Percy Liang
All AI mindsPercy Liang, full AI read
Percy Liang, Founder, Together AI; Professor & Director CRFM, Stanford, Stanford University / Together AI (United States), ranks #12/520 on the AI Advancement Index (85.8). Known for Director of Stanford CRFM which coined 'foundation models'; created the HELM benchmark and co-founded Together AI; influential research on semantic parsing, robustness and reproducible open models. Strongest on Field-building (90.0, Leading).
Dimension read
| Dimension | Value | Standing | What a high vs low value means, and where Percy Liang sits |
|---|---|---|---|
| AAI AI Advancement (AAI) | 85.8 | Leading · #12/520 | High here, among the very top minds advancing AI. ▲ high: among the very top minds advancing AI · ▼ low: lower relative influence within this elite set |
| Research influence Research influence | 88.0 | Leading · #31/520 | High here, field-defining research contributions. ▲ high: field-defining research contributions · ▼ low: limited direct research influence |
| Frontier role Frontier role | 80.0 | Strong · #81/520 | High here, central to building today's frontier AI. ▲ high: central to building today's frontier AI · ▼ low: removed from frontier development |
| Thought leadership Thought leadership | 86.0 | Leading · #26/520 | High here, shapes how the field and public think about AI. ▲ high: shapes how the field and public think about AI · ▼ low: limited public/field influence |
| Field-building Field-building | 90.0 | Leading · #16/520 | High here, builds the field, mentorship, institutions, tools, community. ▲ high: builds the field, mentorship, institutions, tools, community · ▼ low: limited field-building footprint |
| Momentum Momentum | 86.0 | Strong · #81/520 | High here, driving AI's advancement right now. ▲ high: driving AI's advancement right now · ▼ low: less active at the current frontier |
Strengths
- Field-building (90.0, Leading), builds the field, mentorship, institutions, tools, community.
- Thought leadership (86.0, Leading), shapes how the field and public think about AI.
- Research influence (88.0, Leading), field-defining research contributions.
Risk factors
- A significant, well-rounded contributor to AI's advancement.
AI worldview
Ideas & positions
Percy Liang is a proponent of robust and reliable AI systems, emphasizing the importance of transparency, interpretability, and fairness in machine learning models. He has been instrumental in the development of foundation models, which are large-scale models trained on diverse data to serve as versatile bases for various applications. Liang's work on the HELM benchmark aims to evaluate the performance and robustness of these models across different tasks and domains. He advocates for open and collaborative approaches to AI research, as evidenced by his co-founding of Together AI, which focuses on creating open-source models. While he does not explicitly take a strong stance on existential risk, his emphasis on robustness and reliability suggests a cautious approach to the deployment of AI technologies.
What shapes the view
Liang's views are shaped by his academic background in computer science and his experience in leading research initiatives at Stanford. His focus on robustness and interpretability is influenced by the need to ensure that AI systems can be trusted and understood by users. His advocacy for open-source models reflects a belief in the democratization of AI technology and the importance of community-driven innovation. His work also highlights the economic and ethical implications of AI, particularly in ensuring that the benefits of AI are widely distributed.
The AI-powered future they see
Liang envisions a future where AI systems are more transparent, reliable, and accessible. He promotes the development of foundation models that can serve as building blocks for a wide range of applications, from natural language processing to robotics. He warns against the potential pitfalls of over-reliance on black-box models and emphasizes the need for continuous evaluation and improvement of AI systems to ensure they meet the needs of diverse users.