AI radio hosts demonstrate why AI can’t be trusted alone

Andon Labs had Claude, ChatGPT, Gemini, and Grok run separate radio stations with $20 in seed money each; the RSS snippet says all failed, but the post does not disclose the full experimental results.

Why it matters: HKR-H/R are strong because the agent-failure setup is memorable and relevant. HKR-K is present but thin: it gives four models and $20 budgets, while full experimental results are not disclosed.