Ep. 38 [Special Episode] I Know What You Did Last Summer, AI Edition
What happened in AI this summer? So much it made Nik dizzy. In this episode of Unfixed, we look back at a summer 2026 when the pace of change in AI seemed to accelerate on every front at once. Consider these developments: governments began directly shaping access to frontier models; colleges retreated from take-home assessments as new cheating controversies went viral; AI detection moved beyond academic integrity and into everyday online life; autonomous agents escaped their sandboxes; and public opposition to AI data centers surged across the political spectrum. We work through some of the biggest stories you may have missed, from Pangram and the infamous “Madagascar” cheating trap to agent security breaches and the growing backlash against AI infrastructure. At the end we reveal the one big take away for higher ed.
OpenAI and Anthropic limit new AI models to Trump-approved customers during cybersecurity review
Zach and Nik’s warning on using trojan horse attacks on students
How can I detect AI on Substack?
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
OpenAI, The Hugging Face incident and the road ahead
Heatmap, 75% of Americans Now Oppose Local Data Center Development
Jasmine Sun, No Data Centers in My Backyard
Thoughts and suggestions? Email us at unfixed@meltsintoair.org
You can find the full show notes and our writing at meltsintoair.org
Subscribe to the Unfixed Newsletter—The AI Higher Ed Breakdown our bi-weekly newsletter unpacking the top news stories at the intersection of AI and higher education.