where employees collectively petitioned to terminate cooperation. Yes, it’s you, Google.
Nowadays, we can see the cutting-edge AI companies in Silicon Valley staged the "Three Kingdoms Killing" around the contract with DoW.
While eating melon, I couldn't help but sigh, the world has finally changed.
Silicon Valley is less idealistic and more idealistic than before.
01
Origin: Anthropic was kicked out
Originally, everything was calm.
Many AI companies in the United States actually cooperate with DoW, but they are all in the "unconfidential" category. Only one company has a contract in the "confidential" category, and that's Anthropic.
At this point, Anthropic is definitely ahead of its peers.
But in the past few weeks, things have suddenly taken a turn for the worse.
DoW stated that it will unify its contracts with private AI companies in the future and make a direct agreement: allowing "any lawful use" of its technology.
By introducing this default clause, we no longer have to negotiate the scope of technology applications one by one with private companies. If you cannot accept this default clause, we will not negotiate. Those who have already signed the contract must update the contract, otherwise they will be replaced.
This involves the $200 million contract between Anthropic and DoW, and the two parties began to communicate.
Anthropic is unwilling to accept it. Founder Dario Amodei publicly stated that Anthropic has two "red lines", one is large-scale surveillance, and the other is completely autonomous weapons. It will definitely not do these two, and it must be written clearly in the contract not to apply Anthropic's technology to these.
Anthropic has always emphasized safety, and this move is indeed consistent.
At this point, it seems that Anthropic and DoW are still in dispute over the terms of the contract.
But no one expected that Anthropic and DoW would intensify so quickly and to such an exaggerated degree.
To put it simply, DoW was very tough and gave Anthropic three days, basically a "notify you" attitude.
Anthropic is also very tough, repeatedly emphasizing: I can’t do it, I’m sorry.
When the three-day time limit expired, DoW announced the termination of cooperation with Anthropic and started a war of words on social media.
The Minister of DoW directly called Anthropic founder Amoudi a "liar" and "self-righteous" on X. Trump also personally dismissed Anthropic, calling it a "radical left-wing awakening company" that "doesn't understand the real world."
You may not be satisfied with not cooperating with Anthropic. The DoW minister also announced an explosive news: Anthropic will be included in the list of "supply chain risk companies".
You know, this list has never been to an American company before, and this is the only "treatment" Anthropic has received.
What does it mean to be on this list?
The Minister of DoW explained: Once on this list, any company that has a contract with our DoW cannot cooperate with Anthropic.
This is pretty crazy, Anthropic is known for its enterprise users.
The company recently raised $30 billion at a $380 billion valuation and disclosed annualized revenue of $14 billion, driven by growing corporate demand for its products. Currently, more than 500 enterprise customers spend more than $1 million annually on Anthropic's products.
For example, Amazon has invested billions of dollars in Anthropic and is also an important partner of DoW.
02
OpenAI sides with Anthropic, Musk comes out to "take over"
Buying and selling cannot be done without justice, and such drastic actions by DoW instantly ignited controversy in public opinion.
The most surprising thing is that on the cusp of this crisis, OpenAI’s Sam Altman chose to publicly side with Anthropic.
He sent a memo to employees, saying that the two points raised by Anthropic are also the bottom line of OpenAI, and expressed his desire to help Anthropic ease the tense relationship with DoW.
Whether it is the two companies Anthropic and OpenAI, or Ultraman and Amoudi personally, they are old enemies.
At the recent Indian AI Summit, Ultraman and Amodei took a photo with their fists raised high because they did not want to hold hands like others, and it became a meme.
In the blink of an eye, Ultraman stood with his opponent despite the pressure? ! This kind of mind, this kind of consciousness, makes many people look at Ultraman with admiration.
The well-known Ilya Sutskever, the co-founder and former chief scientist of OpenAI, broke from his usual silence and publicly applauded Anthropic for "not backing down", and also mentioned OpenAI for the first time.
It can be seen that for him, this matter means far more than the grudges with his old club.
Of course, it is said that there is a fierce public controversy, and certainly not everyone agrees with Anthropic.
Under Sutskvi's tweet, a high-like comment came from a well-known conservative lawyer with 420,000 followers, exclaiming"Oh my God, you are so arrogant."
His point of view is actually DoW's point of view - as long as DoW abides by the law, why should it listen to the dictates of private companies on how to act and what it can and cannot do?
The AI company that is most involved and high-profile this time, apart from OpenAI, is xAI.
Musk won't miss this drama. He has been chasing Anthropic recently, saying that the company is sanctimonious. While saying that his model has been plagiarized, he has actually been involved in a lawsuit due to copyright issues. In a tweet, Musk even called Anthropic "hating Western civilization."
Even when Anthropic’s official account announced the good news of raising $30 billion, Musk was chasing after him and scolding it.
As if this farce wasn't exciting enough, just as Anthropic was kicked away by DoW, xAI "took over".
xAI has reached an agreement with DoW, and Grok will be used for "confidential" purposes.
In addition, DoW also told the media that communication with companies such as Google and OpenAI is also in progress.
03
Big reversal: OpenAI also entered the game, leaving Anthropic alone and embarrassed.
Perhaps everyone still underestimates the value of the "triangular hatred" between Ultraman, Amodei, and Musk, and underestimates the fists raised by Ultraman and Amodei at the Indian AI Summit.
Just when the media was still talking about the "hero team-up" of Ultraman and Amodei and Musk's cutting off the Hu, a shocking reversal came.
Ultraman suddenly announced on X: OpenAI and DoW have also reached a cooperation.
·
What is confusing is that in the statement released by Ultraman on X, he still mentioned the two "security principles", namely mass surveillance and autonomous weapons, saying:
· DoW demonstrates a strong focus on security during all interactions.
· DoW recognizes these principles and is reflected in laws and policies, and we have also written them into the agreement.
· We will deploy FDEs (Field Deployment Engineers) to assist in model operation and ensure safety.
· We are asking the Department of Defense to provide all AI companies with the same terms that we believe every company should be willing to accept.
In other words, OpenAI not only reached a cooperation with DoW, but DoW just rejected Anthropic, and then agreed to OpenAI on the same terms?
Maybe he was too happy. Not only was Ultraman's statement quite long, he also posted it three times:
As soon as the news came out, everyone was confused. The most important thing was that they didn't know how Ultraman and DoW reached a consensus on those two key controversial points.
Although Ultraman's attitude is still that of a savior, and he wants to benefit Anthropic and reach a cooperation, many people don't buy it.
There is no lack of doubts about Ultraman in the comment area. Firstly, they criticize him for speaking in roundabout ways and not explaining clearly; secondly, they criticize him for always being like this.
In the final analysis, they do not think that Ultraman really made DoW "compromise", and they do not think that Ultraman did what Anthropic's Amoudi did not do. Instead, they think that Ultraman must have compromised, but tried to distort the facts with confusing narratives and expressions.
It is unclear what exactly OpenAI has done. What is certain is that within a week, Anthropic was selected from DoW’s “confidential” cooperation and became a member of the “blacklist”. While xAI and OpenAI will enter DoW’s “confidential circle”, Google is likely to be on the way too.
In the latest half-hour interview, Anthropic’s Amoudi answered several questions carefully and revealed some interesting information——
Amodei said the DoW said Anthropic had been added to the "supply chain risk" list, but the company had not received any formal documentation.
In addition, Amodei said that the DoW Minister’s explanation of the blacklist was misleading. The Minister said that after Anthropic is on the list, any company that has a contract with DoW will not be able to have any cooperation with Anthropic.
As mentioned before, this will definitely be a major blow to Anthropic.
Amodei emphasized that according to legal provisions, if Anthropic is on the list, companies that have contracts with DoW cannot be related to Anthropic in the projects under this contract.
If this were the case, the impact surface would be much smaller.
In other words, DoW either wants to "punish" Anthropic beyond the legal provisions, or it is exaggerating the truth in public opinion.
Amodei still insists on two red lines and believes that DoW's operation is at least "punishment" and "retaliation" in nature.
To translate, it is true that our AI ethics conflicts with DoW. If it doesn't work, we can just not cooperate. You, DoW, can think that private companies should not interfere with how DoW does things. In the same way, you cannot "punish" or retaliate against me because I insist on my AI ethics.
This is not just a contract dispute, it is clearly a unique struggle for power and voice.
It’s not that Silicon Valley is no longer ideal, it’s more spicy. After all, AI is developing too fast.