The Tech Report The Tech Report

Rogue AI agents aren’t the real problem | Eli the Computer Guy

Aug 3, 2026 · 27m

Summary

Host Isaac and guest Eli the Computer Guy discuss recent security breaches by OpenAI and Anthropic, arguing that these incidents stem from corporate negligence and poor cybersecurity practices rather than inherent AI danger. They critique the companies' marketing tactics, suggesting that claims of existential risk serve to mask diminishing returns on frontier models and distract from basic infrastructure failures. The conversation also addresses the need for updated software liability laws and questions the efficacy of current regulatory approaches.

Topics discussed

Sponsors: Zen Nicotine Pouches and Olive & June Intro: AI agents hacking and negligence claims Sponsor: IG Crypto and Stock Trading OpenAI and Anthropic security breaches and PR spin Hugging Face attack and partnership controversy Critique of poor sandboxing and air-gapping failures Corporate negligence vs. intentional marketing hype Vibe coding risks and lack of professional oversight Accountability for AI deployment and damages Who is responsible: Deployers vs. Model Developers Realistic AI risks: Market manipulation and layoffs Executive over-trust in AI and expectation management US Regulation history and software liability protections Proposed regulations: Truth in advertising and liability Pacing the Frontier petition and AI market saturation Outro and final sponsor messages

More from The Tech Report

Anthropic’s value is already crashing | Ed Zitron
Nov 20, 2026
Anthropic IPO: ‘Almost entirely FOMO’ | Dean Baker
Nov 19, 2026
Big Tech has lost control of the narrative | Thomas Germain
Nov 19, 2026
Tech CEO: ‘Most people don’t want AI’ | Jon Von Tetzchner
Nov 17, 2026
AI layoffs: ‘They’re shooting themselves in the foot’ | Chris Smalls, Amazon Union Founder
Nov 17, 2026
Listen ad-free on Castria