2025-07-17 · auto-analysis (draft)
Ex-Google CEO: What Artificial Superintelligence Will Actually Look Like w/ Eric Schmidt & Dave B
The podcast discusses the timeline and impact of digital superintelligence, the energy needs of AI, and the shift towards nuclear power by tech companies, with Eric Schmidt and Peter Diamandis as the main speakers. The podcast discusses the impact of AI on various industries, including enterprise automation, the potential for superintelligence, and the redefinition of AGI. It also provides specific numbers to illustrate the value and scale of AI applications. The podcast discusses the potential impact of AI on national security, the challenges of regulating AI models, and the competition between the US and China in the AI domain. Discussion on the geopolitical implications of AI development, including the need for mutual assured destruction, tracking of AI chip locations, historical lessons from the Cold War, and concerns about model distribution and security. Discussion on the challenge
▶ Watch on YouTube
Topics AI and Superintelligence AI ▶ 0:00 Eric Schmidt discusses the timeline and impact of digital superintelligence.
Energy and AI Energy ▶ 2:47 Discussion on the energy needs of AI and the shift towards nuclear power by tech companies.
Computing and Chips Compute ▶ 5:30 Discussion on the energy efficiency of chips and the need for more efficient computing solutions.
Future of AI AI ▶ 8:58 Peter Diamandis discusses the potential for AI to generate intellectual power and abundance.
AI AI ▶ 9:11 Discusses 10 major tech trends transforming industries over the decade ahead.
AI in Enterprise AI ▶ 11:09 Describes how AI can automate enterprise tasks, potentially disrupting 100,000 enterprise software companies.
Superintelligence AI ▶ 14:44 Predicts the timeline for specialized AI scientists in every field within five years and the potential for superintelligence.
AGI AI ▶ 16:35 Discusses the redefinition of AGI and the potential for models to discover complex theories like relativity.
National Security Geopolitics ▶ 18:11 Concerns about AI becoming a national emergency due to potential biological and cyber attacks.
Chip Bans and Competition Economy ▶ 20:09 Discussion on the impact of chip bans on competition with China and the role of startups in innovation.
Regulation and Open Source Economy ▶ 21:00 Discussion on the regulation of AI models and the challenges of maintaining open source models while preventing proliferation.
AI and geopolitical implications Geopolitics ▶ 26:17 Discussion on the potential for AI to become a geopolitical threat and the need for mutual assured destruction to prevent conflict.
AI chip tracking Compute ▶ 27:00 Recommendation for governments to track AI chip locations and their activities.
Historical and strategic lessons Geopolitics ▶ 28:00 Comparison of current AI race to the Cold War and lessons from historical conflicts to prevent future wars.
Predictions made open Eric Schmidt : The computing needs for AI will come from traditional energy suppliers.
“The computing needs that we name now are going to come from traditional energy suppliers.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Peter Diamandis : Specialized AI scientists in every field within five years.
“That's pretty much in the bag as far as I'm concerned.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Peter Diamandis : AGI will redefine the concept of intelligence beyond the sum of what humans can do.
“That's when we have AGI.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Unknown : China will be able to use architectural changes to build models as powerful as those in the US, even with chip restrictions.
“Will the Chinese be able to use even with um chip restrictions, will they use architectural changes that will allow them to build models as powerful as ours?”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : AI development could lead to a cyber attack to slow down competitors.
“Under this scenario, I would be highly tempted to do a cyber attack to slow you down.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : AI models will be distributed globally, leading to potential security issues.
“If the structure of the world in 5 to 10 years is 10 models... and those models are data centers that are multi-gigawatts.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : AI's ability to generate its own scaffolding will be a 2025 thing.
“pretty much sure that that will be a 2025 thing at least from from their point of view.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : The system of knowledge will be much more distributed, with many smaller but nearly as intelligent models.
“So the system that's on your future phone may be, you know, three orders of magnitude, four order magnitude smaller than the one at the very tippy top, but it will be very, very powerful.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : The second and third scaling laws of machine learning will become highly valuable in the future.
“It's too new. It's too powerful. And at the moment, all of these businesses are incredibly highly valued.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : Blockbusters will still be produced by people with significant AI assistance.
“Um, I think blockbusters are likely to still be put together by people with an awful lot of help from by AI.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : Super empathetic voices with any inflection will be available in the next two months.
“When they see that which will be in the next probably two months.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
open Eric Schmidt : There will be 10 meta-Googlesized companies founded on the principle of learning loops.
“And so, it's likely to me that there will be another 10 fantastic Google scale meta-cale companies.”
Your call:
⏩ Sooner
🎯 On time
🐢 Later
🚫 Never
Numbers that matter 1 gawatt One gawatt is equivalent to one big nuclear power station needed for AI revolution in the US. 10 to $1,000 Value of voice customer service and sales conversations. 10 million Number of concurrent phone calls expected to move to AI in the next year or so. 100,000 Number of enterprise software companies that could be disrupted by AI. 5 years Timeline for specialized AI scientists in every field. very close Major biological understandings are very close. 10 to the 26 flops The threshold above which AI models need to be regulated according to the Biden administration. 200,000 GPUs Estimated number of GPUs used to train a supercomputing cluster for AI model Grock. 2025 Predicted year for AI's ability to generate its own scaffolding. 2.0 The birth rate in India is now down to 2.0 children per two parents. 0.7 The birth rate in Korea is now down to 0.7 children per two parents. 1.0 The birth rate in China is now down to 1.0 child per two parents. 1 The number one shortage in jobs right now in America are truck drivers, a low-status, hard, and low-paying job. 12 minutes Time taken to generate a paper using supercomputers. 2004 Year of the Google IPO prospectus.