Google CEO Page Backs Radical Moonshots to Evade Corporate Decline

Where's Google going next? | Larry Page — TED

TED

Page argues that companies fail when they cling to core businesses instead of anticipating shifts, contending that deploying corporate capital toward untended high-risk problems achieves more societal progress than traditional philanthropy.

AI-assisted summary of Where's Google going next? | Larry Page — original source. Check the original for full context.

Source published:

Key ideas

  • Google CEO Larry Page views personal computing as still fundamentally awkward and clunky, pushing Google toward deep machine learning and context-aware interfaces to truly understand human intent. (0:13)
  • Following the acquisition of DeepMind, Page highlights unsupervised reinforcement learning systems that master arcade games and discover visual concepts like cats purely from raw data without human labeling. (2:29)
  • Project Loon aims to bridge the global digital divide by deploying high-altitude stratospheric balloons navigated via computational wind models to connect the two-thirds of humanity without reliable internet. (8:43)
  • Page critiques secret government surveillance as an existential threat to democracy, while simultaneously urging society to overcome privacy paranoia so anonymized health data can save an estimated 100,000 lives annually. (11:07)
  • Through the lens of 'additionality,' Page defends ambitious corporations as primary engines of societal progress, arguing that companies decline because they miss the future rather than failing at present execution. (17:48)

Detailed notes

## Background This interview was published by the TED channel on YouTube on March 22, 2014. According to the page description, Charlie Rose interviews Google CEO Larry Page onstage at TED2014 about his far-off vision for the company, spanning aerial bikeways, internet-delivering balloons, and the recent acquisition of artificial intelligence lab DeepMind.

## TL;DR Google CEO Larry Page joins Charlie Rose to discuss the evolving mission of Google beyond traditional search, emphasizing the push toward ambient, context-aware computing and unsupervised artificial intelligence. Page highlights audacious moonshots designed to expand global connectivity and reshape urban infrastructure, including Project Loon's high-altitude stratospheric balloons and autonomous driving technology. Addressing contemporary controversies around surveillance and digital privacy, he advocates for transparent governance, arguing that societal overcaution prevents transformative breakthroughs like anonymized health research. Throughout the conversation, Page champions the economic concept of 'additionality' and defends mission-driven corporations as powerful, essential engines for long-term philanthropic and technological progress.

## Atmosphere & style The conversation carries an intimate, cerebral, and contemplative tone, set against the warm staging of TED with an engaged live audience. Larry Page speaks in a gentle, raspy cadence, balancing self-deprecating humor with quiet, unyielding conviction about high-stakes technological futures. Charlie Rose acts as an inquisitive, conversational prober, nudging Page across technical demonstrations, personal health vulnerabilities, and macro-philosophical reflections on corporate mortality. The dynamic feels less like an executive press briefing and more like a rare, unguarded window into the radical mental models of a visionary builder.

## Bulletin - [00:13](https://www.youtube.com/watch?v=mArrNRWQEso&t=13s) Google CEO Larry Page views personal computing as still fundamentally awkward and clunky, pushing Google toward deep machine learning and context-aware interfaces to truly understand human intent. - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) Following the acquisition of DeepMind, Page highlights unsupervised reinforcement learning systems that master arcade games and discover visual concepts like cats purely from raw data without human labeling. - [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) Project Loon aims to bridge the global digital divide by deploying high-altitude stratospheric balloons navigated via computational wind models to connect the two-thirds of humanity without reliable internet. - [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) Page critiques secret government surveillance as an existential threat to democracy, while simultaneously urging society to overcome privacy paranoia so anonymized health data can save an estimated 100,000 lives annually. - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Through the lens of 'additionality,' Page defends ambitious corporations as primary engines of societal progress, arguing that companies decline because they miss the future rather than failing at present execution.

## Key takeaways - [00:13](https://www.youtube.com/watch?v=mArrNRWQEso&t=13s) Fifteen years after Google's founding, organizing the world's information remains an unfinished challenge hampered by computers that lack context about user location, activity, and knowledge. - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) Google turned to unsupervised machine learning after traditional speech recognition hit a wall, proving that neural models could independently learn complex visual concepts without training labels. - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) DeepMind's unified algorithm achieved superhuman video game performance using only raw screen pixels, controls, and score feedback, demonstrating the potential of combining computer science with cognitive neuroscience. - [06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) Kenyan farmer Zack Matere's story illustrates that democratized internet access enables individuals to solve critical local problems, which Matere bridged offline through a handwritten village notice board. - [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) Two out of three people globally still lack reliable internet connectivity, driving Google's Project Loon to deploy stratospheric balloons steered by dynamic altitude adjustments along atmospheric wind currents. - [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) Mass government surveillance conducted in total secrecy undermines democratic governance by forcing technology companies to defend users against covert state programs that never received public debate. - [12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) An overemphasis on medical privacy prevents research into rare and untreated conditions; anonymizing patient records with transparent access audit logs could save an estimated 100,000 lives each year. - [15:12](https://www.youtube.com/watch?v=mArrNRWQEso&t=912s) Autonomous vehicles address a massive global health crisis—with over 20 million road injuries annually—while offering the potential to reclaim vast urban areas currently consumed by parking and roads. - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Nikola Tesla illustrates the limitation of pure inventive genius without commercial and distribution infrastructure, reinforcing the need to pair revolutionary technology with effective business execution. - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Corporations should be evaluated as viable vehicles for philanthropy when they pursue 'additionality'—investing substantial talent and capital into transformative, neglected problems that market incentives otherwise ignore. - [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) Dominant companies primarily fail and accelerate corporate turnover because they miss technological shifts in the future rather than executing poorly on their existing business lines. - [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) Page initially felt guilty about devoting time to Android as an acquired startup outside Google's search core, an experience that reinforced the necessity of prioritizing risky, future-defining platforms.

## Detailed walkthrough ### [00:13](https://www.youtube.com/watch?v=mArrNRWQEso&t=13s) Google’s Evolving Mission and the Limits of Clunky Computing

Charlie Rose opens the conversation by revealing a self-deprecating note Larry Page had sent him prior to taking the stage: Page wrote that they needed to ensure they did not appear like “a couple of middle-aged boring men.” Rose jokes back that he was deeply flattered by the comparison, pointing out that he is quite a bit older than Page, while Page possesses considerably more net worth. Rose sets the agenda for the session: an exploration of the internet, search, privacy, Page’s overarching philosophy, and where Google is heading next.

When Rose asks Page where Google currently stands and where it is going, Page reflects on the company’s original founding mission—to organize the world’s information and make it universally accessible and useful. Page admits that people often ask whether that is still Google’s genuine mission, and he frequently asks himself the same question. Ultimately, he sees search as a profound, unfinished challenge. “We've been at it for 15 years already, but it's not at all done,” Page explains, marveling at how society is still in the earliest days of understanding human intent and global knowledge. Asked what the destination looks like when it is finally done, Page points out that the fundamental roadblock is that personal computing remains awkward and unintegrated:

> “You know, your computer doesn't know where you are, it doesn't know what you're doing, it doesn't know what you know, and a lot we've been trying to do recently is just make your devices work, make them understand your context.”

He notes that efforts like Google Now are designed to anticipate what a user needs based on location and schedule, but confesses that modern computing across the board remains far too clunky and disconnected from actual human behavior.

---

[02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) DeepMind, Unsupervised Learning, and the Path to AI

Rose asks how Google's recent acquisition of DeepMind fits into this broader technological picture. Page explains that the U.K.-based company aligns directly with Google's efforts to move past clunky interfaces. To make computers truly understand speech, Google discovered that existing speech recognition technology was inadequate, prompting them to invest heavily in machine learning research. As they tested machine learning across vast datasets, such as YouTube videos, an unsupervised model famously taught itself the visual concept of a cat completely unassisted, without any labeled training data.

Page shows a video demonstrating DeepMind’s system mastering classic video games without human instructions. Given only the raw screen pixels, the game controls, and the score counter, a single, unified program learned to play multiple games with superhuman skill. Demonstrating a boxing game where the computer takes the left-hand boxer and pins the opponent against the wall to rack up points, Page imagines what will happen when this capability is directed toward managing human schedules and complex information needs.

```mermaid flowchart LR accTitle: DeepMind Unsupervised Learning Architecture accDescr: Diagram showing how raw visual inputs and score feedback train a unified algorithm to take actions and achieve superhuman game performance.

subgraph Inputs["Raw Inputs"] P["Screen Pixels"] S["Game Score"] end

subgraph Agent["DeepMind Unified Program"] ML["Deep Reinforcement<br/>Learning Algorithm"] end

subgraph Output["Autonomous Action"] C["Game Controller Inputs"] P2["Superhuman Gameplay<br/>(e.g., Boxing Pin Strategy)"] end

P --> ML S --> ML ML --> C C --> P2 ```

Discussing the leadership behind DeepMind, Page highlights founder Demis Hassabis, who earned a Ph.D. in cognitive neuroscience after studying computer science in order to map how the human brain functions. Page sees immense promise at the intersection of neuroscience and artificial intelligence. When Rose presents the machine-generated visual composite of a cat on the stage screen, Page stresses why this milestone was so meaningful:

> “And there's no training, no notion of a cat, but this concept of a cat is something important that you would understand, and now that the machines can kind of understand.”

- **Acquisition background**: DeepMind is based in the United Kingdom and founded by Demis Hassabis, who combined computer science with a Ph.D. in neuroscience. - **Unsupervised discovery**: The system identified cats on YouTube purely from raw video feeds without prior labels or human guidance. - **Unified gaming system**: The exact same algorithm tackled different arcade games using only pixels and score feedback to develop optimal, superhuman strategies.

---

[06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) Access to Information: The Story of Zack Matere

Page transitions back to the core power of search with a video feature Google discovered in the news. The video documents Zack Matere, a farmer living in Soy, Kenya. When Matere planted a crop of potatoes, the plants began withering and dying without apparent reason, and local agricultural books provided no answers. Seeking an explanation, Matere traveled to an internet cafe and searched for “potato diseases.” He discovered an online article suggesting ants could be attacking the roots, recommending that he sprinkle wood ash over the crops. Within days of applying the ash, the ants vanished and his harvest was saved.

Inspired by the power of the web, Matere returned to the cyber cafe with an entrepreneurial friend to research renewable energy designs, which led the friend to construct a working windmill at the local school. Realizing that the majority of his community—including his grandmother—lacked the means or literacy to browse the web, Matere devised an offline solution: a simple wooden community notice board. Whenever he retrieves critical health, agricultural, or educational information on his mobile phone, he hand-writes the findings onto the board for the whole village to read. In the clip, Matere reflects on the transformative potential of democratized knowledge:

> “When people have the knowledge, they can find solutions without having to helped out. Information is powerful, but it is how we use it that will define us.”

Page remarks on how finding this gentleman's story underscored Google's responsibility to deliver information directly into communities that lack standard infrastructure.

---

[08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) Project Loon and Worldwide Connectivity

Rose points out that Page has long been driven by the conviction that technology must show the way to world-changing improvements, especially through universal connectivity. Page agrees, noting that if one looks at the future, two out of three people on Earth still lack reliable internet access. To overcome the high cost and slow deployment timelines of terrestrial cables and orbital satellites, Google launched Project Loon.

The initiative uses high-altitude balloons floating in the stratosphere to provide cost-efficient wireless internet coverage. Page recalls that Google had researched the concept for over five years before formal development began. While searching online for precedents, Page discovered historical records showing that high-altitude research balloons had successfully circumnavigated the globe multiple times three to four decades earlier. Google used complex computational weather simulations to solve the navigation challenge: by pumping air into or out of the balloon envelope to alter altitude, the balloons can catch distinct stratospheric wind currents heading in different directions. Page emphasizes that this dynamic altitude control enables Google to steer the fleet and build a resilient planetary communication network:

> “You can actually control roughly where they go, and so I think we can build a worldwide mesh of these balloons that can cover the whole planet.”

As the conversation turns from internet infrastructure toward future technologies, Rose notes Page's long-standing passion for automated cars, bicycles, and transportation, before pivoting to a question about global surveillance, digital security, and Edward Snowden—referencing a photo taken the previous day at the conference showing Google co-founder Sergey Brin alongside Snowden. Page acknowledges that security and privacy have become unavoidable, top-of-mind concerns.

[11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) Government Surveillance and Defending Democracy

Addressing the fallout from Edward Snowden’s disclosures, Larry Page draws a sharp boundary between user privacy and systemic security, emphasizing that privacy is impossible to safeguard without robust underlying security. For Page, the revelations about mass government surveillance were deeply frustrating not because the government seeks to counter terrorism, but because the surveillance apparatus was established in total secrecy without democratic authorization or public discourse. He makes clear that Google should never have been forced into the adversarial role of shielding its own users against covert state surveillance programs that were never put to public debate.

Page clarifies that intelligence agencies do not need to reveal operational specifics or sensitive terrorist threats, but the general parameters, mechanics, and boundaries of surveillance must be publicly defined. When Charlie Rose suggests that the government erred by never coming to Google directly, Page firmly corrects him, pointing out that the fundamental breach was not with tech companies, but with citizens. As Page puts it: “I don’t think we can have a democracy if we’re having to protect you and our users from the government for stuff that we’ve never had a conversation about.” By keeping these programs cloaked in secrecy, he argues, the state has actively undermined democratic governance and done itself a tremendous disservice.

- Page emphasizes that Google's responsibility is protecting users, but insists that doing so against covert government overreach represents an untenable institutional position. - The core requirement for surveillance oversight is public awareness of the scope, methods, and justifications, rather than specific classified targets.

---

[12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) User Privacy and the Untapped Potential of Anonymized Health Data

Turning to consumer privacy, Page acknowledges that mobile smartphones collect unprecedented quantities of personal information, from real-time location to search histories. He views user transparency and control as Google’s primary obligations, citing tools like Google Chrome’s Incognito mode as steps toward giving individuals granular visibility and autonomy over their footprint. However, Page cautions against an overcorrection toward absolute secrecy, warning that society risks throwing out the baby with the bathwater by ignoring the enormous public good that comes from shared information.

To illustrate this, Page shares an anecdote about his own vocal cord paralysis, which initially developed around the time he previously appeared on Charlie Rose’s show. Poking fun at his raspy voice, Page jokes that Rose should pull out a voodoo doll if it might help him recover it. Page admits he was initially terrified to share his medical struggles, but co-founder Sergey Brin urged him to go public. Once he disclosed his condition, Google conducted an informal survey and gathered data from thousands of people with similar unstudied conditions.

```mermaid flowchart TD accTitle: Anonymized Health Records Research Model accDescr: Flow showing how anonymized medical records pass to researchers with audit transparency to save lives.

subgraph Data["Patient Data"] P["Individual Medical Records"] -->|"anonymize & aggregate"| A["Shared Research Database"] end

subgraph Access["Scientific Research"] A --> R["Research Doctors & Scientists"] R -->|"study conditions"| S["Life-Saving Treatments"] end

subgraph Audit["Patient Transparency"] R -->|"audit log: who & why"| T["Patient Notification Log"] T -->|"personal condition insights"| P end ```

Page connects this experience to the broader healthcare system, arguing that privacy paranoia prevents lifesaving discoveries: “I think if we just did that, we’d save 100,000 lives this year.” He believes that if patient records were made anonymously available to medical researchers—paired with an audit log showing patients exactly which doctor viewed their data and why—it would revolutionize modern medicine.

- Medical data isolation severely restricts research into rare and undertreated conditions due to a lack of shared clinical datasets. - Page's public disclosure generated responses from thousands of individuals whose condition had previously lacked systematic data. - User trust in data sharing relies on two pillars: transparency regarding who accesses the data and verifiable assurances that the information will not be abused.

---

[15:12](https://www.youtube.com/watch?v=mArrNRWQEso&t=912s) Autonomous Vehicles and Reclaiming Urban Space

Page’s fixation on transportation began during his undergraduate years at the University of Michigan, where he spent freezing winters waiting for campus buses and digging into the municipal economics of public transit systems. That early frustration sparked an 18-year obsession with autonomous transit. Moving people through automated vehicles, Page explains, addresses a massive public health crisis: automobile accidents injure more than 20 million people worldwide each year and stand as the leading cause of death for Americans under age 34.

Beyond preserving human life, automated transport reimagines urban architecture. Page observes that modern cities dedicate excessive physical footprints to vehicle storage and roadways: “Los Angeles is half parking lots and roads, half of the area, and most cities are not far behind, actually.” Freeing urban centers from parked and idling cars would unlock land for housing and open community spaces. Asked by Rose how near this reality is, Page notes that Google's autonomous fleet had already logged well over 100,000 miles entirely driverless, indicating that deployment could arrive very quickly.

Rose then plays a video illustrating an alternative transit concept that caught Page’s curiosity: elevated bike tracks built above street grids to segregate human-powered travel from motorized traffic at minimal infrastructure cost. While Page clarifies that Google is not actively developing elevated bikeways, the company provides thousands of free campus bicycles that employees ride around the clock until the components wear out. Page continuously studies ideas that encourage cycling and separate bikes from vehicle flow.

- Automobile accidents cause over 20 million injuries annually worldwide and represent the top cause of mortality for people under 34 in the United States. - Land use in major metropolitan centers like Los Angeles is roughly 50 percent consumed by roadways and surface parking lots. - Google’s autonomous driving initiative had safely completed more than 100,000 miles in full autonomous mode by early 2014.

---

[17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Additionality, Nikola Tesla, and Corporations as Agents of Change

Rose asks Page to describe the operating philosophy behind Google X and his ambition to tackle grand moonshots instead of incremental optimizations. Page invokes the economic principle of “additionality”—investing organizational resources solely into problems that would never be solved unless Google directly pursued them. Broadening technological capabilities expands the visible horizon of what is achievable, whether discovering specific balloon materials that unlocked Project Loon to connect offline populations, or developing advanced artificial intelligence.

Page argues that pure scientific discovery is insufficient on its own, citing the historical tragedy of Nikola Tesla. Tesla invented the alternating current power grid that powers modern society, yet lacked the organizational and commercial capability to distribute it, spending decades struggling while others brought his breakthroughs to market. Page argues that real societal progress requires pairing raw inventive genius with world-class commercial execution:

> “We need, especially in technology, we need revolutionary change, not incremental change.”

This imperative informs Page’s defense of corporations as engines for social progress. Dismayed that many people reflexively view corporations as evil, Page concedes that companies often earn that reputation by pursuing tired, 50-year-old incremental playbooks. Yet he contends that well-run commercial enterprises possess the speed, talent, and resources needed to solve humanity's hardest problems.

Page references his past remark that he would rather bequeath his wealth to an entrepreneur like Elon Musk than to traditional philanthropic foundations. Musk’s ambition to establish human life on Mars to back up humanity constitutes a deeply philanthropic mission, even though it operates as a private company. Page observes that many Google engineers and conference attendees have amassed extraordinary fortunes, yet continue working around the clock to improve the world. He poses a fundamental question: if companies command most of society's best talent, time, and financial capital, society must reconsider corporations as legitimate vehicles for philanthropy and systemic improvement. As the session nears its end, Rose sets up a final question regarding which specific mental habit or quality of mind has served Page best throughout his career.

- Additionality drives Google X: pursuing radical technical projects that no existing commercial actor or market incentive would otherwise attempt. - Project Loon provides an example of technical viability following material science breakthroughs, offering internet access to the two-thirds of the globe lacking reliable connectivity. - Nikola Tesla serves as Page’s cautionary tale about the limitations of invention without the commercial infrastructure needed for global distribution. - Page argues for treating mission-driven corporations as philanthropic vehicles, suggesting that leaving capital to audacious builders like Elon Musk can yield a greater global return than conventional non-profit giving.

[21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) Anticipating the Future and the Value of Additionality

As the conversation draws to a close, Charlie Rose invokes Bill Gates and Warren Buffett, noting that both have famously cited "focus" as the ultimate driver of success. Rose asks Larry Page what specific quality of mind has enabled him to simultaneously reimagine the distant future and transform the present. Page reflects on his study of corporate history, observing the increasingly rapid turnover of once-dominant companies and asking himself what they fundamentally did wrong. In his view, the answer is straightforward: “And usually it's just that they missed the future.” Rather than obsessing solely on existing businesses, Page explains that his core imperative is determining what the future will genuinely look like, how to build it, and how to motivate a large organization to pursue that trajectory with intense urgency.

Page credits this mindset to relentless curiosity, examining spaces other people overlook, and committing to areas where nobody else is currently operating. This approach produces genuine "additionality"—making a contribution that would not have happened otherwise—which requires a willingness to take real risks. To illustrate how counterintuitive this can feel inside a running company, Page recalls the early days of mobile development: “I felt guilty about working on Android when it was starting.” Because Android was originally just a tiny acquired startup rather than Google's core search product, he initially felt he was misallocating valuable time. In hindsight, Page dismisses that hesitation bluntly, noting, “That was stupid. That was the future, right? That was a good thing to be working on.” Rose concludes the interview by thanking Page for joining him onstage, prompting an enthusiastic round of applause from the audience.

```mermaid flowchart TD accTitle: Larry Page's Framework for Long-Term Company Survival accDescr: Diagram illustrating how corporate failure from missing the future is countered by curiosity, additionality, and long-term bets like Android.

subgraph FailureMode ["Common Corporate Trap"] A["Focus solely on present core businesses"] --> B["Miss structural technological shifts"] B --> C["Corporate stagnation and turnover"] end

subgraph FutureMindset ["Larry Page's Approach"] D["Lead with curiosity and risk-taking"] --> E["Pursue 'additionality' by tackling untended spaces"] E --> F["Drive organizational execution at high speed"] F --> G["Build the future (e.g., early bet on Android)"] end ```

- **Corporate Mortality**: Page observes that the turnover rate of major corporations is accelerating primarily because leadership teams fail to anticipate and build toward the future. - **The Guilt of Emerging Bets**: When Google first acquired Android, Page felt internal guilt for dedicating executive attention to a small outside project rather than the established core business, only to realize later that mobile was the entire future of computing. - **Additionality**: True impact comes from deliberately deploying engineering talent and resources toward difficult problems that other technology companies are actively ignoring.

## Notable quotes > [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) I don’t think we can have a democracy if we’re having to protect you and our users from the government for stuff that we’ve never had a conversation about. — Larry Page

> [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) We need, especially in technology, we need revolutionary change, not incremental change. — Larry Page

> [06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) Information is powerful, but it is how we use it that will define us. — Zack Matere

## Mentioned Android, Bill Gates, Charlie Rose, Demis Hassabis, DeepMind, Edward Snowden, Elon Musk, Google, Google Chrome, Google Now, Google X, Larry Page, Nikola Tesla, Project Loon, Sergey Brin, TED, University of Michigan, Warren Buffett, YouTube, Zack Matere

## Video description Onstage at TED2014, Charlie Rose interviews Google CEO Larry Page about his far-off vision for the company. It includes aerial bikeways and internet balloons ... and then it gets even more interesting, as Page talks through the company's recent acquisition of Deep Mind, an AI that is learning some surprising things.

TEDTalks is a daily video podcast of the best talks and performances from the TED Conference, where the world's leading thinkers and doers give the talk of their lives in 18 minutes (or less). Look for talks on Technology, Entertainment and Design -- plus science, business, global issues, the arts and much more. Find closed captions and translated subtitles in many languages at http://www.ted.com/translate

Follow TED news on Twitter: http://www.twitter.com/tednews Like TED on Facebook: https://www.facebook.com/TED

Subscribe to our channel: http://www.youtube.com/user/TEDtalksDirector

[Source](https://www.youtube.com/watch?v=mArrNRWQEso)

中文简报

谷歌CEO佩奇力挺激进登月项目以规避企业衰退 佩奇认为，企业因固守核心业务而非预见变革走向失败，并坚称将企业资本投入无人涉足的高风险问题，比传统慈善能带来更多社会进步。 谷歌CEO拉里·佩奇在与查理·罗斯对话时探讨了谷歌超越传统搜索的使命演变，强调了推动环境感知、情境感知计算以及无监督人工智能的发展。佩奇重点介绍了旨在扩大全球互联互通并重塑城市基础设施的大胆登月项目，包括Project Loon的高空平流层气球和自动驾驶技术。在谈及当代围绕监控与数字隐私的争议时，他呼吁透明化治理，指出社会层面的过度谨慎阻碍了如匿名健康研究等变革性突破。在整场谈话中，佩奇力挺“额外性”这一经济概念，并坚称使命驱动型企业是推动长期慈善与技术进步的强大且不可或缺的引擎。 谷歌CEO拉里·佩奇认为个人计算目前仍存在根本上的笨拙与滞后，促使谷歌转向深度机器学习与情境感知界面，以真正理解人类意图。 (0:13) 在收购DeepMind之后，佩奇重点介绍了无监督强化学习系统，该系统能在无需人工标注的情况下，完全依据原始数据精通街机游戏并发现“猫”等视觉概念。 (2:29) Project Loon旨在通过利用计算风力模型导航的高空平流层气球，来弥合全球数字鸿沟，从而为缺乏可靠网络的全球三分之二人口建立连接。 (8:43) 佩奇批评秘密政府监控是对民主的生存威胁，同时敦促社会克服对隐私的偏执，以便匿名健康数据每年能够挽救约10万人的生命。 (11:07) 佩奇从“额外性”的角度辩护称，雄心勃勃的企业是社会进步的核心引擎，并认为企业走向衰落是因为错失了未来，而非未能做好当下的执行。 (17:48)

Reading summary

In a conversation with Charlie Rose, Google co-founder Larry Page argues that technology companies must venture beyond incremental search improvements to tackle foundational global problems, from artificial intelligence and connectivity to public health and urban transit. Page contends that long-term societal progress depends on applying the economic principle of additionality—marshaling corporate capital and technical talent into ambitious moonshots that market forces would otherwise overlook. From Clunky Computing to Unsupervised Machine Learning (0:13) Page reflects that although Google has organized global information for fifteen years, personal computing remains fundamentally disconnected from human behavior because devices fail to understand context, location, and user intent. To bridge this divide, Google invested heavily in machine learning, demonstrating that an unsupervised neural network could autonomously identify the visual concept of a cat from unlabeled YouTube video feeds without human instruction. Google's acquisition of U.K.-based DeepMind, founded by neuroscientist Demis Hassabis, accelerated this agenda. DeepMind developed a unified reinforcement learning algorithm that mastered arcade video games with superhuman capability relying solely on raw screen pixels and score feedback, demonstrating strategies like cornering opponents in a boxing simulation. Page envisions translating this capability from games into managing complex human schedules and daily information retrieval. Democratizing Access through Search and Stratospheric Balloons (6:20) Highlighting the practical value of information, Page points to Zack Matere, a farmer in Soy, Kenya, who searched an internet cafe database to diagnose a potato crop disease, eradicated root-eating ants using wood ash, and subsequently hand-copied online health and farming guidance onto a village notice board. Because two-thirds of the global population still lacked reliable internet access, Google initiated Project Loon to deploy stratospheric balloons as cost-effective alternatives to terrestrial cables and satellites. Drawing on historical flight records from several decades prior, Google paired high-altitude research balloons with computational weather models. By pumping air into and out of balloon envelopes to shift altitude and catch differing wind currents, engineers achieved dynamic directional control. Page explains that this approach allows Google to navigate a coordinated planetary mesh network to deliver connectivity to underserved communities. Democratic Surveillance Boundaries and Anonymized Health Data (11:07) Following the disclosures by Edward Snowden, Page emphasizes that user privacy cannot exist without underlying digital security. While acknowledging that governments must counter security threats, he criticizes the state for conducting bulk surveillance in secrecy without democratic mandate. Page insists Google should never be positioned as an adversarial shield against covert state programs, arguing that democracies require open public discourse regarding the scope and mechanics of government surveillance, even if operational targets remain confidential. At the same time, Page warns against excessive privacy safeguards that impede scientific advancement. Drawing from his own diagnosis of vocal cord paralysis—which generated thousands of patient reports once publicly acknowledged—Page argues that anonymizing patient health records for medical researchers, supported by audit logs displaying which doctors access the data, could save 100,000 lives annually by accelerating treatments for neglected conditions. Urban Redesign and Autonomous Mobility (15:12) Page traces his interest in automated transit to his undergraduate years studying the municipal economics of public transit at the University of Michigan. He highlights that traffic collisions cause more than 20 million injuries each year globally and represent the primary cause of mortality for Americans under 34. Autonomous transport offers a direct safety intervention while liberating metropolitan real estate, noting that roughly half of the surface area in cities like Los Angeles is consumed by roadways and parking lots. By early 2014, Google's autonomous vehicle fleet had completed over 100,000 driverless miles. Alongside self-driving cars, Page studies alternative methods to separate human-powered transit from motorized vehicles, including elevated cycling track concepts and Google's campus-wide bicycle infrastructure, aiming to reclaim paved land for residential housing and community spaces. Corporate Additionality and Avoiding Organizational Mortality (17:48) Page defends corporations as viable engines for broad societal progress when guided by 'additionality'—the principle of allocating resources exclusively to difficult, high-impact problems that would not otherwise be pursued. Citing Nikola Tesla's struggle to commercialize alternating current despite groundbreaking inventive genius, Page stresses that revolutionary technological change requires pairing breakthrough science with corporate scale. Under this framework, Page argues that mission-oriented corporations, such as Elon Musk's initiatives to colonize Mars, can deliver greater philanthropic impact than traditional non-profit foundations. Examining why once-dominant companies frequently collapse, Page concludes that organizational mortality occurs because leadership teams focus exclusively on existing revenue streams and miss fundamental market shifts. He recalls feeling guilty during Google's early work on Android, viewing it as a distraction from the core search business, before realizing mobile was the future of computing. He maintains that corporate longevity requires sustained curiosity, calculated risk-taking, and high-urgency bets on neglected spaces. Page's overarching vision relies on the presumption that large corporations will exercise their capital and technical capabilities for broad public benefit. His model leaves unresolved how societies should govern commercial moonshots that reshape vital public goods—from internet access and healthcare data to urban infrastructure—without relying on the personal discretion of tech executives.

精炼阅读摘要

Google 联合创始人 Larry Page 在与 Charlie Rose 的对谈中指出，计算设备与人类真实需求之间仍存在脱节，技术公司必须突破渐进式改良，通过追求具备“附加值”的根本性创新来重塑未来。 超越笨重计算：非监督学习与人机交互 (0:13) Page 认为，Google 组织全球信息的使命远未完成，核心障碍在于现代计算设备依然笨拙，无法感知用户的具体环境、位置和正在进行的操作。为突破这一瓶颈，Google 持续加大对机器学习的投入。在收购英国人工智能公司 DeepMind 后，Google 探索了基于神经科学的无监督学习模型。在未经过任何预先标注或人工指导的情况下，算法仅凭海量 YouTube 视频数据便自主形成了对“猫”的视觉概念；DeepMind 的统一算法仅凭借屏幕像素和得分反馈，就能在多款经典街机游戏中掌握超人水准的策略（例如在拳击游戏中将对手逼入墙角）。Page 预计，当这类具备自主学习与推理能力的系统被应用于人类日常日程管理与复杂信息处理时，计算机将真正理解用户的真实意图。 弥合数字鸿沟：从偏远乡村到平流层气球 (6:20) 信息获取的平民化对缺乏传统基础设施的地区具有重大现实意义。肯尼亚农户 Zack Matere 在作物受病害侵袭且当地书籍无法给出答案时，前往网吧检索信息，凭借网上“木灰驱蚁”的建议挽救了马铃薯收成，随后又借助网络资讯与朋友为当地学校建造了风车，并设立线下木制公告板向全村转抄健康与农业知识。Page 指出，全球仍有三分之二的人口缺乏可靠的网络连接。为了克服铺设地面光缆与发射轨道卫星的高昂成本和漫长周期，Google 推出了 Project Loon。该项目利用材料科学突破，在平流层部署高空气球，通过计算气象模拟控制气球充气与放气以改变飞行高度，借由不同方向的平流层风流实现编队导航，旨在构建覆盖全球的低成本无线通信网络。 监控争议与数据共享：在安全与隐私之间寻求平衡 (11:07) 在回应 Edward Snowden 披露的政府大规模监控事件时，Page 强调强有力的底层安全机制是保护用户隐私的基石。他表达了对秘密监控体系的强烈不满，明确指出情报机构固然无需公开具体战术行动，但监控的总体范围、机制和法律边界必须公开讨论并获得民主授权；迫使科技公司在缺乏公众辩论的情况下保护用户免受政府侵犯，是对民主机制的削弱。针对商业领域的数据隐私，Page 认为移动设备收集了大量敏感数据，企业必须通过诸如隐身模式等工具向用户提供透明度和控制权。但他也警告，过度防范数据共享会导致巨大的社会损失。Page 以自己公开声带麻痹病症从而收集到数千名罕见病患者数据的经历为例主张，如果能在确保患者知情（包括明确记录哪些医生查阅了记录）的前提下，将匿名化医疗数据开放给科研人员，仅在当年就可能挽救 10 万人的生命。 重构城市空间：自动驾驶的公共健康价值 (15:12) Page 对交通运输的关注始于其在密歇根大学就读时对寒冬等候班车及公交系统市政经济学的观察。他指出，传统汽车交通已构成严重的全球公共健康危机：全球每年因车祸受伤的人数超过 2000 万人，且交通事故是美国 34 岁以下人群的首要死因。此外，现代都市空间被传统车辆过度侵占，例如洛杉矶有近一半的土地被道路和地面停车场占据，其他主要城市也面临类似困境。自动驾驶技术的普及不仅能显著降低死伤率，还能释放大量昂贵用地以用于住宅与公共社区建设。根据 Page 提供的数据，截至 2014 年初，Google 的自动驾驶车队已经在完全无人驾驶模式下安全行驶了超过 10 万英里。除了自动化车辆，Page 亦长期研究能够将人力骑行与机动车流彻底物理隔离的高架自行车道等替代方案。 “附加值”哲学与作为变革主体的商业机构 (17:48) Page 阐述了 Google X 的核心运作原则——经济学中的“附加值”（Additionality），即企业的资源应当专门投向那些若无自身介入就绝不会被他人解决的根本性难题，而非在成熟业务上进行平庸的边际优化。他以 Nikola Tesla 发明交流电系统却因缺乏商业与组织推广能力而陷入长期困顿的历史为鉴，强调科学技术必须与世界级的商业执行力相结合才能推动颠覆性变革。Page 认为，许多人对商业机构持有负面成见，这主要是因为大量企业固守陈旧的渐进式策略；然而，运行良好的商业机构聚集了社会中最优秀的工程人才和雄厚资本，完全可以成为推动社会进步的强效载体。Page 甚至表示，相比传统慈善基金会，他更愿意将个人财富托付给类似 Elon Musk 这样致力于在外太空建立人类定居点、追求宏大愿景的创业者。 Page 认为，历史上主流大企业的衰亡多源于对技术未来的误判与错失，他以自己早期曾因在 Android 项目上分配时间而产生内疚感为例，反思了将精力局限于成熟核心业务的局限性；但该观点未深入探讨在追求高风险、颠覆性创新的同时，企业应如何平衡当期商业回报与公共伦理监管之间的潜在矛盾。

Detailed notes — zh

## Background 该视频由 TED 官方频道于 2014 年 3 月 22 日发布在 YouTube 平台上。内容记录了在 TED2014 大会舞台上，著名主持人 Charlie Rose 深度专访 Google 联合创始人兼首席执行官 Larry Page，围绕 Google 的未来远景展开探讨，话题涵盖平流层通信气球、空中悬挂自行车道以及当时最新收购的人工智能团队 DeepMind。 ## TL;DR Larry Page 在访谈中回顾了 Google 整合全球信息的初心，指出计算机距离真正理解人类情境仍有巨大差距，并重点展示了 DeepMind 在无监督学习和决策智能上的突破。他分享了旨在为全球未联网人群提供平价网络的 Project Loon 高空气球计划，以及自动驾驶技术重塑城市空间与交通安全的潜力。针对棱镜门引发的信任危机，Page 犀利批判了政府秘密监控对民主的伤害，同时倡导在提供透明审计的前提下推进医疗数据匿名共享以拯救生命。最后，他阐释了 Google X 的“额外性”（Additionality）原则，强调企业衰亡往往源于错失未来，必须将远大发明与商业化落地结合，敢于承担风险布局冷门赛道以推动革命性变革。 ## Atmosphere & style 整场访谈在轻松幽默的风趣调侃中开场，随后迅速切入宏大深邃的技术与哲学思辨，兼具前瞻性与真诚感。Larry Page 以其特有略带沙哑但极其沉稳的语调娓娓道来，摒弃了商业领袖常见的宣讲套路，展现出一种质朴、克制却极具野心的极客理想主义。Charlie Rose 的提问兼具温度与锐度，二人在从国家安全监控的严肃反思平滑切换至高空气球与空中自行车道等奇思妙想时，碰撞出一种脚踏实地却敢于仰望星空的思辨张力。 ## Bulletin - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) Larry Page 指出计算机在理解情境方面仍处于初级阶段，并展示了 DeepMind AI 仅凭原始像素和得分机制便在雅达利经典街机游戏中实现超越人类的表现。 - [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) 针对全球三分之二人口缺乏网络连接的现状，Google 通过 Project Loon 研发高空气球通信系统，利用气囊充排气技术在平流层灵活搭乘风层编织低成本浮动网络。 - [12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) Page 呼吁在保护隐私的同时切勿“把婴儿与洗澡水一起倒掉”，若能建立透明且去标识化的医疗病历共享机制，预计每年有望挽救 10 万条生命。 - [15:12](https://www.youtube.com/watch?v=mArrNRWQEso&t=912s) Google 自动驾驶汽车测试里程已突破 10 万英里，Page 强调重塑交通不仅能应对车祸这一全美 34 岁以下年轻人的头号死因，更能释放洛杉矶等城市被道路占用的半数土地。 - [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) Page 提出企业常因“错失未来”而迅速衰败，因此 Google 奉行“额外性”（additionality）哲学，全力投入无人涉足的高风险冷门领域与颠覆性技术。 ## Key takeaways - [00:13](https://www.youtube.com/watch?v=mArrNRWQEso&t=13s) Google 的核心使命始终是整合全球信息并使其人人可及，但当前的计算设备仍缺乏情境感知能力，搜索技术的发展仍处于早期阶段。 - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) 为了克服人机交互中语音识别的瓶颈，Google 投入机器学习研究，并在无人工干预的情况下让算法自主从 YouTube 视频中抽象出“猫”的概念。 - [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) DeepMind 的无监督强化学习展现出惊人的通用决策能力，同一个模型仅凭屏幕像素输入即可自主学会多种不同游戏并达到超越人类的水准。 - [06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) 肯尼亚农民 Zack Matere 利用互联网土方拯救农作物并设立线下信息黑板的故事，证明了信息的普及与利用方式足以改变普通人的现实命运。 - [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) 全球仍有三分之二的人口无法接入互联网，Project Loon 试图通过低成本的高空气球替代耗资高昂的卫星网络以弥合数字鸿沟。 - [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) Project Loon 利用气压泵调节气囊浮力来捕捉不同平流层高度的风向，从而实现对气球漂移轨迹的控制与全球网状通信覆盖。 - [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) 国家监控行动应当保持制度透明，政府在缺乏公众知情与公开讨论的情况下秘密监控，严重损害了民主体制并迫使科技公司充当屏障。 - [12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) 隐私保护与数据开放需要平衡，过度防范隐私风险可能阻碍数据共享带来深远的社会效益。 - [12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) Page 以自身声带患病的经历说明，公开健康数据不仅能填补罕见病学空白，匿名医疗数据共享更有潜力每年挽救 10 万人的生命。 - [15:12](https://www.youtube.com/watch?v=mArrNRWQEso&t=912s) 车祸是美国年轻群体的头号死因，且道路与停车场占用了洛杉矶等城市约 50% 的土地，自动驾驶与立体化骑行方案将极大重塑城市交通与空间格局。 - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Google X 的核心原则是经济学中的“额外性”（additionality），即专注于“如果 Google 不去做，世界上就没有其他人会去做”的难题。 - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) 光有技术发明是远远不够的，必须将前沿创新与强大的工程商业化能力相结合，才能像交流电普及一样将突破真正交付给大众。 - [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) 运作良好的商业实体同样能承担深远的公益使命，科技公司不应局限于渐进式改良，而应成为推动社会革命性变革的核心引擎。 - [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) 绝大多数企业衰亡的根本原因在于“错失了未来”，领导者的核心职责应是洞察并创造未来，而非仅局限于眼前的日常业务。 - [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) Page 反思早期曾因在 Android 上花费精力而心生内疚，但事实证明勇于押注非核心的未知领域才是把握时代未来的关键。 ## Detailed walkthrough ### [00:13](https://www.youtube.com/watch?v=mArrNRWQEso&t=13s) 幽默开场与重审 Google 的根本使命 访谈在主持 Charlie Rose 与 Google 联合创始人兼首席执行官 Larry Page 的风趣调侃中拉开帷幕。Charlie Rose 透露，Larry Page 事前曾给他发邮件叮嘱道：“我们得确保在台上看起来不像两个无聊的中年男人。”对此，Charlie Rose 调侃回应称自己感到受宠若惊，因为事实不仅是他比 Page 更年长，而且 Page 的身家资产也远比他丰厚。随后，Rose 为本场对话定下基调：全面探讨互联网、Google 自身、搜索技术、隐私哲学、过去的发展轨迹，以及最为核心的议题——科技的未来。 ```mermaid flowchart TD accTitle: Google 从早期使命到情境化计算的演化 accDescr: 展示 Google 创立初期的搜索愿景，经过十余年演化至情境化计算与机器学习的路径。 subgraph S1["创始使命与现实差距"] A["整合全球信息<br/>使人人皆可访问并从中受益"] --> B["计算设备体验依然笨拙<br/>无法真正理解情境与意图"] end subgraph S2["进化方向：情境感知"] B --> C["突破传统关键词搜索"] C --> D["情境化计算工具<br/>（如 Google Now）"] D --> E["融合机器学习与深度理解"] end ``` 当被问及“Google 如今走到了哪里，又将去向何方”时，Larry Page 回归了公司最初的初心。他表示，团队长期以来的核心使命始终是“整合全球信息，使人人皆可访问并从中受益”。常有人好奇探问这是否依然是 Google 现阶段的工作重心，Page 坦言自己也常反思这一点。在他看来，搜索技术承载着极为深邃的内涵，要真正理解用户心中所想并融会贯通全球信息，目前的进展其实仍然处在令人难以置信的初级阶段：“We've been at it for 15 years already, but it's not at all done.”（我们已经努力了15年，但根本还没有完成）。 至于为何尚未完成，Page 尖锐地指出了当前计算机系统的落后与繁琐。他认为现代计算机在很大程度上依然处于“一团乱麻”（a mess）的笨拙状态：设备不知道用户的物理位置、不知道用户正在做什么、更不知道用户掌握了什么知识。Google 近期的发力重点正是消除这种笨重感，让设备能够感知人类身处的情境（Context）。例如 Google Now，不仅能根据位置提供帮助，还能预判用户即刻所需。真正的智能计算应当深入理解人和信息，而行业目前距离这个目标仍有极大差距。 --- ### [02:29](https://www.youtube.com/watch?v=mArrNRWQEso&t=149s) DeepMind、无监督学习与在 YouTube 上“认出猫” 随着对人机交互理解的深入，话题转向了 Google 对英国人工智能公司 DeepMind 的收购及其背后的技术路径。Larry Page 解释道，探索搜索的本质必然要求计算机摆脱笨拙，而实现顺畅的人机交互首先要攻克语音技术。然而，当时业界的语音识别能力并不高，无法真正听懂用户的意图，Google 因而启动了机器学习研究以提升语音识别性能。在这一探索过程中，团队开始尝试让机器学习算法直接分析海量的 YouTube 视频，结果算法在毫无人工干预的情况下，自主归纳并发现了“猫”这一概念。Page 强调，这种从无到有的自主抽象代表了真正关键的突破。 ```mermaid flowchart LR accTitle: DeepMind 的端到端强化学习决策循环 accDescr: 展示 AI 系统如何直接从游戏原始像素获取输入，并在无额外编码的情况下输出操作指令与刷新高分。 subgraph Input["原始视觉输入"] P["游戏屏幕像素<br/>(Pixels)"] end subgraph Model["单一强化学习程序"] M["无监督 / 自主试错学习<br/>（仅凭操作按键与分数反馈）"] end subgraph Output["超人类表现"] A["自主摸索高级策略<br/>（如将拳击对手逼入角落痛击）"] end P --> M M --> A ``` DeepMind 最震撼之处在于其卓越的无监督学习能力。Page 现场展示了一段 AI 运行经典街机游戏的视频，演示中同一个程序能够在仅接收基础屏幕像素（Pixels）、操作控制按键以及最终得分反馈的前提下，完全自主学会玩各种不同的老游戏，并且表现达到了超越人类的水平（Superhuman performance）。现场特别播放了拳击游戏的画面：计算机控制的左侧角色自主领悟出了压制战术，将对手牢牢堵在角落反复击打并不断狂揽分数。Page 畅想道，一旦将这种强大的自主决策与策略优化智能应用于现实生活中的日程管理或个性化信息检索，将彻底颠覆人们的生活方式。 随后，现场大屏幕展示了早期神经网络自主从 YouTube 图像中提炼出的“猫脸”特征图。Page 强调指出，系统此前从未接受过“猫”的任何先验知识与人工标注训练，但它自主构建了这一概念：“And there's no training, no notion of a cat, but this concept of a cat is something important that you would understand, and now that the machines can kind of understand.”（这里没有任何训练，没有任何关于猫的概念，但猫的概念是你能够理解的重要事物，如今机器也开始能理解了）。谈及 DeepMind 创始人 Demis Hassabis 时，Page 赞赏其兼具神经科学与计算机科学的双重视角，不仅攻读了脑科学博士学位，还将这两个领域交叉融合，探索智能的终极本质。 - DeepMind 创立于英国，后被 Google 战略收购。 - 核心技术特性：仅凭相同的算法程序与无监督/强化机制，在多款老游戏中全面刷新超人类纪录。 - 知识发现验证：在未标记任何标签的 YouTube 视频海洋中，系统自发聚类出猫的视觉抽象模型。 --- ### [06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) 肯尼亚农夫的故事：搜索重塑现实生活的力量 为了诠释搜索在现实世界中如何深远地改变普通人的生活轨迹，Larry Page 在台上播放了一段关于肯尼亚索伊（Soy）地区的真实短片。短片主角是当地的一位普通农民 Zack Matere。 视频详细记录了 Zack Matere 利用互联网改变命运的过程： - **挽救农作物**：Zack 种下的马铃薯突然一株接一株枯死。他翻阅了手中所有书籍却毫无所获，于是前往网吧搜索“马铃薯疾病”（Potato diseases）。网络资料提示蚂蚁可能是元凶，并给出了一个极为平民化的土方——将草木灰撒在农作物上。几天后，蚂蚁彻底消失，庄稼得以保全。 - **邻里赋能**：深感互联网震撼的 Zack 随后带领一位想扩大业务的朋友去网吧查阅资料；后来这位朋友成功在当地学校架设起了一座风力发电机，让原本贫瘠的土地诞生出崭新的生机。 - **信息黑板**：意识到身边绝大多数人无法直接触达数字世界后，Zack 想出了一种“连外婆也能使用”的物理互联网——他在路旁立起一块简易的木质信息公告栏。每当手机接收到天气、农业或市场资讯，他便亲手将其誊写在黑板上，让它充当社区的公共“计算机屏幕”。 Zack 在短片中动情地表达了对信息价值的领悟：“Information is powerful, but it is how we use it that will define us.”（信息是强大的，但正是我们使用它的方式界定了我们是谁）。Page 补充道，这并非官方事先策划的营销宣传，而是 Google 团队在新闻报道中偶然读到这个动人故事后，顺藤摸瓜找到 Zack 并制作了这段纪实影像。 --- ### [08:43](https://www.youtube.com/watch?v=mArrNRWQEso&t=523s) Project Loon：高空气球构建全球平价网络 针对 Zack 的故事，Charlie Rose 指出，许多深知 Page 的人都深信“Larry 渴望改变世界，且坚信科技能够指引道路”，这必然要求所有人都能获得平等的网络接入权限。Page 对此高度赞同，并坦承自己目前在思考未来时，正把极大精力放在突破网络接入限制（Access）上。他随即向观众介绍了 Google 刚公开不久的平价联网探索计划——Project Loon。 ```mermaid flowchart TD accTitle: Project Loon 高空气球环球通信架构 accDescr: 展示 Project Loon 如何通过气压泵调节飞行高度、捕捉平流层风层，构建全球平价浮动网状通信覆盖。 subgraph AltitudeControl["平流层动态高度调节"] A["双气囊气泵系统<br/>（充入或排出空气）"] --> B["升高或降低飞行高度"] end subgraph WindSteering["风向利用与路径规划"] B --> C["捕捉不同层级平流层风向<br/>（依托大规模气象仿真模拟）"] end subgraph GlobalMesh["全球覆盖形态"] C --> D["大致控制移动轨迹与漂移路线"] D --> E["跨球际平流层浮动网状网络<br/>（高性价比覆盖全球未联网人群）"] end ``` Page 抛出了一组发人深省的数据：当前全球依然有三分之二的人口无法享有稳定可靠的互联网连接。Project Loon 旨在利用高空漂浮气球作为信号基站，以极具成本效益的方式向偏远地区输送网络。在探讨其工程实现时，Page 分享了研发背后的创新过程： - **立项历程**：团队就低成本高空接入点这一构想论证推演了至少五年。传统的通信覆盖高度往往依赖发射周期漫长、耗资巨大的同步轨道卫星，而气球的释放与升空则极其简便快捷。 - **历史灵感**：Page 曾通过搜索发现，早在三四十年前就曾有人释放气球成功环绕地球多次，这让他坚信现代科技完全能在这个方向上大有作为。 - **高空风向操控**：面对外界关于气球是否只能“任由风摆布”（at the mercy of the wind）的质疑，Page 透露团队进行了此前罕见的大规模超精细气象仿真模拟。工程师通过向气球内部泵入或排出空气来调节其浮力与所处高度，进而精准搭乘不同层级平流层的风流。借助这一手段，团队能够大致掌控气球的走向，最终在全球上空编织出一张覆盖全人类的平流层网状通信网（Mesh network）。 在话题逐步迈向交通运输自动化之前，Charlie Rose 将焦点引入了当时 TED 大会上最核心的争议焦点——就在前一天，爱德华·斯诺登（Edward Snowden）通过远程视频亮相并引发轰动，而 Sergey Brin 也刚与斯诺登完成合影。Rose 借此顺理成章地将对话引向了数字化时代风口浪尖上的核心议题：网络安全与用户隐私。 ### [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) 政府监控与民主：安全与隐私的界限 Larry Page 在回应关于 Edward Snowden 与国家监控的提问时，首先区分了安全（security）与隐私（privacy）的概念。他强调这两者密不可分，没有安全就谈不上隐私。Page 对美国政府秘密进行大规模监控且向公众隐瞒表达了极大的失望。他认为，政府在完全没有经过公众讨论的情况下开展此类行动，迫使 Google 处于必须挺身保护用户免受政府侵犯的境地，这严重损害了民主体制的基石。 对于政府以反恐为由开展秘密行动的辩护，Page 进一步明确了公众知情权的界限：民众不需要知道具体在防范哪一次特定的恐怖袭击细节，但必须知晓监控的整体边界、类型、手段以及背后的原因。当 Charlie Rose 提到政府甚至从未向 Google 提出沟通要求时，Page 强调问题不仅仅关乎 Google，而关乎全体公众。他尖锐地指出，政府秘密行事其实是对自身的巨大损害：“I don't think we can have a democracy if we're having to protect you and our users from the government for stuff that we've never had a conversation about.” - **核心观点**：民主社会的正常运转需要公开讨论，缺失监督的秘密监控让 Google 被迫充当阻挡政府侵犯公民权利的屏障。 - **透明度诉求**：无需公开敏感的反恐情报细节，但必须公开监控行为的“参数、类型、实施方式及正当理由”。 --- ### [12:35](https://www.youtube.com/watch?v=mArrNRWQEso&t=755s) 隐私权与数据共享：以医疗数据与声带疾病为例 谈到个人隐私时，Page 指出智能手机等设备的普及使个人数据剧增，人们产生疑虑完全合情合理。Google 的应对策略是为用户提供透明度与控制权，例如通过展示搜索历史与位置记录、并在 Chrome 中推行无痕模式（incognito mode）。然而，Page 最担心的趋势是社会在隐私讨论中“把婴儿与洗澡水一起倒掉”（throw out the baby with the bathwater），因过度防范隐私风险而彻底抹杀了数据共享所能带来的巨大社会价值。 Page 以自己的亲身经历作为佐证：他的声带曾严重受损并失去声音。当 Charlie Rose 开玩笑说愿意做任何事帮他恢复时，Page 也幽默地接梗说“拿上你的巫毒娃娃，用上所有手段吧”。Page 随后讲述，在联合创始人 Sergey Brin 的鼓励下，他克服恐惧将自己的声带健康问题公之于众。这一举动引发了数以千计拥有类似病症的患者回应并参与调查，填补了此前该领域缺乏数据的空白。由此，Page 提出了对医疗记录匿名开放的愿景：如果科研医生能够匿名调阅患者病历，且患者能清楚查看是哪位医生因何原因调阅了自己的记录，这种透明的双向机制不仅能帮助患者了解自身病情，还将极大地推进医学研究。Page 强调：“I think if we just did that, we'd save 100,000 lives this year.” ```mermaid flowchart TD accTitle: 医疗数据匿名共享的价值循环 accDescr: 展示患者匿名共享医疗记录至研究人员，再通过研究反馈与挽救生命的良性闭环 subgraph UserSide["患者与用户端"] A["个人健康与病历数据"] -->|"匿名化脱敏处理"| B["开放共享数据库"] D["患者查阅审计日志<br/>(知晓哪位医生/因何目的访问)"] -.->|"建立信任与了解病情"| A end subgraph ResearchSide["科研与社会效益"] B -->|"合法合规调阅"| C["医学科研人员"] C -->|"疾病研究与临床发现"| E["预计每年可挽救 100,000 条生命"] E -.->|"回馈医疗进步"| D end ``` - **数据控制与透明度**：强调给予用户选择权，知晓数据何时被收集及用途。 - **医疗记录开放的潜在收益**：若实现去标识化的病历共享，每年有望挽救十万人的生命。 - **亲身案例**：克服病史公开的恐惧后，获得了成千上万例此前完全缺失的相关病理反馈。 --- ### [15:12](https://www.youtube.com/watch?v=mArrNRWQEso&t=912s) 交通系统的重塑：自动驾驶汽车与悬挂式自行车道 当话题转向交通出行时，Page 回忆起自己在密歇根大学读大学时的经历。当年他在冰天雪地中苦等公交车，为此深感沮丧，随后深入研究了公交系统的运营成本，从而对交通系统的优化产生了浓厚兴趣。大约在 18 年前，他接触到了研发自动驾驶汽车的团队并深受吸引。Page 指出，全球每年有超过 2,000 万人因交通事故受伤，车祸是美国 34 岁以下人群的头号死因。自动驾驶不仅为了拯救生命，更能大幅优化城市空间——洛杉矶一半的面积都被停车场与公路占据，大多数城市也相差无几。他透露，Google 的完全自动驾驶测试里程已经突破了 10 万英里。 除了自动驾驶，Page 还对自行车交通充满热情。Google 园区内免费为员工提供自行车，自行车 24 小时高频运转甚至经常磨损报废。为了进一步推广自行车出行并确保安全，Page 思考如何以极低成本将自行车与机动车流隔离开来。他在现场播放了一段概念视频，展示了一种架设在城市上空的悬挂式封闭轨道自行车道。虽然 Google 并未正式开发该项目，但这类异想天开的设想展示了如何打破常规、解决城市交通拥堵与空间浪费难题。 - **交通事故现状**：全球每年超过 2,000 万人受伤；车祸是全美 34 岁以下人群的首要死因。 - **城市空间利用**：洛杉矶等大城市约 50% 的土地被道路和停车场占用。 - **研发进展**：Google 自动驾驶汽车此时已完成超过 100,000 英里的全自动驾驶里程。 - **空中自行车概念**：探索低成本实现非机动车与机动车立交隔离的交通方案。 --- ### [17:48](https://www.youtube.com/watch?v=mArrNRWQEso&t=1068s) Google X 哲学：额外性原则与重塑企业使命 Charlie Rose 随后探寻了 Google X 背后的思维哲学，提问 Page 为何不满足于微小且可衡量的渐进式进步。Page 借用了经济学中的“额外性”（additionality）概念来解释：Google X 旨在做那些“如果 Google 不去做，世界上就没有其他人会去做”的事。他认为做别人认为不可能的事才能产生最深远的影响。随着对技术前沿了解的加深，人们所能看到的可能性也会不断拓宽，例如 Project Loon（平流层热气球网络计划）之所以可行，正是因为发现了适合高空气球的新型材料。 谈到历史上 Nikola Tesla 发明了交流电却难以自行推广、最终依靠他人耗费漫长时间才普及的教训，Page 总结指出单纯的发明远远不够：“I think invention is not enough.” 他强调必须将前沿创新、发明精神与强大的商业化落地能力相结合，才能真正把技术交付给大众，并给欠缺优质互联网服务的全球三分之二人口带去希望。针对大众普遍认为“企业生来邪恶”的偏见，Page 坦言多数公司因循守旧、仍在重复二三十年前微小的渐进式改良，确实难辞其咎；但他坚信若运营得当，企业能成为推动革命性变革的最大力量。谈及曾有人提及 Page 宁愿把遗产留给 Elon Musk 也不愿捐作常规慈善的传闻，Page 回应称 Musk 创立公司去殖民火星以备份人类文明，本质上是兼具商业与崇高公益属性的追求。现代科技界聚集了大量财富与顶尖人才，与其仅仅将工作视为谋生，不如让汇聚了人类最多时间与资本的企业成为推动社会向好变革的核心引擎。 - **额外性原则（Additionality）**：聚焦于若无自身投入便无法实现的突破性技术与方案。 - **商业与发明结合**：吸取特斯拉的教训，将发明与工程商业化能力紧密结合。 - **企业角色重塑**：科技界需要的不是渐进式改良（incremental change），而是革命性变革（revolutionary change）；运作良好的商业实体同样能承担深远的公益使命。 ### [21:47](https://www.youtube.com/watch?v=mArrNRWQEso&t=1307s) 错失未来：企业的败局与专注冷门赛道的远见 查理·罗斯（Charlie Rose）在访谈尾声抛出了一个深具哲理性与领导力内核的问题。他提到，比尔·盖茨（Bill Gates）与沃伦·巴菲特（Warren Buffett）在被问及成功的秘诀时，都不约而同地给出了“专注”（focus）这一答案。罗斯借此提问佩奇，究竟是一种怎样的思维品质，能够让佩奇在深刻改变当下的同时，始终放眼并构想遥远的未来。 拉里·佩奇对此给出了深刻的剖析。他指出，自己一直在观察大量企业为何无法在时间的洗牌中长久立足，尤其是在当今企业更迭节奏日益加快的背景下。佩奇反思道：“通常仅仅是因为他们错失了未来。”（“And usually it's just that they missed the future.”）为了避免重蹈覆辙，他的首要目标就是将全副精力集中在洞察未来的真正形态上，并思索如何去亲手创造这一未来，进而推动整个组织聚焦并高速迈进。 佩奇将这种思维特质归结为好奇心、关注他人忽略的事物，以及勇于投入没有人着手的方向。他强调，这种“额外增量”（additionality）往往蕴藏在无人涉足的未知之中，即便伴随着巨大的风险也值得承担。他还坦率地分享了关于 Android 的往事：在早期刚收购这家初创公司时，它与 Google 当时的核心搜索业务大相径庭，以至于自己花时间在上面时甚至心生内疚。然而佩奇反思道：“那是愚蠢的。那就是未来，对吧？那是一件值得去投入的好事。”（“That was stupid. That was the future, right? That was a good thing to be working on.”）随着访谈在全场的热烈掌声中走向收尾，罗斯对佩奇的到来与深刻分享表达了由衷的感谢。 ```mermaid flowchart TD accTitle: 企业兴衰与未来探索的思维逻辑 accDescr: 对比传统企业错失未来的淘汰路径，与佩奇倡导的聚焦未知与承担风险的创新路径。 subgraph FailureMode["传统企业的衰败路径"] A["只聚焦眼前核心业务"] --> B["忽视外部技术演进"] B --> C["错失未来浪潮并迅速被淘汰"] end subgraph InnovationMode["Google 面向未来的创新逻辑"] D["保持强烈好奇心"] --> E["布局无人涉足的冷门领域"] E --> F["承担风险并快速驱动组织实现"] F --> G["孕育颠覆性未来 (如 Android)"] end ``` - **企业衰落的核心症结**：拉里·佩奇认为企业被时代快速淘汰的根本原因，不在于日常运营失误，而在于领导层“错失了未来”（missed the future）。 - **“额外增量”（Additionality）法则**：真正的价值突破来自于去攻克其他人未曾涉足的冷门与未知课题，而不是在已有共识的赛道里过度内耗。 - **Android 的早期反思**：Android 最初只是一家不起眼的小型初创公司，佩奇曾因在其身上耗费精力而感到内疚，但最终事实证明，它正是移动互联网时代的未来所向。 ## Notable quotes > [11:07](https://www.youtube.com/watch?v=mArrNRWQEso&t=667s) I don't think we can have a democracy if we're having to protect you and our users from the government for stuff that we've never had a conversation about. — Larry Page > [06:20](https://www.youtube.com/watch?v=mArrNRWQEso&t=380s) Information is powerful, but it is how we use it that will define us. — Zack Matere ## Mentioned Google, Larry Page, Charlie Rose, Google Now, DeepMind, Demis Hassabis, YouTube, Zack Matere, Project Loon, Edward Snowden, Sergey Brin, Google Chrome, Google X, Nikola Tesla, Elon Musk, Android, Bill Gates, Warren Buffett ## Video description Onstage at TED2014, Charlie Rose interviews Google CEO Larry Page about his far-off vision for the company. It includes aerial bikeways and internet balloons ... and then it gets even more interesting, as Page talks through the company's recent acquisition of Deep Mind, an AI that is learning some surprising things. TEDTalks is a daily video podcast of the best talks and performances from the TED Conference, where the world's leading thinkers and doers give the talk of their lives in 18 minutes (or less). Look for talks on Technology, Entertainment and Design -- plus science, business, global issues, the arts and much more. Find closed captions and translated subtitles in many languages at http://www.ted.com/translate Follow TED news on Twitter: http://www.twitter.com/tednews Like TED on Facebook: https://www.facebook.com/TED Subscribe to our channel: http://www.youtube.com/user/TEDtalksDirector [Source](https://www.youtube.com/watch?v=mArrNRWQEso)

Source quotes

[11:07] Larry Page: I don't think we can have a democracy if we're having to protect you and our users from the government for stuff that we've never had a conversation about. [6:20] Zack Matere: Information is powerful, but it is how we use it that will define us.

Source quotes

[11:07] Larry Page: I don’t think we can have a democracy if we’re having to protect you and our users from the government for stuff that we’ve never had a conversation about. [17:48] Larry Page: We need, especially in technology, we need revolutionary change, not incremental change. [6:20] Zack Matere: Information is powerful, but it is how we use it that will define us.

Original transcript

[0:13] Charlie Rose: So Larry sent me an email [0:17] and he basically said, [0:18] we've got to make sure that we don't seem like we're [0:22] a couple of middle-aged boring men. [0:27] I said, I'm flattered by that -- [0:30] (Laughter) — [0:32] because I'm a bit older, [0:36] and he has a bit more net worth than I do. [0:40] Larry Page: Well, thank you. [0:42] CR: So we'll have a conversation about [0:45] the Internet, and we'll have a conversation Google, [0:48] and we'll have a conversation about search [0:50] and privacy, [0:51] and also about your philosophy [0:52] and a sense of how you've connected the dots [0:55] and how this journey that began [0:57] some time ago [0:58] has such interesting prospects. [1:00] Mainly we want to talk about the future. [1:03] So my first question: Where is Google [1:04] and where is it going? [1:06] LP: Well, this is something we think about a lot, [1:08] and our mission we defined a long time ago [1:11] is to organize the world's information [1:14] and make it universally accessible and useful. [1:17] And people always say, [1:19] is that really what you guys are still doing? [1:21] And I always kind of think about that myself, [1:23] and I'm not quite sure. [1:26] But actually, when I think about search, [1:30] it's such a deep thing for all of us, [1:32] to really understand what you want, [1:35] to understand the world's information, [1:37] and we're still very much in the early stages of that, [1:40] which is totally crazy. [1:42] We've been at it for 15 years already, [1:45] but it's not at all done. [1:48] CR: When it's done, how will it be? [1:51] LP: Well, I guess, [1:54] in thinking about where we're going -- [1:56] you know, why is it not done? -- [1:58] a lot of it is just computing's kind of a mess. [2:01] You know, your computer doesn't know where you are, [2:03] it doesn't know what you're doing, [2:05] it doesn't know what you know, [2:06] and a lot we've been trying to do recently [2:09] is just make your devices work, [2:12] make them understand your context. [2:15] Google Now, you know, knows where you are, [2:17] knows what you may need. [2:19] So really having computing work and understand you [2:23] and understand that information, [2:25] we really haven't done that yet. [2:27] It's still very, very clunky. [2:29] CR: Tell me, when you look at what Google is doing, [2:31] where does Deep Mind fit? [2:34] LP: Yeah, so Deep Mind is a company [2:36] we just acquired recently. [2:38] It's in the U.K. [2:41] First, let me tell you the way we got there, [2:44] which was looking at search [2:46] and really understanding, [2:48] trying to understand everything, [2:50] and also make the computers not clunky [2:52] and really understand you -- [2:54] like, voice was really important. [2:56] So what's the state of the art on speech recognition? [2:59] It's not very good. [3:01] It doesn't really understand you. [3:03] So we started doing machine learning research [3:05] to improve that. [3:06] That helped a lot. [3:08] And we started just looking at things like YouTube. [3:10] Can we understand YouTube? [3:12] But we actually ran machine learning on YouTube [3:15] and it discovered cats, just by itself. [3:19] Now, that's an important concept. [3:21] And we realized there's really something here. [3:24] If we can learn what cats are, [3:26] that must be really important. [3:28] So I think Deep Mind, [3:31] what's really amazing about Deep Mind [3:33] is that it can actually -- [3:35] they're learning things in this unsupervised way. [3:39] They started with video games, [3:41] and really just, maybe I can show the video, [3:44] just playing video games, [3:46] and learning how to do that automatically. [3:48] CR: Take a look at the video games [3:50] and how machines are coming to be able [3:52] to do some remarkable things. [3:55] LP: The amazing thing about this [3:56] is this is, I mean, obviously, [3:58] these are old games, [3:59] but the system just sees what you see, the pixels, [4:04] and it has the controls and it has the score, [4:06] and it's learned to play all of these games, [4:09] same program. [4:10] It's learned to play all of these games [4:12] with superhuman performance. [4:14] We've not been able to do things like this [4:16] with computers before. [4:17] And maybe I'll just narrate this one quickly. [4:20] This is boxing, and it figures out it can [4:23] sort of pin the opponent down. [4:25] The computer's on the left, [4:27] and it's just racking up points. [4:30] So imagine if this kind [4:32] of intelligence were thrown at your schedule, [4:34] or your information needs, or things like that. [4:39] We're really just at the beginning of that, [4:41] and that's what I'm really excited about. [4:44] CR: When you look at all that's taken place [4:46] with Deep Mind and the boxing, [4:49] also a part of where we're going [4:51] is artificial intelligence. [4:54] Where are we, when you look at that? [4:57] LP: Well, I think for me, [4:59] this is kind of one of the most exciting things [5:00] I've seen in a long time. [5:02] The guy who started this company, Demis, [5:05] has a neuroscience and a computer science background. [5:07] He went back to school [5:09] to get his Ph.D. to study the brain. [5:12] And so I think we're seeing a lot of exciting work [5:15] going on that sort of crosses computer science [5:18] and neuroscience [5:20] in terms of really understanding [5:22] what it takes to make something smart [5:24] and do really interesting things. [5:26] CR: But where's the level of it now? [5:28] And how fast do you think we are moving? [5:31] LP: Well, this is the state of the art right now, [5:34] understanding cats on YouTube [5:36] and things like that, [5:38] improving voice recognition. [5:40] We used a lot of machine learning [5:42] to improve things incrementally, [5:45] but I think for me, this example's really exciting, [5:48] because it's one program [5:50] that can do a lot of different things. [5:52] CR: I don't know if we can do this, [5:53] but we've got the image of the cat. [5:55] It would be wonderful to see this. [5:56] This is how machines looked at cats [5:59] and what they came up with. [6:00] Can we see that image? [6:01] LP: Yeah. CR: There it is. Can you see the cat? [6:03] Designed by machines, seen by machines. [6:05] LP: That's right. [6:07] So this is learned from just watching YouTube. [6:09] And there's no training, [6:11] no notion of a cat, [6:12] but this concept of a cat [6:15] is something important that you would understand, [6:18] and now that the machines can kind of understand. [6:20] Maybe just finishing [6:22] also on the search part, [6:24] it started with search, really understanding [6:27] people's context and their information. [6:29] I did have a video [6:31] I wanted to show quickly on that [6:33] that we actually found. [6:35] (Video) ["Soy, Kenya"] [6:40] Zack Matere: Not long ago, [6:42] I planted a crop of potatoes. [6:45] Then suddenly they started dying one after the other. [6:48] I checked out the books and they didn't tell me much. [6:51] So, I went and I did a search. [6:53] ["Zack Matere, Farmer"] [6:57] Potato diseases. [7:00] One of the websites told me [7:02] that ants could be the problem. [7:04] It said, sprinkle wood ash over the plants. [7:06] Then after a few days the ants disappeared. [7:08] I got excited about the Internet. [7:11] I have this friend [7:13] who really would like to expand his business. [7:16] So I went with him to the cyber cafe [7:20] and we checked out several sites. [7:22] When I met him next, he was going to put a windmill [7:25] at the local school. [7:27] I felt proud because [7:29] something that wasn't there before [7:31] was suddenly there. [7:33] I realized that not everybody [7:36] can be able to access [7:37] what I was able to access. [7:39] I thought that I need to have an Internet [7:40] that my grandmother can use. [7:42] So I thought about a notice board. [7:45] A simple wooden notice board. [7:47] When I get information on my phone, [7:49] I'm able to post the information [7:51] on the notice board. [7:53] So it's basically like a computer. [7:56] I use the Internet to help people. [8:00] I think I am searching for [8:03] a better life [8:05] for me and my neighbors. [8:09] So many people have access to information, [8:13] but there's no follow-up to that. [8:15] I think the follow-up to that is our knowledge. [8:18] When people have the knowledge, [8:19] they can find solutions [8:21] without having to helped out. [8:23] Information is powerful, [8:25] but it is how we use it that will define us. [8:30] (Applause) [8:34] LP: Now, the amazing thing about that video, [8:37] actually, was we just read about it in the news, [8:38] and we found this gentlemen, [8:41] and made that little clip. [8:43] CR: When I talk to people about you, [8:44] they say to me, people who know you well, say, [8:47] Larry wants to change the world, [8:49] and he believes technology can show the way. [8:53] And that means access to the Internet. [8:55] It has to do with languages. [8:56] It also means how people can get access [8:59] and do things that will affect their community, [9:02] and this is an example. [9:04] LP: Yeah, that's right, and I think for me, [9:08] I have been focusing on access more, [9:10] if we're talking about the future. [9:13] We recently released this Loon Project [9:15] which is using balloons to do it. [9:18] It sounds totally crazy. [9:19] We can show the video here. [9:22] Actually, two out of three people in the world [9:23] don't have good Internet access now. [9:26] We actually think this can really help people [9:29] sort of cost-efficiently. [9:31] CR: It's a balloon. LP: Yeah, get access to the Internet. [9:34] CR: And why does this balloon give you access [9:36] to the Internet? [9:37] Because there was some interesting things [9:39] you had to do to figure out how [9:40] to make balloons possible, [9:43] they didn't have to be tethered. [9:44] LP: Yeah, and this is a good example of innovation. [9:46] Like, we've been thinking about this idea [9:49] for five years or more [9:51] before we started working on it, [9:52] but it was just really, [9:54] how do we get access points up high, cheaply? [9:57] You normally have to use satellites [9:59] and it takes a long time to launch them. [10:02] But you saw there how easy it is to launch a balloon [10:04] and get it up, [10:06] and actually again, it's the power of the Internet, [10:08] I did a search on it, [10:10] and I found, 30, 40 years ago, [10:12] someone had put up a balloon [10:14] and it had gone around the Earth multiple times. [10:17] And I thought, why can't we do that today? [10:20] And that's how this project got going. [10:22] CR: But are you at the mercy of the wind? [10:24] LP: Yeah, but it turns out, [10:26] we did some weather simulations [10:28] which probably hadn't really been done before, [10:30] and if you control the altitude of the balloons, [10:32] which you can do by pumping air into them [10:35] and other ways, [10:37] you can actually control roughly where they go, [10:40] and so I think we can build a worldwide mesh [10:42] of these balloons that can cover the whole planet. [10:45] CR: Before I talk about the future and transportation, [10:47] where you've been a nerd for a while, [10:49] and this fascination you have with transportation [10:52] and automated cars and bicycles, [10:54] let me talk a bit about what's been the subject here [10:55] earlier with Edward Snowden. [10:58] It is security and privacy. [11:01] You have to have been thinking about that. [11:03] LP: Yeah, absolutely. [11:05] I saw the picture of Sergey with Edward Snowden yesterday. [11:07] Some of you may have seen it. [11:10] But I think, for me, I guess, [11:14] privacy and security are a really important thing. [11:17] We think about it in terms of both things, [11:19] and I think you can't have privacy without security, [11:22] so let me just talk about security first, [11:25] because you asked about Snowden and all of that, [11:27] and then I'll say a little bit about privacy. [11:30] I think for me, it's tremendously disappointing [11:34] that the government [11:35] secretly did all this stuff and didn't tell us. [11:37] I don't think we can have a democracy [11:41] if we're having to protect you and our users [11:44] from the government [11:46] for stuff that we've never had a conversation about. [11:49] And I don't mean we have to know [11:50] what the particular terrorist attack is they're worried [11:52] about protecting us from, [11:54] but we do need to know [11:56] what the parameters of it is, [11:58] what kind of surveillance the government's [12:00] going to do and how and why, [12:02] and I think we haven't had that conversation. [12:05] So I think the government's actually done [12:07] itself a tremendous disservice [12:09] by doing all that in secret. [12:12] CR: Never coming to Google [12:13] to ask for anything. [12:15] LP: Not Google, but the public. [12:17] I think we need to have a debate about that, [12:20] or we can't have a functioning democracy. [12:23] It's just not possible. [12:24] So I'm sad that Google's [12:27] in the position of protecting you and our users [12:29] from the government [12:31] doing secret thing that nobody knows about. [12:33] It doesn't make any sense. [12:35] CR: Yeah. And then there's a privacy side of it. [12:38] LP: Yes. The privacy side, [12:40] I think it's -- the world is changing. [12:42] You carry a phone. It knows where you are. [12:46] There's so much more information about you, [12:49] and that's an important thing, [12:52] and it makes sense why people are asking [12:54] difficult questions. [12:56] We spend a lot of time thinking about this [13:00] and what the issues are. [13:02] I'm a little bit -- [13:04] I think the main thing that we need to do [13:05] is just provide people choice, [13:08] show them what data's being collected -- [13:10] search history, location data. [13:15] We're excited about incognito mode in Chrome, [13:18] and doing that in more ways, [13:20] just giving people more choice [13:21] and more awareness of what's going on. [13:25] I also think it's very easy. [13:27] What I'm worried is that we throw out [13:28] the baby with the bathwater. [13:30] And I look at, on your show, actually, [13:33] I kind of lost my voice, [13:35] and I haven't gotten it back. [13:36] I'm hoping that by talking to you [13:38] I'm going to get it back. [13:40] CR: If I could do anything, I would do that. [13:41] LP: All right. So get out your voodoo doll [13:44] and whatever you need to do. [13:46] But I think, you know what, I look at that, [13:48] I made that public, [13:50] and I got all this information. [13:51] We got a survey done on medical conditions [13:54] with people who have similar issues, [13:57] and I look at medical records, and I say, [14:02] wouldn't it be amazing [14:04] if everyone's medical records were available [14:06] anonymously [14:07] to research doctors? [14:10] And when someone accesses your medical record, [14:13] a research doctor, [14:15] they could see, you could see which doctor [14:17] accessed it and why, [14:19] and you could maybe learn about [14:21] what conditions you have. [14:22] I think if we just did that, [14:24] we'd save 100,000 lives this year. [14:26] CR: Absolutely. Let me go — (Applause) [14:29] LP: So I guess I'm just very worried that [14:32] with Internet privacy, [14:34] we're doing the same thing we're doing with medical records, [14:36] is we're throwing out the baby with the bathwater, [14:38] and we're not really thinking [14:40] about the tremendous good that can come [14:42] from people sharing information [14:45] with the right people in the right ways. [14:47] CR: And the necessary condition [14:49] that people have to have confidence [14:51] that their information will not be abused. [14:54] LP: Yeah, and I had this problem with my voice stuff. [14:55] I was scared to share it. [14:57] Sergey encouraged me to do that, [14:59] and it was a great thing to do. [15:01] CR: And the response has been overwhelming. [15:02] LP: Yeah, and people are super positive. [15:04] We got thousands and thousands of people [15:07] with similar conditions, [15:08] which there's no data on today. [15:11] So it was a really good thing. [15:12] CR: So talking about the future, what is it about you [15:16] and transportation systems? [15:19] LP: Yeah. I guess I was just frustrated [15:21] with this when I was at college in Michigan. [15:24] I had to get on the bus and take it [15:25] and wait for it. [15:27] And it was cold and snowing. [15:29] I did some research on how much it cost, [15:32] and I just became a bit obsessed with transportation systems. [15:38] CR: And that began the idea of an automated car. [15:41] LP: Yeah, about 18 years ago I learned about [15:42] people working on automated cars, [15:46] and I became fascinated by that, [15:47] and it takes a while to get these projects going, [15:50] but I'm super excited about the possibilities of that [15:55] improving the world. [15:57] There's 20 million people or more injured per year. [16:01] It's the leading cause of death [16:03] for people under 34 in the U.S. [16:05] CR: So you're talking about saving lives. [16:07] LP: Yeah, and also saving space [16:09] and making life better. [16:13] Los Angeles is half parking lots and roads, [16:17] half of the area, [16:19] and most cities are not far behind, actually. [16:22] It's just crazy [16:24] that that's what we use our space for. [16:25] CR: And how soon will we be there? [16:28] LP: I think we can be there very, very soon. [16:29] We've driven well over 100,000 miles [16:33] now totally automated. [16:37] I'm super excited about getting that out quickly. [16:41] CR: But it's not only you're talking about automated cars. [16:43] You also have this idea for bicycles. [16:45] LP: Well at Google, we got this idea [16:48] that we should just provide free bikes to everyone, [16:51] and that's been amazing, most of the trips. [16:54] You see bikes going everywhere, [16:56] and the bikes wear out. [16:57] They're getting used 24 hours a day. [16:59] CR: But you want to put them above the street, too. [17:01] LP: Well I said, how do we get people [17:02] using bikes more? [17:04] CR: We may have a video here. [17:05] LP: Yeah, let's show the video. [17:07] I just got excited about this. [17:10] (Music) [17:16] So this is actually how you might separate [17:18] bikes from cars with minimal cost. [17:26] Anyway, it looks totally crazy, [17:28] but I was actually thinking about our campus, [17:30] working with the Zippies and stuff, [17:32] and just trying to get a lot more bike usage, [17:35] and I was thinking about, [17:36] how do you cost-effectively separate [17:39] the bikes from traffic? [17:40] And I went and searched, [17:42] and this is what I found. [17:43] And we're not actually working on this, [17:45] that particular thing, [17:46] but it gets your imagination going. [17:48] CR: Let me close with this. [17:50] Give me a sense of the philosophy of your own mind. [17:52] You have this idea of [Google X]. [17:55] You don't simply want [17:58] to go in some small, measurable arena of progress. [18:03] LP: Yeah, I think [18:05] many of the things we just talked about are like that, [18:07] where they're really -- [18:10] I almost use the economic concept of additionality, [18:14] which means that you're doing something [18:16] that wouldn't happen unless you were actually doing it. [18:19] And I think the more you can do things like that, [18:22] the bigger impact you have, [18:24] and that's about doing things [18:27] that people might not think are possible. [18:31] And I've been amazed, [18:33] the more I learn about technology, [18:35] the more I realize I don't know, [18:37] and that's because this technological horizon, [18:40] the thing that you can see to do next, [18:43] the more you learn about technology, [18:45] the more you learn what's possible. [18:48] You learn that the balloons are possible [18:50] because there's some material that will work for them. [18:52] CR: What's interesting about you too, though, for me, [18:55] is that, we have lots of people [18:56] who are thinking about the future, [18:58] and they are going and looking and they're coming back, [19:02] but we never see the implementation. [19:04] I think of somebody you knew [19:05] and read about, Tesla. [19:08] The principle of that for you is what? [19:12] LP: Well, I think invention is not enough. [19:14] If you invent something, [19:15] Tesla invented electric power that we use, [19:18] but he struggled to get it out to people. [19:21] That had to be done by other people. [19:23] It took a long time. [19:24] And I think if we can actually combine both things, [19:28] where we have an innovation and invention focus, [19:32] plus the ability to really -- a company [19:35] that can really commercialize things [19:37] and get them to people [19:38] in a way that's positive for the world [19:40] and to give people hope. [19:42] You know, I'm amazed with the Loon Project [19:45] just how excited people were about that, [19:48] because it gave them hope [19:50] for the two thirds of the world [19:51] that doesn't have Internet right now that's any good. [19:54] CR: Which is a second thing about corporations. [19:56] You are one of those people who believe [19:59] that corporations are an agent of change [20:01] if they are run well. [20:03] LP: Yeah. I'm really dismayed [20:04] most people think companies are basically evil. [20:08] They get a bad rap. [20:09] And I think that's somewhat correct. [20:12] Companies are doing the same incremental thing [20:15] that they did 50 years ago [20:16] or 20 years ago. [20:18] That's not really what we need. [20:19] We need, especially in technology, [20:22] we need revolutionary change, [20:24] not incremental change. [20:25] CR: You once said, actually, [20:26] as I think I've got this about right, [20:28] that you might consider, [20:30] rather than giving your money, [20:31] if you were leaving it to some cause, [20:35] just simply giving it to Elon Musk, [20:37] because you had confidence [20:38] that he would change the future, [20:40] and that you would therefore — [20:42] LP: Yeah, if you want to go Mars, [20:43] he wants to go to Mars, [20:45] to back up humanity, [20:47] that's a worthy goal, but it's a company, [20:49] and it's philanthropical. [20:51] So I think we aim to do kind of similar things. [20:54] And I think, you ask, we have a lot of employees [20:57] at Google who have become pretty wealthy. [21:00] People make a lot of money in technology. [21:03] A lot of people in the room are pretty wealthy. [21:05] You're working because you want to change the world. [21:07] You want to make it better. [21:09] Why isn't the company that you work for [21:13] worthy not just of your time [21:14] but your money as well? [21:17] I mean, but we don't have a concept of that. [21:18] That's not how we think about companies, [21:21] and I think it's sad, [21:22] because companies are most of our effort. [21:26] They're where most of people's time is, [21:28] where a lot of the money is, [21:30] and so I think I'd like for us to help out [21:33] more than we are. [21:34] CR: When I close conversations with lots of people, [21:35] I always ask this question: [21:37] What state of mind, [21:39] what quality of mind is it [21:41] that has served you best? [21:42] People like Rupert Murdoch have said curiosity, [21:45] and other people in the media have said that. [21:47] Bill Gates and Warren Buffett have said focus. [21:51] What quality of mind, [21:52] as I leave this audience, [21:53] has enabled you to think about the future [21:57] and at the same time [21:58] change the present? [22:01] LP: You know, I think the most important thing -- [22:02] I looked at lots of companies [22:04] and why I thought they don't succeed over time. [22:07] We've had a more rapid turnover of companies. [22:10] And I said, what did they fundamentally do wrong? [22:13] What did those companies all do wrong? [22:15] And usually it's just that they missed the future. [22:18] And so I think, for me, [22:21] I just try to focus on that and say, [22:23] what is that future really going to be [22:25] and how do we create it, [22:27] and how do we cause our organization, [22:32] to really focus on that [22:34] and drive that at a really high rate? [22:38] And so that's been curiosity, [22:39] it's been looking at things [22:41] people might not think about, [22:42] working on things that no one else is working on, [22:45] because that's where the additionality really is, [22:49] and be willing to do that, [22:50] to take that risk. [22:52] Look at Android. [22:53] I felt guilty about working on Android [22:56] when it was starting. [22:57] It was a little startup we bought. [22:59] It wasn't really what we were really working on. [23:02] And I felt guilty about spending time on that. [23:04] That was stupid. [23:05] That was the future, right? [23:07] That was a good thing to be working on. [23:09] CR: It is great to see you here. [23:10] It's great to hear from you, [23:12] and a pleasure to sit at this table with you. [23:14] Thanks, Larry. [23:15] LP: Thank you. [23:17] (Applause) [23:21] CR: Larry Page.