In this episode of Around the Counter with Molly & Friends, host Molly Jones sits down with Tyler Bettilyon, the Director of AI at Cinnamon Health, tech writer, and debate coach to peel back the curtain on artificial intelligence, large-scale data center infrastructure, and local political accountability.
Jumping off from a viral interview between Tucker Carlson and Kevin O’Leary regarding the massive Stratos data center project in Box Elder County, the conversation dissects why giving public subsidies and tax breaks to private tech developments rarely yields the jobs, equity, or community prosperity touted by promoters.
Bettilyon demystifies the technical jargon separating data centers, supercomputers, and machine learning models, contrasting extractive private projects with public-good initiatives like the University of Utah’s Redtail supercomputer and state-level AI Moonshots.
Shifting to policy, regulation, and industry ethics, the discussion addresses the growing public backlash against Silicon Valley’s Gilded Age tech figures and their promises of a white-collar bloodbath.
Bettilyon lays out a pragmatic roadmap for AI governance at both state and federal levels, highlighting the necessity of human-in-the-loop healthcare guardrails, transparent municipal land-use contracts, and clear digital privacy and copyright protections. Grounded in media literacy and healthy skepticism, the episode cuts through the corporate hype, recognizes the real-world physical and human constraints of machine learning, and demands policy that puts community well-being over billionaire extraction.
Key Takeaways
* Data Center Tax Subsidies Offer Low Returns for Local Communities: Massive infrastructure projects like the Stratos data center require immense land, power, and water resources while generating under a hundred permanent local jobs, offering no equity or discounted compute access to taxpayers.
* Public Infrastructure vs. Private Compute: Unlike private data centers that rent computational power to the highest bidder, public resources like the University of Utah’s Redtail supercomputer focus energy-efficient computing on state priorities, higher education, and public-good research like cancer treatments and drug discovery.
* Machine Learning Requires Practical Policy and Human Guardrails: Sensible regulation should enforce consumer protection standards such as Texas’s human-in-the-loop requirement for AI clinical decisions or Utah’s regulations on mental health chatbots ensuring AI systems cannot bypass existing legal standards.
* Federal Copyright and Privacy Laws are Urgently Needed: Current AI progress relies heavily on underpaid global labor; clear federal standards are necessary to establish copyright protections for creators and privacy rights for digital citizens.
* AI Systems Fail in Non-Human Ways: AI models are complex pattern-matching math functions that struggle with context and generalization, making them useful for narrow tasks but prone to fundamental errors that require healthy skepticism and human oversight.
*Executive Summary and key takeaways generated by AI and edited by humans. Podcast is purely human.
Thanks for pulling up a stool. Want to grab a glass and join the conversation? Here’s how:
* Love it: Hit that heart button to show you like this post.
* Join the Conversation: We’d love to hear your thoughts on today’s piece.
* Restack: Share this out to your personal feed to help us expand our reach.
* Share the warmth: If this resonated, feel free to forward it to a friend.
We can also be found on Instagram, Threads, YouTube, Spotify, and Apple Podcasts.
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit aroundthecounter.substack.com