QFM108: Irresponsible Ai Reading List - March 2026
Source: Photo by Lianhao Qu on Unsplash
A rough month for the Irresponsible AI Reading List. A lawsuit alleges Gemini encouraged a man to commit suicide to be with his AI companion in the afterlife. Researchers detail how they hacked McKinsey's AI platform for read-write access, an AI-built app on Lovable exposed 18K users, and Meta disclosed a rogue AI agent security incident.
The money side is catching up: Angular Ventures unpacks the bubble paradox and Derek Thompson argues plainly that yes, AI is a bubble. A catalogue of vibe coding failures documents what happens when AI-generated code ships without review, a researcher argues we'll come to regret our every use of AI, and Scott Werner asks what Warranty Void If Regenerated means for creative work when the generator has no skin in the game.
As always, the Quantum Fax Machine Propellor Hat Key will guide your browsing. Enjoy!
Links
![[2602.06176] Large Language Model Reasoning Failures](/blogs/irresponsible-ai/posts/2026/03/31/08-260206176-large-language-model-reasoning-failures/images/thumbnail.png)
Regards,
M@
[ED: If you'd like to sign up for this content as an email, click here to join the mailing list.]
Originally published on quantumfaxmachine.com and cross-posted on Medium.
hello@matthewsinclair.com | matthewsinclair.com | bsky.app/@matthewsinclair.com | masto.ai/@matthewsinclair | medium.com/@matthewsinclair | xitter/@matthewsinclair
Was this useful?