AGI Update

GuardFall: When Decades-Old Shell Injection Tricks Beat Modern AI Safety Guardrails

Decades-old shell injection tricks can bypass modern AI safety guardrails, highlighting the need for more robust security measures

Dev.to · Cor E

What 116 court judgments taught me about the limits of AI

Learn about the limitations of AI in professional settings through an analysis of 116 court judgments and a personal project using consumer AI tools

Medium · AI

Your ChatGPT History Is a Liability. I Fixed That With a $80 Chip and a Pi5.

Protect your ChatGPT history from being used as evidence with a simple hardware solution using a $80 chip and a Pi5

Medium · AI

Your Skepticism About AI Is an Asset. Here’s How to Use It.

Learn to leverage skepticism about AI to improve its adoption and implementation in your team and organization, and why it matters for responsible AI development

Medium · Programming