@FinanceYF5: Meta CTO confirms that over the past year, Zuck has had to enter 'founder mode'. He also greatly appreciates Alexandr Wang joining the team last year. To build the latest model, Meta made an emergency pivot: pulling thousands of employees from their original roles to generate expe…
Summary
Meta CTO confirms that Zuckerberg entered 'founder mode' last year, and appreciates Alexandr Wang's joining. To build the latest model, Meta urgently mobilized thousands of employees to generate human-guided expert traces for post-training, highlighting the reliance of cutting-edge models on high-quality validated data.
View Cached Full Text
Cached at: 07/14/26, 08:19 AM
Meta CTO confirms that over the past year, Zuck has had to enter “founder mode.”
He also highly appreciates Alexandr Wang joining the team last year.
To build the latest model, Meta made an urgent pivot: pulling thousands of employees from their original roles to generate expert traces.
These expert traces are manually guided code examples used for post-training the model.
This also confirms one thing: the bottleneck for frontier models is not just raw compute, but high-quality, human-verified reinforcement learning data.
Worth a listen.
Similar Articles
@_FORAB: So surreal. Recently, Meta founder Zuckerberg, on the eve of massive layoffs, had an internal public talk that was recorded by an employee. He revealed that all employees' devices are routinely tracked by the company to train AI models. The reason is to turn everyone's working methods into training data, so that fewer human workers are needed. This is too strang…
Meta founder Zuckerberg revealed in an internal talk that the company routinely tracks employee device data to train AI models, aiming to turn working methods into training data to reduce labor needs. The talk was recorded and exposed by an employee on the eve of layoffs.
Meta's AI Chief On AI Beef, New Models And Life With Zuck (3 minute read)
Meta's AI chief Alex Wang discusses the company's new Muse Spark model, efforts to catch up with rivals like OpenAI and Anthropic, and his transition from Scale AI in an exclusive interview.
@FinanceYF5: Leaked audio from Meta all-hands meeting shows Zuckerberg telling employees that they are being used to train AI. Before laying off 8,000 people, he explained why their work data is more valuable than outsourced data.
Leaked Meta all-hands meeting audio reveals Zuckerberg acknowledged using employee data to train AI, explained why it's more valuable than outsourced data, then laid off 8,000 people.
Zuckerberg 'Admits' Meta's Layoffs Were Ineffective
Mark Zuckerberg reportedly admitted that Meta's recent layoffs and restructuring centered around AI agents have not produced the expected acceleration, highlighting ongoing strategic failures and a 'vibes-based' management approach.
@FinanceYF5: 1/ Meta is dismantling its engineering base. The most striking point of this article is: Meta didn't start cutting when business was bad. On the contrary, at a time when revenue and profits were strong, it turned engineers from "profit creators" into "people whose costs need to be squeezed."
Meta proactively adjusted its engineering team structure when revenue and profits were strong, transforming engineers from profit creators into targets for cost control, reflecting a deep organizational change.