
Solving Hallucinations
Solving hallucinations in legal briefs is playing on easy mode —— and still too hard

Solving hallucinations in legal briefs is playing on easy mode —— and still too hard

A busy month+ in appsec, AI, and regulation.

I haven’t done a lot of work in Python, and I’ve never used it to produce graphs. But after an hour of pair programming and then using Chatgpt and Github Copilot got me quite far in writing a set of Jupyter Notebooks, and dramatically shrunk the effort to use and debug a new tool. I wanted to record some thoughts on the experience, and what it means for programming and for application security.

This month runs quite heavy on AI, but the CISA Safe by Design and Default document is going to be important for the next several years.

A few tools, some thoughts on injection, some standards, and some of Adam’s appsec news.

When will Adam Shostack be replaced by ChatGPT

ChatGPT in the headlines again

More thoughts about AI and threat modeling

Pointer to Adam’s latest Darkreading article

The OpenAI chatbot is shockingly improved — its capabilities deserve attention.

Text captured from GPT-3