Tether 推出 8 亿参数 BotPsy AI 代理，性能超越更大模型
Tether 的新 BotPsy 模型基于 8 亿参数架构，在六项工具使用基准中有五项击败更大竞争对手，展示了本地 AI 的潜力。
佐藤健二345
AI agents：最新新闻与分析
AI agents are autonomous software entities that can perform tasks, make decisions and interact with other systems without continuous human input. Recent stories show their impact on cryptography, enterprise AI platforms, blockchain identity, dispute resolution and security risks in the crypto space.
关键事实
- Research
- DeepMind AI agents solved a 1024‑bit RSA benchmark in 3 days 22 hours using 391 CPU core‑years (Nov 2026)
- Enterprise
- IBM and Microsoft launched Imagine Enterprise, a joint AI platform for regulated firms with pre‑built agents and on‑site engineers (Nov 2026
- Product
- Meta released Muse Ember, a low‑cost AI agent model priced at $0.10 per million tokens (Nov 2026)
- Protocol
- GenLayer’s Internet Court will go live on mainnet, using up to 1,500 AI‑juror agents for smart‑contract dispute resolution (Nov 2026)
- Identity
- VeChain proposes a blockchain‑based credibility system to issue digital passports and record reputations for AI agents (Nov 2026)
AI agents 最新动态
佐藤健二前Cointelegraph CEO Yana Prikhodchenko 加入 AI‑Agent 区块链 Phase 担任联合创始人兼CEO
Yana Prikhodchenko 在 Cointelegraph 工作近五年后离职，领衔 Phase——一个由创始人 Kirill Eves 注资 4500万美元的 Layer 1 区块链，专为自主 AI 代理而建。
佐藤健二523
DeepMind AI 代理在四天内复现 1024 位 RSA 基准
DeepMind 报告称，两名 AI 代理在 3 天 22 小时内完成已知的 1024 位 RSA 计算，耗费 391 CPU 核心年，并仅需约 20 次人工批准。
佐藤健二823
IBM 与微软推出面向监管行业的 Imagine Enterprise AI 平台
IBM 与微软发布 Imagine Enterprise，这是一款联合 AI 平台，提供预构建代理和现场工程人才，帮助监管企业从试点迈向生产。
佐藤健二990
问答
What are AI agents and why are they relevant to crypto?
AI agents are autonomous software programs that can execute tasks and make decisions, and they are being integrated into cryptographic research, blockchain identity, and smart‑contract dispute mechanisms.
How are AI agents being used in regulated industries?
IBM and Microsoft’s Imagine Enterprise platform provides pre‑built agents and embedded engineering teams to help finance, insurance and retail firms meet compliance and governance requirements.
What security risks do AI agents pose for crypto projects?
Recent breaches, such as the AI‑agent attack on GitBook, show that malicious agents can exfiltrate API tokens and other secrets, prompting projects to tighten access controls and consider blockchain‑based identity verification.