CHM Live | This Time It's Different: AI Startups Across Three Generations

CHM Live | This Time It's Different: AI Startups Across Three Generations

🎙 Computer History Museum 👥 177K 📅 October 9, 2025 ⏱ 69 min 👁 1K 📄 panel discussion 🧭 2026-08-06
Available in: English (current) Français

Keywords

AI boomexpert systemsSiriliquid neural networksentrepreneurship

Summary

The panel, moderated by Marc Weber, brings together Jerry Kaplan (Teknowledge), Adam Cheyer (Siri), and Daniela Rus (Liquid AI) to discuss AI startups across three generations. Kaplan describes the 1980s expert systems boom, highlighting the role of Lisp machines and the incorrect assumptions that led to the AI winter. Cheyer recounts the founding of Siri, initially pitched as a ‘do engine’ to avoid the AI label, and its acquisition by Apple, noting Steve Jobs’ role in popularizing the term AI. Rus contrasts the current MIT startup culture with the past, emphasizing the shift from academic aspirations to entrepreneurial ambitions. The discussion explores common themes: the cyclical nature of AI hype, the importance of practical applications, and the evolution of AI from rule-based systems to modern neural networks. The panel also touches on the potential of liquid neural networks and the future of AI, offering lessons from past booms and busts for today’s entrepreneurs.

153 words

Critical Evaluation

The panel provides a valuable historical perspective on AI entrepreneurship, drawing on the personal experiences of three pioneers from different eras. Jerry Kaplan’s account of the 1980s expert systems boom is particularly insightful, highlighting the technical and business challenges that led to the AI winter. Adam Cheyer’s story of Siri illustrates the importance of framing AI in terms of user value (‘do engine’) and the strategic decision to avoid the AI label during a period of skepticism. Daniela Rus brings a contemporary view, discussing the shift in academic culture towards entrepreneurship and the potential of liquid neural networks. The discussion is well-moderated, with questions that draw out comparisons and contrasts between eras. However, the panel is largely anecdotal, lacking rigorous data or formal citations. The speakers’ perspectives are informed but may be subject to hindsight bias. The adéquation between title and content is strong, as the panel directly addresses the theme of ’this time it’s different’ by examining similarities and differences across generations. The technical depth is moderate, suitable for a general audience but not delving deeply into the underlying technologies. Overall, the panel offers a credible and engaging overview of AI startup history, though it could benefit from more critical analysis of current AI hype.

206 words

Title / Content Match

The title accurately reflects the content: a panel discussion comparing AI startups across three generations, with a focus on the theme 'this time it's different'.

Quality & Reliability

8/10

The panel features three highly credible AI entrepreneurs with direct experience across different eras of AI, moderated by a CHM curator. The discussion is grounded in historical context and personal experience, but lacks formal citations or data. The information is reliable but primarily anecdotal and retrospective.

Key Moments

Cited Sources

  • Startup: A Silicon Valley Adventure — Jerry Kaplan's book about his experience with GO Corporation.
  • Knowledge Navigator — Apple's concept video that inspired Siri.

Concurring Sources

  • AI winter — Supports the discussion of AI booms and busts.
  • Expert systems — Relevant to Jerry Kaplan's description of Teknowledge.

Dissenting Sources

  • None — No discordant sources were identified in the discussion.

Contribution & Novelties

The panel provides a unique comparative perspective on AI entrepreneurship across three distinct eras, offering insights into the cyclical nature of AI hype and the strategies used to navigate it. The discussion highlights the importance of framing AI in terms of practical value and the challenges of building sustainable business models. For those interested in exploring further, the following resources are relevant:

Pour aller plus loin :

  • AI winter — Historical context on the periods of reduced funding and interest in AI.
  • Expert systems — Overview of the rule-based systems that drove the 1980s boom.
  • Liquid neural networks — Daniela Rus’s research area, offering a flexible alternative to traditional neural networks.

111 words

Radar Profile

The radar profile shows high scores in quality of information and reliability, reflecting the credibility of the speakers and the historical depth. The quantity of information is moderate, and the technical level is moderate, indicating a balance between depth and accessibility.

Reliability 8/10

💬 No comments were provided for analysis.