本周在20VC节目中，Rory O'Driscoll 和 Jason Lampkin 讨论了科技界最大的新闻，涵盖了人工智能监管、新兴的人工智能助手市场、重大的并购活动以及主要的融资轮次。 对话始于 Dario Amodei 呼吁“放慢”人工智能发展“前沿”的步伐，这一想法得到了 Sam Altman 和 Elon Musk 的支持。Rory 对“人类灭绝有10%的可能性”的说法表示怀疑，他认为，如果这种风险真实存在，政府将会更积极地介入。他指出了“世界末日”式人工智能的言论与实际市场反应（半导体股票小幅下跌，网络安全股票上涨）之间的差异。Jason 将 Amodei 的声明视为一项战略举措，旨在 Anthropic 潜在的 S1 文件提交之前解决风险因素，强调了 CEO 管理此类担忧的责任。Rory 承认网络风险真实存在，但驳斥了经济风险的说法，他认为对人工智能“失去控制”（递归式自我改进）是唯一真正引人入胜且尚未解决的危险。两人都同意 David Sacks 的观点：如果一个产品如此危险，开发它的公司要么修复它，要么停止开发。讨论也触及了人工智能能力的“黑暗版本”不可避免地会泄露，将其比作核技术的扩散。 接下来，“人工智能助手之战”成为焦点，Meta 推出了 Muse，直接与 Instinct 和 Town 等初创公司竞争。Jason 赞扬了 Muse 的无缝功能，将其速度和效率归因于 Meta 庞大的基础设施和专有的 Muse LLM。然而，他对这类助手的“杀手级应用”提出了疑问，质疑仅仅依靠便利性是否足以推动其在调度或预订等基本任务之外的广泛采用。Rory 反驳说，“超级应用”的方法，类似于微信，通过在许多功能上提供渐进式改进，仍然可能是一条可行的道路。 谈话随后转向了 Instinct 传闻中的10亿美元融资，估值达到100亿美元。Jason 扮演“反对者”角色，提出了支持和反对投资的论点。他指出了 Meta 的局限性（非跨平台，以及在最初炒作之后可能缺乏持续投入），这对 Instinct 来说是一个优势。然而，他警告不要与 Meta 庞大的资源和初创公司高昂的基础设施成本竞争，暗示这样的投资可能完全取决于并购结果。Rory 承认这些投资的“高风险、高回报”性质，特别是考虑到万亿美元级公司对世代人才和战略收购的追求，尽管他强调，随着估值飙升，风险回报比正在迅速下降。 Miro 以13.5亿美元被 Bending Spoons 收购，相比其175亿美元的估值大幅下降，这突显了2021年牛市中被高估的独角兽公司所面临的残酷现实。Rory 将其称为一场“不可避免的清理行动”，认为这对早期投资者和创始人来说是一个积极的结果，而后期投资者则获得了较少的回报。Jason 用“鸭子、鸭子、鹅”的比喻来解释，暗示对于前人工智能时代的公司来说，剩下的“椅子”（收购机会）已经不多了。Bending Spoons 追求无情效率、提高价格和削减成本的策略，凸显了与风险投资支持的增长模式不同的世界观。 Mistral 的30亿欧元融资，这是欧洲有史以来最大的一笔科技融资，更多地被视为不是与美国前沿实验室的直接竞争，而是一项旨在实现欧洲“人工智能主权”的战略举措。Rory 将其与空客（Airbus）相提并论，暗示欧洲决心建立自己可行的人工智能参与者，以避免对美国技术的依赖，尤其是在美国模型被限制在欧洲使用的情况下。Jason 指出三星（Samsung）在本轮融资中领投，这表明其估值是战略驱动的，而非纯粹由市场驱动。 其他话题包括 SBF 潜在的最高法院上诉，Jason 认为，由于在罚款和赔偿方面的宪法论点，这可能还有一线希望。Automattic (WordPress 的母公司) 的董事会风波，涉及创始人 Matt Mullenweg，被描述为围绕一个正在缩水的资产而展开的“争吵”，这分散了公司在人工智能时代进行创新的需求。最后，Jason 认为 Adobe 的 CEO 交接是“无足轻重”的事情，标志着现状的持续而非大胆的人工智能驱动的变革，这突显了在需要快速发展的时代保持增长的挑战。主持人总结说，当前的科技格局是前所未有的，领先的人工智能公司与传统科技公司之间存在巨大分歧，处于“列车前端”意味着一切皆有可能，而处于“列车末端”则意味着生活要艰难得多。

This week on 20VC, Rory O'Driscoll and Jason Lampkin discussed the biggest news in tech, covering AI regulation, the burgeoning AI assistant market, significant M&A activity, and major funding rounds. The conversation kicked off with Dario Amodei’s call to "pace the frontier" of AI development, an idea backed by Sam Altman and Elon Musk. Rory expressed skepticism about claims of a "10% chance of human extinction," arguing that if such a risk were real, governments would intervene more aggressively. He pointed out the disparity between the rhetoric of "world-ending" AI and the actual market reaction (a minor dip in semi-stocks and a jump in cyber stocks). Jason viewed Amodei's statement as a strategic move to address risk factors ahead of a potential S1 filing for Anthropic, highlighting the CEO's responsibility to manage such concerns. While acknowledging cyber risk as real, Rory dismissed economic risk claims and found the "loss of control" over AI (recursive self-improvement) as the only truly intriguing and unaddressed danger. Both agreed with David Sacks' sentiment: if a product is that dangerous, the company developing it should either fix it or stop. The discussion also touched on the inevitability of "dark versions" of AI capabilities escaping, likening it to the spread of nuclear technology. Next, the "AI assistant war" took center stage, with Meta launching Muse as a direct competitor to startups like Instinct and Town. Jason lauded Muse's seamless functionality, attributing its speed and efficiency to Meta's vast infrastructure and proprietary Muse LLM. However, he questioned the "killer app" for such assistants, wondering if mere convenience would be enough to drive widespread adoption beyond basic tasks like scheduling or reservations. Rory countered that a "super app" approach, akin to WeChat, offering incremental improvements across many functions, could still be a viable path. The conversation then turned to Instinct's rumored $1 billion funding round at a $10 billion valuation. Jason, playing devil's advocate, presented arguments for and against investing. He noted Meta's limitations (not cross-platform, potential lack of sustained commitment beyond initial hype) as an advantage for Instinct. However, he cautioned against competing with Meta's immense resources and the high infrastructure costs for startups, suggesting such an investment might hinge solely on an M&A outcome. Rory acknowledged the high-risk, high-reward nature of these bets, particularly given the pursuit of generational talent and strategic acquisitions by multi-trillion-dollar companies, though he stressed the rapidly diminishing risk-return profile as valuations soar. Miro's acquisition by Bending Spoons for $1.35 billion, a stark drop from its $17.5 billion valuation, underscored the brutal reality for overvalued unicorns from the 2021 bull market. Rory called it an "inevitable cleanup operation," arguing it was a positive outcome for early investors and founders, while late-stage investors saw less return. Jason used a "Duck, Duck, Goose" analogy, suggesting few "chairs" (acquisition opportunities) are left for pre-AI era companies. Bending Spoons' strategy of ruthless efficiency, increasing prices, and cutting costs highlighted a different worldview from venture-backed growth models. Mistral's €3 billion funding round, Europe's largest ever tech funding, was seen less as a direct competitor to US frontier labs and more as a strategic move for European "AI sovereignty." Rory drew parallels to Airbus, suggesting Europe is determined to build its own viable AI player to avoid reliance on US technology, especially after instances of US models being restricted from European use. Jason noted Samsung's lead in the round, indicating a strategic rather than purely market-driven valuation. Other topics included the potential Supreme Court appeal for SBF, with Jason believing it has a chance due to constitutional arguments regarding the fine and restitution. The board drama at Automattic (WordPress's parent company) involving founder Matt Mullenweg was framed as a "pissing match" over a diminishing asset, distracting from the company's need to innovate in the AI era. Finally, Adobe's CEO transition was deemed a "nothing burger" by Jason, signaling continued status quo rather than bold AI-driven change, highlighting the challenge of maintaining growth in an era demanding rapid evolution. The hosts concluded that the current tech landscape is unprecedented, with extreme divergence between leading-edge AI companies and older tech, where being at the "head of the train" means everything is possible, and being at the "tail end" means life is much harder.