-
サマリー
あらすじ・解説
Industrial Talk is onsite at OMG, Q1 Meeting and talking to Chuck Byers, CTO with Industry IoT Consortium about "Developing AI standards to insure market trustworthiness!" Chuck Byers, a seasoned industry professional, discussed the integration of AI and IoT technologies in the context of the Object Management Group (OMG) and the Digital Twin Consortium. Byers highlighted the importance of AI in real-time systems, emphasizing the need for prompt engineering and meta AI to ensure trustworthiness. He detailed the computational requirements of training large language models, noting that retraining the ChatGPT 3 model would require 195 NVIDIA DGX servers, costing $80 million and consuming 2.5 million watts. Byers also stressed the importance of sustainable energy sources and efficient cooling solutions for future data centers. Action Items
Outline Introduction and Welcome
Chuck Byers' Background and Contributions
Digital Twin Consortium and Industry IoT Consortium
Challenges and Opportunities in AI
- [ ] Explore ways to detect when AI results are not as trustworthy as they need to be, and develop methods to validate AI outputs
- [ ] Investigate the use of small modular nuclear reactors to power data centers for AI model training
- [ ] Promote the work of OMG and the Digital Twin Consortium in addressing the challenges of AI and the Internet of Things
Outline Introduction and Welcome
- Scott MacKenzie introduces the podcast and its focus on industry innovations and trends.
- Scott welcomes listeners and mentions the broadcasting location at the OMG SQ1 meeting in Reston, Virginia.
- Scott introduces Chuck Byers, highlighting his extensive background in electrical engineering and his contributions to Bell Labs and Cisco.
- Chuck Byers joins the conversation, expressing his pleasure in being invited and his admiration for Scott's hosting style.
Chuck Byers' Background and Contributions
- Chuck Byers shares his educational background, including a Master's degree in electrical engineering and teaching at the University of Wisconsin.
- He discusses his work at Bell Labs and his invention of a US patent for a device that alerts users to approaching badness.
- Chuck talks about his tenure at Cisco, where he worked on computer control, IoT, edge computing, and drone technology.
- He mentions his current role as Chief Technical Officer of the Industry IoT Consortium and its recent integration into the Digital Twin Consortium.
Digital Twin Consortium and Industry IoT Consortium
- Chuck explains the Digital Twin Consortium's focus on pairing simulation models with real-time physical systems.
- He describes the Industry IoT Consortium's focus on sensors, actuators, and the computation necessary for IoT systems.
- Chuck highlights the synergy between the two consortiums, emphasizing their combined potential to sense, network, and actuate the physical world into the digital world.
- He provides an example of using simulation models to optimize oil refinery processes without physical experimentation.
Challenges and Opportunities in AI
- Scott MacKenzie and Chuck Byers discuss the challenges of AI, particularly the lack of standards and trustworthiness.
- Chuck explains the different types of AI, including machine vision and generative AI, and the unique challenges of generative AI.
- He describes the use of large language models in generative AI and their applications in tools like ChatGPT and Microsoft 365.
- Chuck emphasizes the importance of prompt engineering and fact-checking to ensure