Comment Comment by piker > I'm flabbergasted that Anthropic and OpenAI aren't more worried about these attack vectorsYep. We tricked them both trivially with malicious fonts in Docx files. Documented it here: https://tritium.legal/blog/norobotoI wonder if prompt injection (and the thousands of vectors for hiding injection attempts) is actually un solvable. Discussing it may be existential to the business model. parent piker 5 months ago
Yep. We tricked them both trivially with malicious fonts in Docx files. Documented it here: https://tritium.legal/blog/noroboto
I wonder if prompt injection (and the thousands of vectors for hiding injection attempts) is actually un solvable. Discussing it may be existential to the business model.