In April, DeepSeek unveiled its V4 technical report, which disclosed that among the nearly 300 authors acknowledged, 10 were identified as 'former employees.' This group included key contributors such as Wang Bingxuan and Guo Daya, who have since transitioned to prominent corporations like Tencent and ByteDance, spanning various pivotal technical domains within DeepSeek. Since its inception three years ago, DeepSeek has adhered to a policy of eschewing external funding. While this approach has shielded the company from capital-induced deviations in its technical trajectory, it has also placed DeepSeek at a competitive disadvantage in attracting and retaining AI talent. The lucrative salaries and vested stock options offered by rival firms starkly contrast with DeepSeek's employee stock options, which lack a valuation basis and are challenging to liquidate. Presently, the AI sector is grappling with a severe talent deficit, with major players leveraging generous compensation packages and marketable stock options to lure core talent away from startups. DeepSeek's decision to publicly disclose the names of former employees is a relatively uncommon practice. Notably, at least five key R&D personnel have been poached over the past year. There are now reports indicating that DeepSeek is contemplating its inaugural round of external financing, with the objective of bolstering the value of its stock options to retain staff. This development has prompted industry-wide introspection: How can AI startups retain their core talent while staying true to their original vision? How should teams dedicated to independent R&D navigate the high-stakes market environment? The responses to these queries could significantly shape the future of China's AI landscape.