Meta's individual-hiring approach is the wrong way to build AI teams

“We now have a stronger pre-trained base model than the model that currently powers GPT 5.6 Sol.”

“And in fact, we used one-quarter of the compute that OpenAI used to train their model.”

“12 times less computation than was used for DeepSeek, for example, on compute efficiency.”

““In the coming years, it will directly influence the remuneration of senior leaders, including the Board of Management,” Fouquet wrote.”

“I think we should all be worried about sending intentionally or inadvertently or otherwise, signals that send the most talented next wave of AI software engineers to other countries rather…”

“We really set out to have the best transition ever, a textbook example that people could look back to and learn from. It actually started years ago.”