GLM-5.3 open release sees no major attacks despite Anthropic warnings
#open-weights#ai-safety#glm-5.3#cyber-risk
Nathan Lambert of Interconnects AI reports that GLM-5.3's open release has not yet resulted in major cyberattacks, despite Anthropic's warnings about its Mythos-level cyber risk. This development undercuts arguments for banning open models, highlighting the trade-offs between open-weights distribution and security concerns.
Coverage timeline
Techmeme
Nathan Lambert / Interconnects AI : 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 — Open-weights, ideology, and acknowledging trade-offs. … Audio playback is not supported on your browser. Please upgrade.
