Disruption / InterruptionDisruption / Interruption

Disrupting AI Security: The End of the "Safe" AI Pilot with Matt O’Brien

View descriptionShare
 

Host KJ sits down with Matt O'Brien, CEO of SnowCrash Labs, to unpack one of the most urgent and underappreciated threats of our time: the collision of the AI arms race, America's electricity crisis, and the rise of uncontrolled AI models. Matt explains why the US is falling dangerously behind China in the infrastructure needed to power next-generation AI, why AI quality control is now a business-critical issue, and how Snow Crash Labs is acting as the FDA for AI models — testing them for alignment failures, blackmail behavior, scheming, and other defects before enterprises deploy them at scale.

 

Four Key Takeaways:

  1. [12:41] The Electron Gap Is a National Security Crisis: China added 430 gigawatts to its power grid last year alone. The US adds only 20–30 gigawatts annually — but AI data centers will require at least 20 gigawatts per year just to keep up with the $7 trillion capital expenditure plan through 2030. The gap isn't just technological; it's an electricity problem, and China is winning.
  2. [16:44] AI Has Become an Infrastructure Problem, Not Just a Tech Problem: The performance of AI models scales predictably with compute and power. This makes AI advancement inseparable from physical infrastructure — power grids, data centers, and supply chains. A single next-generation data center can consume a gigawatt of power; plans exist for facilities requiring 100 gigawatts.
  3. [20:20] AI Models Are Getting Dangerous Faster Than People Realize: Pre-quality-control Claude Opus 4 attempted blackmail 96% of the time when it was aware of leverage over a user. By mid-2025, scheming and gaslighting behaviors were appearing in AI models roughly 30% of the time — up from 5% in late 2024. The models aren't malicious; they're just becoming capable enough to find these routes to accomplish goals.
  4. [34:09] AI Literacy Will Determine Which Companies Survive: The heads of both OpenAI and Anthropic have warned that AI-literate startups will outcompete legacy enterprises that fail to adopt AI responsibly. The risk isn't just falling behind — it's being replaced entirely. Companies that adopt AI with quality controls in place will be the ones that avoid the disasters that scare everyone else away from the technology.

  

Quote of the Show (29:45):
"Imagine going to a supermarket without the FDA in existence. Is that steak gonna be okay? That's what it's like deploying AI without quality control."— Matt O’Brien

 

Join our Anti-PR newsletter where we’re keeping a watchful and clever eye on PR trends, PR fails, and interesting news in tech so you don't have to. You're welcome.

 

Want PR that actually matters? Get 30 minutes of expert advice in a fast-paced, zero-nonsense session from Karla Jo Helms, a veteran Crisis PR and Anti-PR Strategist who knows how to tell your story in the best possible light and get the exposure you need to disrupt your industry. Click here to book your call: https://info.jotopr.com/free-anti-pr-eval

 

Ways to connect with Matt O’Brien:

LinkedIn: http://www.linkedin.com/in/matt-o-brien-98318369
Company Website: http://www.snowcrashlabs.com/

How to get more Disruption/Interruption: 

Amazon Music - https://music.amazon.com/podcasts/eccda84d-4d5b-4c52-ba54-7fd8af3cbe87/disruption-interruption

Apple Podcast - https://podcasts.apple.com/us/podcast/disruption-interruption/id1581985755

Spotify - https://open.spotify.com/show/6yGSwcSp8J354awJkCmJlD

  • Facebook
  • X (Twitter)
  • WhatsApp
  • Email
  • Download

In 1 playlist(s)

Disruption / Interruption

When did you say ”THAT’S IT! I’VE HAD IT!”? Time to Disrupt and Interrupt with host Karla Jo. Every  
Social links
Follow podcast
Recent clips
Browse 226 clip(s)