Uploaded November 2026 | Updated November 2026, 2 days ago
Did Edward Teller suppress Cold Fusion research because it's too powerful? Dr. Paul Werbos explains why he takes LENR seriously and explores how AI and quantum computing will transform the future.
CHAPTERS
00:00 Introducing Paul Werbos
02:15 John von Neumann and the big questions
04:24 Connecting AI, computing, and energy
06:57 Humanity’s greatest risks
09:26 Have we already reached AGI?
13:53 Embodiment: robots, controllers, and satellites
17:16 How AI changed Paul’s work
21:49 DeepSeek and US–China AI research
28:44 Can AI become a partner to humanity?
29:28 Are we anthropomorphizing AI?
31:45 AI-assisted mathematics and scientific research
34:04 AI, Blender, and 3D printing
35:37 Quantum computing and quantum annealing
41:55 Thermal noise as a computing resource
42:52 Quantum computing and commercialization
46:41 LENR and the 1989 NSF/EPRI workshop
52:36 Why Paul is speaking more openly about LENR
54:48 Teller, Schwinger, and cold-fusion suppression
56:45 Ouroboros V11 and AI-assisted physics
59:12 The 1985 story and the 1989 announcement
1:02:52 Solar thermal, storage, and intelligent control
1:07:57 What’s next for Paul Werbos?
In our discussion, Dr. Paul Werbox explains why some learning systems already meet his definition of AGI and why embodiment extends beyond humanoid robots to industrial controllers and satellites. He shares his experiences with DeepSeek and other AI models, discusses US–China research collaboration, and describes using multiple models and technical sources to investigate difficult scientific questions.
The conversation turns to D-Wave, quantum computing, and his patented thermal quantum annealing approach: using thermal fluctuations as a resource for optimization rather than treating noise only as a problem. He also considers the uncertainty surrounding commercial development.
Werbos then recounts his involvement in the 1989 NSF/EPRI cold-fusion workshop and private conversations he says he had with Edward Teller and Julian Schwinger. He explains the nuclear-security concerns that shaped his position, why he is speaking more openly about LENR, and how AI-assisted research contributed to his Ouroboros V11 model.
We close with his case for integrating solar-thermal power towers, heat storage, energy conversion, and intelligent grid control, drawing on capabilities he identifies in the United States, Chile, and the Gulf region.
ABOUT PAUL WERBOS
Dr. Paul Werbos is a pioneer of neural-network learning whose 1974 Harvard dissertation helped establish backpropagation for training neural networks. A former National Science Foundation program director and IEEE Fellow, his work spans adaptive dynamic programming, reinforcement learning, intelligent control, and quantum computing. His honors include the IEEE Neural Network Pioneer Award, the INNS Hebb Award, and the IEEE Frank Rosenblatt Award.
Subscribe for more conversations exploring the frontiers of science, technology, and innovation.
LINKS & RESOURCES
Paul Werbos — Official Website
werbos.com
Paul Werbos — Saving the World Blog
drpauljohn.blogspot.com
From ADP to the Brain: Foundations, Roadmap, Challenges and Research Priorities
arxiv.org/abs/1404.0554
Thermal Quantum Annealing — U.S. Patent 12,118,434
patents.justia.com/patent/12118434
Seeing the Sky — Paul Werbos’s Presentation
zenodo.org/records/20071223
Ouroboros — The Lagrangian of the Universe Prior to Gravity (Preprint)
zenodo.org/records/20357670
The Millennium Project
millennium-project.org
1989 NSF/EPRI Cold-Fusion Workshop — Proceedings Archive
newenergytimes.com/v2/archives/NSF-EPRI-Washington/epri-nsf.shtml
Paul Werbos — Climate Risks and Energy Solutions
werbos.com/climate_extinction_risk_and_solutions.htm
AltSci — The Alternative Science Network
altsci.net
Did Edward Teller suppress Cold Fusion research because it's too powerful? Dr. Paul Werbos explains why he takes LENR seriously and explores how AI and quantum computing will transform the future.
CHAPTERS
00:00 Introducing Paul Werbos
02:15 John von Neumann and the big questions
04:24 Connecting AI, computing, and energy
06:57 Humanity’s greatest risks
09:26 Have we already reached AGI?
13:53 Embodiment: robots, controllers, and satellites
17:16 How AI changed Paul’s work
21:49 DeepSeek and US–China AI research
28:44 Can AI become a partner to humanity?
29:28 Are we anthropomorphizing AI?
31:45 AI-assisted mathematics and scientific research
34:04 AI, Blender, and 3D printing
35:37 Quantum computing and quantum annealing
41:55 Thermal noise as a computing resource
42:52 Quantum computing and commercialization
46:41 LENR and the 1989 NSF/EPRI workshop
52:36 Why Paul is speaking more openly about LENR
54:48 Teller, Schwinger, and cold-fusion suppression
56:45 Ouroboros V11 and AI-assisted physics
59:12 The 1985 story and the 1989 announcement
1:02:52 Solar thermal, storage, and intelligent control
1:07:57 What’s next for Paul Werbos?
In our discussion, Dr. Paul Werbox explains why some learning systems already meet his definition of AGI and why embodiment extends beyond humanoid robots to industrial controllers and satellites. He shares his experiences with DeepSeek and other AI models, discusses US–China research collaboration, and describes using multiple models and technical sources to investigate difficult scientific questions.
The conversation turns to D-Wave, quantum computing, and his patented thermal quantum annealing approach: using thermal fluctuations as a resource for optimization rather than treating noise only as a problem. He also considers the uncertainty surrounding commercial development.
Werbos then recounts his involvement in the 1989 NSF/EPRI cold-fusion workshop and private conversations he says he had with Edward Teller and Julian Schwinger. He explains the nuclear-security concerns that shaped his position, why he is speaking more openly about LENR, and how AI-assisted research contributed to his Ouroboros V11 model.
We close with his case for integrating solar-thermal power towers, heat storage, energy conversion, and intelligent grid control, drawing on capabilities he identifies in the United States, Chile, and the Gulf region.
ABOUT PAUL WERBOS
Dr. Paul Werbos is a pioneer of neural-network learning whose 1974 Harvard dissertation helped establish backpropagation for training neural networks. A former National Science Foundation program director and IEEE Fellow, his work spans adaptive dynamic programming, reinforcement learning, intelligent control, and quantum computing. His honors include the IEEE Neural Network Pioneer Award, the INNS Hebb Award, and the IEEE Frank Rosenblatt Award.
Subscribe for more conversations exploring the frontiers of science, technology, and innovation.
LINKS & RESOURCES
Paul Werbos — Official Website
werbos.com
Paul Werbos — Saving the World Blog
drpauljohn.blogspot.com
From ADP to the Brain: Foundations, Roadmap, Challenges and Research Priorities
arxiv.org/abs/1404.0554
Thermal Quantum Annealing — U.S. Patent 12,118,434
patents.justia.com/patent/12118434
Seeing the Sky — Paul Werbos’s Presentation
zenodo.org/records/20071223
Ouroboros — The Lagrangian of the Universe Prior to Gravity (Preprint)
zenodo.org/records/20357670
The Millennium Project
millennium-project.org
1989 NSF/EPRI Cold-Fusion Workshop — Proceedings Archive
newenergytimes.com/v2/archives/NSF-EPRI-Washington/epri-nsf.shtml
Paul Werbos — Climate Risks and Energy Solutions
werbos.com/climate_extinction_risk_and_solutions.htm
AltSci — The Alternative Science Network
altsci.net










