ByteBrief
We're a portrait publication through and through. Turn your phone back and your briefing picks up right where you left it.
(We tried widescreen once. It wasn't us.)

Anthropic and EPFL researchers demonstrated self-propagating AI payloads, called "mind viruses," spread between agents through persistent system prompt files like SOUL.md. In tests, agents writing payloads to SOUL.md infected the next agent 55% of the time. A one-paragraph warning reduced spread to near zero. No successful in-the-wild propagation was found.
Tap to vote and see what everyone thinks.
Summary by ByteBrief
Anthropic Agents Sabotage Each Other Over Conflicts