下一个前沿：人工智能协调、社交智能和团队协作的未来

Humans& 获得 4.8 亿美元资金，用于建立人类与人工智能协作的“中枢神经系统”，缩小与社会智能之间的协调差距。

In a move that signals a seismic shift in the artificial intelligence landscape, a stealthy startup named Humans& has recently bagged a jaw-dropping $480 million seed round. Their mission? To construct nothing less than a 'central nervous system' for human-AI collaboration, ushering in an era where AI isn't just smart, but truly socially intelligent. This ambitious endeavor, backed by a dream team of alumni from top AI labs, aims to tackle the thorniest challenge yet: coordinating people with conflicting priorities, tracking complex decisions, and maintaining team alignment over the long haul. It's less about answering questions and more about understanding the room.

一家名为 Humans& 的隐秘初创公司最近获得了令人瞠目结舌的 4.8 亿美元种子轮融资，此举标志着人工智能领域发生了巨大的转变。他们的使命？构建一个人类与人工智能协作的“中枢神经系统”，开创一个人工智能不仅智能，而且真正具有社交智能的时代。这项雄心勃勃的努力得到了顶级人工智能实验室校友梦想团队的支持，旨在解决迄今为止最棘手的挑战：协调优先事项相互冲突的人员、跟踪复杂的决策以及长期保持团队的一致性。与其说是回答问题，不如说是了解房间。

Bridging the 'Coordination Gap' with Social Intelligence

用社交智能弥合“协调差距”

For too long, our sophisticated AI chatbots, while brilliant at crunching numbers or drafting emails, have operated largely as individual virtuosos. They excel in solitary tasks but falter when confronted with the messy, nuanced world of group dynamics – the very essence of 'team collaboration.' Humans& founders argue this 'coordination gap' is the next great frontier for foundation models. Current AI, for all its prowess, simply lacks the 'social intelligence' required to navigate the intricacies of human interaction, consensus-building, and conflict resolution. It's like having a brilliant researcher who can't facilitate a meeting to save their life.

长期以来，我们复杂的人工智能聊天机器人虽然在处理数字或起草电子邮件方面表现出色，但在很大程度上却以个人艺术大师的身份运作。他们在单独任务中表现出色，但在面对混乱、微妙的群体动态世界时却步履蹒跚——这正是“团队协作”的本质。人类和创始人认为这种“协调差距”是基础模型的下一个伟大前沿。当前的人工智能尽管拥有强大的能力，但只是缺乏驾驭人类互动、建立共识和解决冲突等错综复杂的问题所需的“社交智能”。这就像一位才华横溢的研究人员无法召开会议来挽救自己的生命一样。

A New Brain for Team Collaboration

团队协作的新大脑

So, what's the secret sauce? Humans& isn't just tweaking existing models; they're building a fundamentally new architecture. Their approach leverages cutting-edge techniques like long-horizon and multi-agent reinforcement learning. Think of 'long-horizon RL' as teaching an AI to play a strategic game over many moves, not just reacting to the immediate play. It learns to plan, execute, revise, and follow through on long-term goals. 'Multi-agent RL,' on the other hand, trains systems to thrive in environments where multiple AIs and humans are all interacting simultaneously. Crucially, this new breed of AI will have a much deeper, contextual memory – it'll 'remember things about itself, about you,' fostering an understanding that feels less like a bot and more like a seasoned colleague.

那么，秘诀是什么？ Humans& 不仅仅是调整现有模型；他们正在构建一个全新的架构。他们的方法利用了长视野和多智能体强化学习等尖端技术。将“长期强化学习”视为教导人工智能通过多种动作进行战略游戏，而不仅仅是对即时游戏做出反应。它学习计划、执行、修改和贯彻长期目标。另一方面，“多智能体强化学习”训练系统在多个人工智能和人类同时交互的环境中蓬勃发展。至关重要的是，这种新型人工智能将拥有更深层次的情境记忆——它会“记住关于自己、关于你的事情”，从而培养一种理解，感觉不像机器人，而更像经验丰富的同事。

Beyond Bots: Redefining the Collaborative Layer

超越机器人：重新定义协作层

The vision extends beyond a mere application; Humans& aims to 'own the collaboration layer' entirely. Imagine a system that intuitively understands the skills, motivations, and needs of every individual within an organization, from a ten-person startup to a global enterprise, and then deftly balances those factors for collective benefit. It's a bold claim, positioning AI not as a job replacement, but as an unparalleled enhancer of human potential, empowering teams to achieve collective goals with unprecedented efficiency. As LinkedIn founder Reid Hoffman sagely noted, AI truly lives at the workflow level, and Humans& is building the connective tissue to make that a reality.

这一愿景不仅仅局限于应用； Humans& 的目标是完全“拥有协作层”。想象一个系统，它可以直观地了解组织内每个人的技能、动机和需求，从十人初创公司到全球性企业，然后巧妙地平衡这些因素以实现集体利益。这是一个大胆的主张，将人工智能定位为人类潜力的无与伦比的增强器，而不是工作的替代者，使团队能够以前所未有的效率实现集体目标。正如 LinkedIn 创始人 Reid Hoffman 明智地指出的那样，人工智能真正存在于工作流程层面，而 Humans& 正在构建结缔组织以使其成为现实。

The Path Forward: Challenges and Potential

前进之路：挑战与潜力

Naturally, such an audacious undertaking comes with its own set of hurdles. Developing AI that can truly engage in 'social intelligence' – asking questions with genuine understanding, navigating ethical considerations, ensuring data privacy – is no small feat. The company, despite its record-breaking $480 million seed round (one of the largest in AI history), will be competing with tech titans for talent and computational resources. Yet, the investment itself underscores a growing industry consensus: the future of AI isn't just about individual brilliance, but about collective synergy. If Humans& can pull it off, we might just be looking at a future where our digital partners don't just answer our questions, but help us orchestrate our grandest collaborations. It's enough to make you want to high-five your future AI colleague, isn't it?

当然，如此大胆的事业也有其自身的一系列障碍。开发能够真正参与“社交智能”的人工智能——以真正理解的方式提出问题、引导道德考虑、确保数据隐私——绝非易事。尽管该公司在种子轮融资中获得了创纪录的 4.8 亿美元（人工智能历史上规模最大的融资之一），但仍将与科技巨头争夺人才和计算资源。然而，这项投资本身凸显了行业日益增长的共识：人工智能的未来不仅仅取决于个人的才华，还取决于集体的协同作用。如果 Humans& 能够成功，我们可能会看到这样一个未来：我们的数字合作伙伴不仅会回答我们的问题，还会帮助我们精心策划最宏伟的合作。这足以让你想要向你未来的人工智能同事击掌，不是吗？