AIInterconnects AI49m ago
GLM-5.3's open release has yet
GLM-5.3's open release has yet to produce major attacks despite Anthropic's warnings about its Mythos-level cyber risk, undercutting calls to ban open models
TL;DROpen AI model released without predicted major cyberattacks, contradicting safety warnings.
Why it matters: Challenges arguments for banning open models; raises questions about AI risk assessment accuracy.
Open-weights, ideology, and acknowledging trade-offs. … Audio playback is not supported on your browser. Please upgrade.
Read full articleSource: Interconnects AI · Opens in new tab