新员工立刻拿期权 ，奥特曼抢人不和张一鸣拼现金

Core Viewpoint - OpenAI has eliminated the waiting period for stock options for new employees, allowing immediate access to equity, reflecting a competitive strategy to attract and retain talent in a highly competitive AI job market [1][4][7]. Group 1: Policy Changes - This is the second adjustment to OpenAI's stock option policy this year, having previously reduced the waiting period from 12 months to 6 months before completely removing it [2]. - Other companies, such as xAI, have also shortened their waiting periods, although specific details have not been publicly disclosed [2][14]. - The traditional model of a one-year waiting period followed by a multi-year vesting schedule has been a standard in Silicon Valley for decades, but OpenAI's approach is seen as a radical departure from this norm [6][16]. Group 2: Talent Retention Challenges - OpenAI is facing significant talent retention issues, with a two-year retention rate of only 67%, which is lower than competitors like Anthropic, which boasts an 80% retention rate [3][8]. - The company has experienced notable talent losses, including the departure of eight core researchers to Meta in a single week, prompting OpenAI to reassess its compensation strategies [7][8]. - The aggressive changes in stock option policies are part of a broader effort to combat the pressures of talent competition in the tech industry, particularly in AI [7][19]. Group 3: Financial Implications - OpenAI is projected to spend approximately $6 billion on stock option incentives this year, which constitutes nearly half of its expected revenue of around $13 billion [23]. - The shift to immediate stock option access may increase costs for OpenAI, as it allows employees to potentially leave with significant equity after only a few months of employment [20][23]. - This financial strategy contrasts with the original intent of stock options, which aimed to align employee interests with long-term company growth, now seemingly inverted due to the intense demand for AI talent [23].