### Beyond the Hype: A Look Ahead at Infrastructure with GenAI and What Lies Ahead
Just last week, **Ars Technica** Editor-in-Chief **Ken Fisher** and I ventured to sunny **San Jose** to launch an engaging event named *”Beyond the Hype: A Look Ahead at Infrastructure with GenAI and What Lies Ahead,”* organized in collaboration with **IBM**. This gathering attracted a crowd of eager Ars readers and tech experts to explore the future of infrastructure, artificial intelligence (AI), security, and environmental sustainability. For those who were unable to join us, stay tuned—we’ll be hosting another event next month in **Washington, DC**. Detailed information will be provided at the end of this article.
The gathering took place at the **Computer History Museum** (CHM), an ideal setting for a conversation focused on the path of technology. As Ken Fisher mentioned in his opening remarks, “Today’s lineup of speakers and topics illustrates the intricate and swiftly changing tech environment we all navigate.” The discussions spanned the potential of generative AI (GenAI) as well as the challenges it poses regarding infrastructure requirements, security risks, and environmental consequences.
### The Panels
#### 1. **Environmental Consequences of Growing Data Centers**
The initial panel addressed the environmental repercussions of data centers, particularly as they continue to grow to facilitate AI services. The panel featured **Jeff Ball**, scholar-in-residence at Stanford University’s **Steyer-Taylor Center for Energy Policy & Finance**; **Joanna Wong**, solutions architect for AI & Storage at IBM; and **Dr. John Timmer**, Senior Science Editor at Ars Technica.
A significant takeaway from the conversation was Jeff Ball’s statement that “not all energy is created equal.” He elaborated that the ecological impact of cloud resources significantly hinges on the geographical location of data centers. For instance, a data center in **Iceland** might operate using geothermal energy, while one in **China** could depend on coal, resulting in substantially different carbon footprints.
Joanna Wong pointed out that infrastructure often experiences “unknown failure points”—situations that might not lead to immediate breakdowns but still waste additional computational resources and energy. She highlighted the necessity of pinpointing these inefficiencies to minimize waste and enhance performance.
#### 2. **Security Threats and AI-Generated Code**
The second panel shifted gears to the changing landscape of security threats, particularly relating to AI-generated or AI-validated code. The panel included **Stephen Goldschmidt**, Global Platform Security Architect at Box; **Patrick Gould**, Director of the Cyber & Telecom Portfolio for the **Defense Innovation Unit** (DIU) at the Department of Defense; and **Ram Parasuraman**, Executive Director for Data & Resiliency at IBM.
This discussion covered concerns surrounding AI-generated code, especially the propensity of large language models (LLMs) to “confabulate” or produce erroneous information. Nevertheless, the panelists concurred that the most fitting role for AI in coding would be to complement human endeavors rather than replace them. AI can assist in identifying potential vulnerabilities, such as typographical errors, and refine code after human programmers have finished. As Ram Parasuraman noted, “The issue of how to trust AI output will persist. What will evolve are the methods we employ to verify and oversee that output.”
#### 3. **Strategizing for Infrastructure Sustainability**
The concluding panel focused on the long-term planning of infrastructure, specifically in anticipating unforeseen challenges. The panelists included **Ashwin Ballal**, Chief Information Officer at Freshworks; **Karun Channa**, Director of Product AI at Roblox; and **Pete Bray**, Global Product Executive at IBM.
The panelists stressed the significance of thorough requirement gathering, resilience, and adaptability in infrastructure planning. Proper planning of requirements is essential for constructing resilient systems capable of managing unexpected demand surges. Moreover, maintaining emergency operational funds can provide the flexibility necessary to tackle unforeseen obstacles.
### The Celebration
Following the panels, attendees were invited to network and explore CHM’s permanent exhibit, *”Revolution: The First 2000 Years of Computing.”* Special thanks to long-time Ars contributor **Brian Won** for guiding an informal tour of the exhibit. The evening was alive with engaging discussions, cocktails, and a shared enthusiasm for technology.
### Coming Soon: Washington, DC
If you were unable to make it to the San Jose event, fear not—we’ll be doing it again next month in **Washington, DC**. On **October 29**, we’ll host an event at the **International Spy Museum** titled *”AI in DC: Privacy, Compliance, and Enhancing Infrastructure.”* This gathering will focus on:
– Navigating compliance with budding technologies
– Safeguarding data in the era of AI-driven cyber-espionage
– Optimal infrastructure solutions for AI/ML strategies
We’re in the process of finalizing the lineup of speakers, but you can express your interest in attending [
