Skip to main content

Search: result

OKHK 👀
OKHK 👀
个人数字泔水\(⁠◔⁠‿⁠◔⁠)
✨ Thinking...

  2. 张晋涛👀TIL
    tg 新特性： Bot-to-Bot Communication On Telegram, bots generally cannot see messages from other bots. However, in specific contexts, Bot-to-Bot communication is allowed – unlocking complex agentic flows and AI-powered use cases https://core.telegram.org/bots/features#bot…
    https://t.me/BotNews/117
    还是需要开发者自行处理多 Bot 间的消息循环 ♻
    #Telegram
    https://core.telegram.org/bots/features
    Loop Prevention Requirements
    Bot-to-bot communication can easily result in infinite interaction loops. When enabling this feature, you must implement safeguards to ensure interactions terminate.
    Recommended Safeguards
    Deduplicate repeated messages.
    Apply rate limits (e.g., no more than one reply every few seconds per bot).
    Enforce maximum interaction depth or timeouts, both globally and per sender/receiver.
    Important
    Your bot must remain stable even if another bot intentionally responds instantly and continuously.
    Failure to handle loops properly may lead to degraded performance or platform restrictions.
    Telegram
    BotNews
    🦀 AI Agents and Bot-to-Bot Communication

    Historically, bot-to-bot interaction was restricted on Telegram to prevent infinite message loops.

    Starting today, in specific contexts, Bot-to-Bot communication is allowed – unlocking complex agentic flows and AI…

  3. Claude Code Channels 远程控制会话
    https://code.claude.com/docs/en/channels
    官方支持的 Telegram / Discord 远程会话 #MCP 插件
    https://github.com/anthropics/claude-plugins-official/blob/main/external_plugins/telegram/README.md

    https://github.com/anthropics/claude-plugins-official/blob/main/external_plugins/discord/README.md
    Claude Code Docs
    Push events into a running session with channels - Claude Code Docs
    Use channels to push messages, alerts, and webhooks into your Claude Code session from an MCP server. Forward CI results, chat messages, and monitoring events so Claude can react while you're away.

  4. Lark（飞书海外版） “优化”（删号）
    去看了下个人版账户还在，团队的给我干掉了

    导出数据： https://www.larksuite.com/trust/privacy-center/support
    link
    LINUX DO
    一觉醒来 Lark 账号被关了
    早上收到了 Lark 官方邮件，标题为“Announcement of scheduled deletion of accounts”，内容如下： Dear customer, As part of our ongoing efforts to optimize our platform infrastructure and ensure service efficiency, we are currently consolidating our system resources. As a result…

  7. SearXNG - 免费搜索引擎

    https://github.com/searxng/searxng

    SearXNG 是一个免费的互联网元搜索引擎，它通过整合不同的搜索服务和数据库来提供搜索结果，同时保障用户隐私不被追踪和分析。

    ● 聚合来自各种搜索服务和数据库的结果
    ● 保护用户隐私, 不跟踪和分析个人资料
    ● 提供可定制和可扩展的搜索引擎

    #GitHub #Tool
    GitHub
    GitHub - searxng/searxng: SearXNG is a free internet metasearch engine which aggregates results from various search services and…
    GitHub - searxng/searxng: SearXNG is a free internet metasearch engine which aggregates results from various search services and…
    SearXNG is a free internet metasearch engine which aggregates results from various search services and databases. Users are neither tracked nor profiled. - searxng/searxng

  8. Windows Subsystem for Linux （WSL）Windows Linux 子系统正式开源

    https://blogs.windows.com/windowsdeveloper/2025/05/19/the-windows-subsystem-for-linux-is-now-open-source

    https://github.com/microsoft/WSL

    #Microsoft #GitHub
    Windows Developer Blog
    The Windows Subsystem for Linux is now open source
    The Windows Subsystem for Linux is now open source
    Today we’re very excited to announce the open-source release of the Windows Subsystem for Linux. This is the result of a multiyear effort to prepare for this, and a great closure to the first ever issue raised on the Microsoft/WSL repo:

  9. Qwen3 发布了，简单看了一下创新的地方：

    训练数据：这次使用了 36T 的预训练数据，可以说相当大了， DeepSeek V3 是 14.8T。这些数据覆盖了 100 多种语言，还包含了上一代模型生成的数学和代码内容。另一个点是还包含了从图片识别出来的文本。

    预训练：这块目前没看出太多特别的，采用的是 128 选 8 的 MoE 并没有使用 shared expert。

    后训练：这里是模型产生混合思考能力的关键步骤，在经过 CoT 的 RL 后又加入了一次 SFT 让模型在通用问题上直接出结果。粗略可以理解为有了个类似 DeepSeek R1 那样啥都要思考的模型后又加了些指导，让模型不要什么问题都思考。

    整体看下来是个在数据上下了大功夫，局部有创新的模型。但是从架构上看也很难有超预期的表现，还是要看实际使用的体验了。

    https://qwenlm.github.io/blog/qwen3/
    Qwen
    Qwen3: Think Deeper, Act Faster
    QWEN CHAT GitHub Hugging Face ModelScope Kaggle DEMO DISCORD
    Introduction Today, we are excited to announce the release of Qwen3, the latest addition to the Qwen family of large language models. Our flagship model, Qwen3-235B-A22B, achieves competitive results…

  10. Bybit遭黑客攻击：主钱包15亿美元ETH被盗，CEO承诺平台有能力1:1赔付用户

    知名加密货币交易所Bybit近日遭遇黑客攻击，其以太坊（ETH）主钱包被盗取约14亿美元的资产。据Bybit联合创始人兼CEO Ben Zhou在X平台（原推特）上表示，此次攻击是通过伪装交易界面实施的，所有签名者看到的都是合法的地址和URL，但实际上签署的信息更改了ETH冷钱包的智能合约逻辑，导致黑客控制了特定的ETH冷钱包，并将其中的ETH转移至未确认的地址。

    Ben Zhou强调，其他冷钱包和热钱包均未受到影响，平台的提现功能正常运行。他呼吁用户放心，Bybit有能力确保所有用户资产的安全


    https://fxtwitter.com/benbybit/status/1892963530422505586
    🧵 Thread • FxTwitter / FixupX
    Ben Zhou (@benbybit)
    Bybit ETH multisig cold wallet just made a transfer to our warm wallet about 1 hr ago. It appears that this specific transaction was musked, all the signers saw the musked UI which showed the correct address and the URL was from @safe . However the signing…

  14. https://ig.ft.com/generative-ai

    以滚动交互的形式，十分生动地解释了大模型推理过程中几个重要的概念，认真看完以后，你就能够理解什么是Token，什么是词群，如何通过概率推理出下一个 Token；什么是自注意力，如何通过自注意力理解上下文了；什么是词搜索和束搜索，如何生成高质量的文本等等。

    https://m.okjike.com/originalPosts/65541cb184bcbc4a4679c60d

    #AI #RePost
    Ft
    Generative AI exists because of the transformer
    Generative AI exists because of the transformer
    The technology has resulted in a host of cutting-edge AI applications — but its real power lies beyond text generation
OKHK