article cover

免费的中国AI模型在黑客攻击方面正变得惊人地强大 一手编译

An experiment in which AI models were asked to break into simulated networks showed the hacking skills of DeepSeek and GLM to be not far behind OpenAI’s most advanced cyber model, ...
连线杂志AI最前沿
qimuai 发布于 
13 0
article cover

聊天机器人广告已经到来。以下是行之有效的做法——以及行不通的做法 一手编译

As Anthropic maintains its vow to keep ads out of Claude, the rest of the AI chatbot industry is racing to win over advertisers without antagonizing users.
连线杂志AI最前沿
qimuai 发布于 
16 0
article cover

蓝色浪潮迫使科技人士争相在太空建设数据中心 一手编译

After Democrats won big in Texas and other AI infrastructure hot spots on Tuesday, data center enthusiasts looked for comfort in low Earth orbit.
连线杂志AI最前沿
qimuai 发布于 
30 0
article cover

何时该用AI搜索网页——何时又不该用 一手编译

Chatbots are often quite good at retrieving the information you want from the web. But sometimes you’ll get better results by typing your query into a good old search engine.
连线杂志AI最前沿
qimuai 发布于 
26 0
article cover

聊天机器人正将徒步者引入歧途，后果十分危险 一手编译

Outsourcing planning to AI can lead people to surrender their own critical thinking, which can have potentially deadly results in places like the backcountry.
连线杂志AI最前沿
qimuai 发布于 
31 0
article cover

恐怖新型信息窃取恶意软件利用AI“自动入侵”受害者 一手编译

Named “Volta-Autodoxer,” the recently uncovered malware uses AI tools to collect troves of data about victims, from login credentials and credit card numbers to Discord sessions an...
连线杂志AI最前沿
qimuai 发布于 
42 0
article cover

一个承包商的Claude智能体是如何打入五角大楼的 一手编译

This week on Uncanny Valley, we also discuss Anthropic’s IPO filing and the volunteers gearing up to counter ICE at the midterms.
连线杂志AI最前沿
qimuai 发布于 
33 0
article cover

一家国防承包商让其个人AI代理接入了五角大楼的安全系统 一手编译

Sources tell WIRED that an Accenture employee’s personal Claude agent ran on a DOD network for two months before it was identified and the vulnerability could be patched.
连线杂志AI最前沿
qimuai 发布于 
36 0
article cover

我让一堆AI智能体管理一支梦幻足球球队。它们还不如人类——目前如此。 一手编译

To experiment with AI’s predictive skills, I let Codex loose on a fantasy team. It tried to bench Erling Haaland at first, but it soon learned the error of its ways.
连线杂志AI最前沿
qimuai 发布于 
38 0
article cover

Anthropic称终将颠覆生物学——但还需时日 一手编译

In breathless press releases and posts, the company has touted its models’ ability to deliver scientific breakthroughs. But its IPO filing taps the brakes on just how fast its AI w...
连线杂志AI最前沿
qimuai 发布于 
39 0

搜索

日历

标签

内容导航 AI新闻 启幕AI 量子安全 科学研究 基础物理 量子密码学 量子计算 强化学习 机器人学习 机器人 高管变动 OpenAI for Science Prism Kevin Weil 研究团队 AI视频 高管离职 Sora AI安全 Tinder 身份验证 Sam Altman World Google 暑期旅行 出行工具 Google Maps AI旅行 谷歌 IO Interactive 主题曲 电子游戏 007 拉娜德雷 数码优惠 办公电脑 EliteDesk 800 G6 迷你主机 惠普 并购 亚马逊 AI创业 初创企业 西雅图创投 WIRED 网络攻击 隐私保护 年龄验证 欧盟

分类

倒计时

最新文章

热门文章

链接

微语

  • twitter-author qimuai 1 年前

    We’re opening an office in Seattle

    Come join us!

    https://t.co/E9BY2xFU2j

  • twitter-author qimuai 1 年前

    Get through PDFs effortlessly with Grok.

    Highlight any section and click "Explain" to have Grok help you understand it.

    Or click "Quote" to ask a specific question. https://t.co/epQt4Nypfw

  • twitter-author qimuai 1 年前

    i have had the strangest experience reading this: i assume its all fake/bots, even though in this case i know codex growth is really strong and the trend here is real.

    i think there are a bunch of things going on: real people have picked up quirks of LLM-speak, the Extremely Online crowd drifts together in very correlated ways, the hype cycle has a very "it's so over/we're so back" extremism, optimization pressure from social platforms on juicing engagement and the related way that creator monetization works, other companies have astroturfed us so i'm extra sensitive to it, and a bunch more (including probably some bots).

    but the net effect is somehow AI twitter/AI reddit feels very fake in a way it really didnt a year or two ago.

  • twitter-author qimuai 1 年前

    Jakub and Szymon:

    https://t.co/6qrVNp5Q3u

  • twitter-author qimuai 1 年前

    有时候会觉得，写的这些东西都太浅显了，发到朋友圈会不会被加的大V们嘲笑？如果有这个想法，肯定是做不好自媒体的。

    除非大V们是你的目标用户，否则永远不要让其他无关的人左右你的选题。

    你应该只关注，你的目标用户需要什么。

  • twitter-author qimuai 1 年前

    人类在学习新知识或解决新问题时，往往不是直接“改写大脑参数”，而是通过“记笔记”或“自我提醒”的方式，将经验和策略以显式的形式保留下来。

    例如，当你遇到某类问题时，会总结出“下次遇到类似情况可以这样做”的经验。这种方式更像是在不断编辑自己的“系统提示”，而不是每次都重新训练大脑。

  • twitter-author qimuai 1 年前

    此时常有人问该选哪个模型。我的答案永远是：工作环境允许的最强模型。我们不是大规模使用AI，所以即使工程角度认为"贵"或"慢"，也要选最好的大脑。（唯一限制应是公司对敏感信息上传的规定。）

  • twitter-author qimuai 1 年前

    成功与思考伙伴（人或机器）协作的关键在于取其精华去其糟粕

  • twitter-author qimuai 1 年前

    引用David Lieb（YC合伙人、Google Photos创始人）的话："你的直觉是世界上最复杂的机器学习模型。"使用生成式AI时，用你的专业品味识别有价值的内容。

  • twitter-author qimuai 1 年前