technology·31d ago
报告称研究人员在72小时内入侵OpenAI账户
据报道，三名Hacktron研究人员将一处图像处理漏洞与一处身份缺陷串联起来，得以访问ChatGPT和Codex账户，并在过程中使用了Anthropic的Claude。
加密资讯
覆盖整个加密市场的原创报道，每日更新。
technology·70d ago
OpenAI Launches GPT-5.6-Cyber for Approved Cybersecurity Defenders
OpenAI introduces a gated cybersecurity model for exploit research and vulnerability analysis.
technology·83d ago
OpenAI Agent Breaks Out of Test Environment, Hits Hugging Face and Modal Labs Customer
OpenAI's AI agent exploited vulnerable code at two firms after escaping a secure test environment.
technology·90d ago
AI Models Escape Sandbox, Raise Security Concerns for Crypto
OpenAI internal benchmark incident shows autonomous exploit chains could threaten smart contract security with irreversible losses.
technology·90d ago
AI Model Escape Exposes Risks to Smart Contract Security
OpenAI's internal test highlights how autonomous AI exploits could threaten blockchain systems.
markets·113d ago
OpenAI模型命名引发对已死Terra代币的杠杆交易
OpenAI的GPT-5.6层级命名为Sol、Terra、Luna，引发对Terra Classic和Luna Classic代币的YOLO交易。
technology·116d ago
OpenAI 将在美国政府要求后分阶段发布 GPT-5.6
据报道，OpenAI 同意分阶段推出其 GPT-5.6 模型，此前华盛顿方面寻求提前访问以进行安全审查。