All videos
0:00 / 0:00
policy

The Red Queen Problem with Anthony Scriffignano, PhD, Data Scientist

Modern CTO20 August 2026Watch on YouTube

What you'll learn

  • The 'Red Queen Problem' forces tech leaders to constantly innovate faster just to maintain their competitive position in AI adoption
  • LLMs see far less of the internet than most people assume, limiting their actual capabilities and increasing hallucination risks
  • The biggest mistake in enterprise AI is leading with technology rather than starting with the problem you want to solve
  • Data governance requires attention to four critical elements: privacy, permission, provenance, and mindset
  • CTOs should be cautious about giving AI agents access to sensitive systems like email due to security and reliability concerns

Frequently asked questions

What is the 'Red Queen Problem' Anthony discusses?
The Red Queen Problem describes how technology leaders must constantly accelerate their AI innovation and adoption just to maintain their current competitive position, similar to the character in Alice in Wonderland who must run fast just to stay in place.
How much of the internet can LLMs actually see?
Much less than people assume. Anthony emphasizes that most people wildly overestimate how much of the internet LLMs can actually access and understand, leading to unrealistic expectations and increased hallucination risks.
What are the four essential elements of data governance?
Anthony identifies the four Ps of data governance: Privacy, Permission, Provenance, and Mindset. These elements together are necessary for responsible data use in enterprise AI.
Why does Anthony advise against giving AI agents access to email?
AI agents with email access pose significant security and reliability risks. Anthony cautions that email is too sensitive a system to entrust to autonomous AI without strict controls in place.

Topics

enterprise aiai policyllm limitationscto strategydata governance

Read next

Description from the channel

Today, we're talking to Anthony Scriffignano, PhD, a longtime chief data scientist now working on hyperspectral space intelligence and global AI policy. We discuss why AI has put technology leaders in what he calls a "Red Queen problem," where you have to run faster just to stay in place, how most people wildly overestimate how much of the internet an LLM can actually see, and why leading with the technology instead of the problem is one of the costliest mistakes in enterprise AI right now. All of this right here, right now, on the Modern CTO Podcast! 00:00 – Cold Open: The Chair Keyboard Interface 00:37 – Catching Up: Life After Dun & Bradstreet 03:12 – New Jersey, DC, and The Sopranos 04:51 – Today’s Topic: What CTOs Get Wrong About AI 06:56 – The Red Queen Problem Explained 09:53 – Avoiding the Echo Chamber: Failing Smarter 12:31 – Different Kinds of Smart: The Salad Bar 14:07 – How Anthony Actually Uses LLMs Daily 17:01 – Why He Won’t Hand Agents His Email 19:10 – The Myth of AI Knowing Everything 20:37 – Veracity, Truth, and Hallucination 23:44 – Using AI for Blind Spot Detection 26:41 – Biggest Mistakes CTOs Are Making Now 28:34 – The Global AI Regulatory Patchwork 30:23 – Assume You’ve Already Been Infiltrated 32:32 – National Security vs. Privacy Worldwide 34:13 – US Federal vs. State Regulation Tangle 37:24 – The Four Ps of Data Governance 40:54 – Mindset: The Forgotten Fourth Lever 44:01 – Leading With “Help Me Understand” 46:43 – Startups, Hubris, and Vibe Coding 49:26 – QR Codes and the New Attack Surface 51:44 – Will Intelligence Get Locked Down? 53:34 – Star Trek vs. The Jetsons Future 55:32 – Technology, Free Time, and Human Choice 57:21 – Uneven Access and Lessons From COVID 1:00:04 – The Next Generation’s Undefined Path 1:00:47 – Parenting, Time, and Doing Better 1:03:34 – Kids Mimic Us, Not What We Teach 1:04:39 – The Vanishing Library and Future Research 1:06:31 – Spotting Disruption AI Can’t Predict 1:09:21 – Peer Reviewed With Johns Hopkins APL 1:10:58 – The Pentagon Pizza Tracker Example 1:12:19 – Why Missing Data Doesn’t Break the Model 1:13:52 – Closing Thoughts and Sign Off