Agent (AI)
An AI system that takes natural language input and produces natural language output to execute complex multi-step workflows; agents can be chained together to automate business processes (e.g., house search → design → contracting → oversight → payment).
As defined in Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
Where it comes up · 1
Most important claims · 10 of 45Search all claims for “Agent (AI)” →
- AlphaGo's victory over the top Go player in Shanghai spurred China into action and catalyzed their AI investment because Go is central to Chinese culture (comparable to chess in the West), and the defeat was significant enough that the Chinese government restricted televised coverage to prevent citizens from witnessing what they saw as a national loss.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Satya Nadella, upon becoming CEO of Microsoft, implemented three transformative changes: (1) shifted corporate culture from internal competition to external competition orientation and adopted a 'growth mindset,' (2) repositioned the company's focus toward cloud computing and mobile environments rather than relying on legacy personal-computer software dominance, and (3) negotiated and executed the partnership with OpenAI, which was unconventional but strategically important; these changes reversed Microsoft's decade-long stagnation.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Nvidia is the most valuable company in the world with a market capitalization of approximately $4.5 trillion because in 2007 it invented CUDA, a microcode library that enables efficient parallel computation for AI and deep learning; CUDA remains a monopoly lock-in because all major technology stacks are built on top of it, and building AI systems quickly requires using Nvidia chips or its partners, despite enormous efforts to create competing alternatives.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Google had done significant research in language models but did not achieve public recognition equivalent to OpenAI's ChatGPT because Google lacked the key innovation of reinforcement learning from human feedback (RLHF), which ChatGPT pioneered.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- RLF (Reinforcement Learning from Human Feedback) was the critical innovation that allowed ChatGPT to outperform Google's prior AI research; RLF involved humans rating responses (good paragraph vs. bad paragraph), and this human signal during training was the final breakthrough needed for the current generation of AI systems.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Henry Kissinger, who was already 95 years old, was introduced to AI at a European meeting and became immediately impressed with the implications of an intelligence that exceeded human capability; as a philosopher and historian (before becoming a diplomat), he then devoted the bulk of his last five years of life to learning about AI not at a technical level but to think deeply about what it would mean for humanity.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Regardless of how AI training evolves, the end game will require immense infrastructure for deployment—not just chips and data centers, but also electrical generation capacity; the US electrical grid cannot supply the power needed at current AI growth rates, making energy availability and permitting the binding constraint.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- The US is currently slightly ahead of China in AI, but the advantage is narrower and more contestable than many assume; China has been aggressive in algorithm development (which reduces reliance on superior hardware), and recent Chinese announcements have shown surprisingly positive results, making the AI race between the US and China an emerging competitive threat rather than an established US lead.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Schools and educators cannot reliably detect when students use ChatGPT to answer exam questions, and while various detection methods are under development (such as embedding coded markers in AI-generated text), no reliable public-facing solution currently exists, meaning this is an ongoing systemic problem for education.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis
- Whether AI should be used to decide investments is debatable, but in the next decade, human-computer collaboration will be essential because AI systems still get confused, and human experts will need to partner with AI to provide oversight and judgment; this human-in-the-loop model will remain the best approach short-term.Eric Schmidt and Craig Mundie with David Rubenstein: Genesis