Jensen Huang Dismisses AI Doomsaying, Backs Open Models

Jensen Huang asserted that safety and innovation are complementary, not contradictory elements in technology development. He rejected the notion that progress must sacrifice security, stating that quick innovation and execution can coexist with robust safety measures. America's leadership in technology, Huang explained, relies on its ability to advance rapidly while maintaining high safety standards.

Huang cataloged a series of past AI predictions that proved false, using them to challenge current apocalyptic forecasts. He pointed to the prediction that radiologists would be entirely replaced by AI within five years, a forecast that proved to be the exact opposite of reality. Similarly, predictions of 90% code generation by AI last year also failed to materialize, and expectations of 50% job loss have been disproven. Huang noted that safety concerns surrounding the releases of models like Llama and GPT were hyperbolic, demonstrating a pattern of exaggerated fear.

Jensen Huang advocated for a corporate culture where company building occurs without public fanfare or excessive discourse. He suggested that businesses should be built 'in silence,' emphasizing a focus on internal development rather than external pronouncements. Nvidia, in line with this philosophy, maintains an apolitical stance internally and restricts employees from making unauthorized public statements.

He questioned what concrete issues have arisen that necessitate immediate regulatory intervention. He further explained that most current AI issues originate from frontier labs due to their massive compute resources, while startups and individuals typically lack the computational power to pose significant systemic risks.

He noted that the term has been "weaponized" in public rhetoric to create an impression of an uncontrollable spiral. Huang assured that rigorous testing, evaluation, and engineering controls ensure RSI remains safe, and companies will maintain control over internal RSI processes before product releases.

He highlighted that 80% of AI-native companies rely on open models to develop their specific innovations. The demand for open models is driven by needs for sovereignty, privacy, and proprietary development, ensuring that every industry can benefit from accessible AI technologies.

Huang pointed out that China is a significant contributor to global open source software, largely due to its high volume of engineering graduates. He explained that once open source software or models are downloaded, they are typically forked and improved locally by users, becoming their own version. He concluded that the true competition is not about the origin of a model, but rather who can best utilize and exploit the technology.

Jensen Huang suggested that AI would redefine the nature of engineering work, moving it beyond its current reliance on constant coding. He posited that historically, engineering did not depend on typing code. Huang believes that future engineering will concentrate on high-level architecture and problem-solving, rather than manual coding, suggesting that engineers will engage in significant innovation 'after typing.'

Trump emphatically dismissed concerns about AI leading to societal collapse, branding such fears as 'a hoax,' while still urging prudence in its development.

Trump also drew comparisons between his administration's investment in AI and that of the Biden administration, consistently reiterating the transformative scale of AI development.

Jensen Huang emphasized AI's role in creating an enormous number of jobs, connecting its development to the broader goal of re-industrializing the United States. He noted the recent $400 billion in venture financing for AI and highlighted the surging demand for compute and data centers, all contributing to job growth.

This involves supporting every layer of infrastructure, from foundational models to hardware. Nvidia has forged partnerships with financial institutions like BlackRock and Goldman Sachs to secure financing and actively identifies and solves bottlenecks within the supply chain to ensure smooth development.

Huang explained Nvidia's strategic shift from supporting a single model, like OpenAI, to empowering many new frontier labs such as Anthropic, Groq, and Physical Intelligence. He articulated the company's philosophy: "go up as far as we need to, but as low as possible." This approach prioritizes helping the entire ecosystem succeed rather than carving out a slice of the market for themselves.

Jensen Huang stated his surprisingly uncompetitive nature, explaining that his focus is on nurturing the AI ecosystem rather than market dominance. He expressed a preference for multiple hyperscalers and believes that current market volatility favors agile, regional NeoClouds over slower, larger hyperscalers. Nvidia is actively scaling up global infrastructure to support these diverse regional players, promoting a collaborative environment.

Huang confirmed personal discussions with Musk about these plans and characterized Musk's unwavering commitment to large-scale projects as his 'superpower.' He noted that he and his team closely study process technology, even though Nvidia designs chips rather than fabricating them.

Huang indicated that 2030 is "just around the corner" in strategic terms, highlighting the country's rapid progress.