AI Insights with Bill Frischling: Deep Dive on Recent Developments
Feb. 24, 2025
STACK Cybersecurity welcomed back Bill Frischling, Distinguished Scientist & VP at FiscalNote, to our weekly AI discussion with Managed Security Service Providers (MSSPs) from throughout the U.S. and Canada. Bill's expertise and entrepreneurial background make him an invaluable resource as we navigate the rapidly evolving AI landscape.
HIGHLIGHTS FROM OUR DISCUSSION
Deep Seek Controversy
Bill shared his perspective on Deep Seek, noting that while there's been controversy around its development, it represents another option that's "cheaper, faster" and gets you "within six months of the state-of-the-art." He emphasized that while the open-source version has been extensively tested and appears safe, businesses should exercise caution with the hosted model from Hangzhou.
Multimodality Becoming Standard
One of the most exciting developments is how multimodal capabilities are becoming commonplace in leading AI models. Bill highlighted practical applications where AI can analyze screenshots of web designs and immediately identify adjustments needed for alignment, margins, and other visual elements.
Data Privacy Concerns
For business owners concerned about data privacy, Bill emphasized the importance of using paid versions of AI tools with proper privacy settings enabled. He warned against using "fly by night" aggregator services that offer suspiciously cheap access to multiple AI models, as their business models often rely on data collection or unrealistic usage estimates.
Practical Business Applications
Bill shared fascinating examples of how AI is transforming workflows, including:
- Creating custom accounting agents that process invoices and financial transactions
- Using deep research capabilities to quickly analyze SEC filings and corporate data
- Generating complex commands for video processing that would take hours to write manually
Future Challenges
The discussion touched on upcoming legal challenges around copyright and intellectual property in AI. As Bill noted, we're likely "a decade plus away" from resolving the philosophical and legal questions around machines learning from copyrighted content.
Looking Ahead
As AI continues to evolve at breakneck speed, Bill's insights remind us to stay pragmatic, focus on real-world applications, and maintain appropriate security measures. For MSPs and business leaders, the key takeaway is balancing innovation with caution as we incorporate these powerful tools into our workflows.
Related Information
Use these resources to learn more about artificial intelligence.
- Start Your AI Journey, or Get Left Behind STACK blog post exploring the myriad ways AI is reshaping the business landscape and what companies can do to harness its potential responsibly.
- Report Highlights Deficiencies in Training, Security Technologies The “Cybersecurity Attitudes and Behaviors, Oh Behave! 2024-2025” report highlights the importance of robust cybersecurity measures and the risks posed by human behavior.
- Phishing 2.0: How AI is Making Cyber Attacks Smarter and More Dangerous STACK blog post details the evolution of phishing and how AI has amplified this type of cyberattack.
- AI and Machine Learning, The Worldwide Problem No One is Discussing This article is an engineering/MBA technical case analysis, based on published worldwide events.
Want to learn more? Watch the video with Bill Frischling. Call STACK Cybersecurity at (734) 744-5300 or Contact Us to explore practical applications of artificial intelligence for your business.