How unbelievable! QuantumBit has exclusively learned from multiple sources that Meta has officially announced a shocking piece of news internally:
Effective immediately, if FAIR wishes to publish papers externally, they must first undergo review by the TBD lab.
Behind this is Meta's Chief AI Officer, Alexandr Wang, who is sparing no effort to consolidate stronger AI forces into his hands.
Not through externally poaching talent with sky-high salaries, nor through cultivating them via campus recruitment.
This time, his approach is a brazen "seizure and appropriation."
No one expected that the hand of appropriation would this time reach directly into Meta's own internal colleagues—FAIR.
Speechless and utterly bewildered, after Meta stirred up the Silicon Valley talent market with exorbitant salaries, it is now throwing its internal affairs into disarray.
The specific details of Meta's new internal regulation are as follows:
Any paper FAIR intends to publish must first pass through TBD.
If, after TBD's review and evaluation, the paper is deemed to have high value, then the paper will be "held back," temporarily halting the external publication process.
Concurrently, the paper's author will also be "held" by TBD.
The author of the paper will be required to apply the new technology proposed in the paper to Meta's products, and only after it has truly been implemented can they return to FAIR to continue their regular work.
As for how these "valuable" papers, as determined by TBD, will be handled subsequently, or whether they will still be published, we are still unclear.
However, one thing is very clear: this announcement, which is tantamount to directly snatching people from FAIR, caused an uproar.
Sources indicate that Alexandr Wang came up with this method because his attempts to transfer some excellent talent from FAIR to TBD were rejected.
Several insiders close to Meta have informed us that some individuals within FAIR, unable to tolerate such regulations, have quickly jumped ship to other AI startups.
To be fair, this matter is indeed baffling.
Because according to the latest results of Meta's internal AI department reorganization, TBD and FAIR are parallel, not hierarchical, relationships—
Ten days ago, Mark Zuckerberg carried out the 4th reorganization of Meta's internal AI departments within six months.
After the reorganization, Meta's AI departments are ostensibly divided into the following four parts:
TBD, the "to be determined" lab, which now reviews FAIR papers, is led by Meta's Chief AI Officer Alexandr Wang.
This division focuses on cutting-edge exploratory research, with its specific development direction constantly evolving and being explored.
FAIR (Facebook AI Research), long known to everyone, focuses on long-term, technical AI scientific research and is more research-oriented.
FAIR's current head is Rob Fergus; in May this year, he just returned to Meta from Google DeepMind.
The reason for saying "returned" is that he and LeCun co-founded FAIR in 2013 and then worked there for six years.
It was thought that after the reorganization, the four departments would be on equal footing, each performing its duties, but unexpectedly, Alexandr Wang's hand reached directly into another's bowl...
This further confirmed the rumors that circulated during the AI department reorganization:
At the time, some speculated that among the "Four Heavenly Kings" who manage the four departments, 28-year-old Alexandr Wang seemed to have been placed in a higher position by Mark Zuckerberg.
Because the internal memo announcing the reorganization results was issued by him in his capacity as Meta's Chief AI Officer (not by Mark Zuckerberg).
Currently, it is unclear whether TBD interferes with the daily operations of the other two departments in a "superior-subordinate" manner.
But the news that has already come to light is enough to evoke a sense of regret and irony.
FAIR has a long history, remarkable achievements, and has always had Yann LeCun at its helm, but frankly, its fate has not been particularly kind:
In the metaverse era, it was assigned to Reality Labs; in the generative AI era, it was assigned to GenAI; and in the AGI era, it became one of the four MSL departments.
But who knew its current predicament could be even worse than outsiders imagined—
It must endure interference in paper publication from a parallel department and face forceful "secondment" of talent by a statistical department to implement technology...
Does Mark Zuckerberg still remember the original intention behind founding FAIR?
Perhaps he hasn't forgotten, but in Zuckerberg's mind, "doing everything possible to revive Meta's AI" is currently the highest priority.
This answer might have been evident when he chose to heavily invest in a 49% stake in Scale AI and brought Alexandr Wang in as Chief AI Officer.
Because he, who proactively dropped out of MIT at 19, started a business young, and became the world's youngest self-made billionaire at 24, has had his conduct widely criticized, with outsiders consistently evaluating him as "unworthy of his position."
To put it nicely, he achieved results through ambition.
To put it bluntly, he is unscrupulous in achieving his goals.
More specifically, this person is a spitting image of Mark Zuckerberg in his early days.
For example, in 2018, his co-founder at Scale AI, Lucy Guo, left him due to disagreements over the company's development direction.
Lucy Guo stated that the core of the disagreement was:
"I believed we needed to ensure workers were paid on time, while Wang was more focused on growth."
For instance, Scale AI employs data annotators worldwide and has been widely criticized for issues concerning wages, working hours, content, and employee mental health.
But Alexandr Wang's response was merely a casually delivered sentence:
"I believe AI should bring together the collective intelligence of as many people as possible."
And now, as the head of TBD, he is once again directly interfering with the research work of the peer-level FAIR.
Research is no longer independent; papers must take a backseat. That ideal bastion in a top tech company, "to advance artificial intelligence technology through open research," has completely vanished with a single announcement.
The one who personally tore down this flag of idealism is a young man with a cold demeanor and a results-oriented approach.
A young man who is almost a carbon copy of Mark Zuckerberg:
Both are programming geniuses, both dropped out of prestigious schools, both betrayed partners, both became billionaires in their early twenties, and both are equally firm in pursuing "I want results."
Therefore, at this critical juncture of the AGI era, Mark Zuckerberg did not hesitate to hand over the reins of Meta's internal AI strategy to this person, who is even "more Zuckerberg" than himself.
And so, the new official's first act was to target FAIR.