FindtheDifference (@wangh920) profile picture

Plant Population Geneticist, Bioinformatist, Scientific Idea Explorer, Soccer and Basketball Lover, and Nature Lover.

New Zealand
Joined May 2011
Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)Photo by FindtheDifference (@wangh920)
最新消息:印度正在准备发射一颗新的导航卫星,以支持本国的区域卫星导航系统 NavIC。 这颗名为 NVS-03 的卫星旨在支持印度及周边地区的导航服务。 10 月 26 日的发射信息显示,目标发射日期为 10 月 30 日,但发射时间仍可能调整。 NavIC 是印度自主研发的卫星导航系统,设计覆盖印度及其周边约 1,500 公里的区域。 #印度 #ISRO #太空 #NavIC
1
23
JUSTIN: India is preparing to launch a new satellite for its own regional navigation system, NavIC. The NVS-03 mission is intended to support navigation services over India and surrounding areas. An October 26 launch listing targeted October 30, but launch schedules can change. NavIC is India’s own satellite navigation system, designed to cover the country and an area extending roughly 1,500 km beyond it. #India #ISRO #Space #NavIC
14
最新消息:NASA 的 Swift 太空观测卫星在太空工作近 22 年后,完成了最后的观测。 Swift 于 2004 年发射,原计划工作两年,但最终观测了 2,000 多次伽马射线暴和 1,400 多颗超新星。 随着大气阻力将其不断拉向地球,卫星高度正在下降。10 月 22 日的一项预测预计它将在 11 月 2 日左右重返大气层,但具体时间仍可能变化。 #NASA #太空 #天文学
6
JUSTIN: NASA’s Swift space observatory has completed its final observations after nearly 22 years in space. Launched in 2004 for a planned two-year mission, Swift went on to observe more than 2,000 gamma-ray bursts and over 1,400 supernovae. The spacecraft has been losing altitude as atmospheric drag pulls it closer to Earth. A forecast issued on October 22 estimated re-entry around November 2, although the timing can change. #NASA #Space #Astronomy
5
最新消息:一篇新研究指出,软件开发正在面临一个问题:AI 写代码越来越快,但代码仍然需要审查。 研究人员分析了 50 份资料,了解 AI 编程代理如何改变开发者提交和审查代码的方式。 被提及最多的问题是:面对不断增加的代码修改,审查人员能否跟得上? 一些项目正在尝试新工具或限制外部贡献,但大多数方案仍保留某种形式的传统代码审查流程。 这篇预印本论文于 10 月 26 日发布。 #人工智能 #编程 #软件开发
1
16
JUSTIN: A new research paper highlights a growing challenge in software development: AI can produce code faster, but humans still need to review it. Researchers examined 50 documents discussing how coding agents are changing the way developers propose and review code changes. The most frequently mentioned problem was review capacity: can reviewers keep up with the growing volume of changes? Some projects are experimenting with new tools or restricting contributions, but most approaches still retain the traditional pull request in some form. The paper was released as a preprint on October 26. #AI #Coding #SoftwareDevelopment
9
最新消息:Google Cloud 与 NASA 正在测试机器人助手,未来可能用于月球和火星任务。 这些概念验证机器人旨在帮助宇航员处理日常工作和紧急情况,例如取来并准备医疗工具。 机器人使用 Google DeepMind 的物理模拟系统进行训练,AI 模型则借助 Google Cloud 技术进行微调。 这项测试的目标是探索:机器人如何在远离地球的任务中协助宇航员。 #Google #NASA #人工智能 #机器人 #太空
8
JUSTIN: Google Cloud and NASA are testing robotic assistants for future missions to the Moon and Mars. The proof-of-concept robots are designed to help astronauts with everyday tasks and emergencies, including retrieving and preparing medical tools. The robots are trained using a physics simulation engine developed by Google DeepMind and use AI models fine-tuned with Google Cloud technology. The goal is to explore how robotic assistants could support crews far from Earth. #Google #NASA #AI #Robotics #Space
7
最新消息:一艘欧洲探测器正在接近一颗曾被人类故意撞击的小行星。 欧洲航天局的 Hera 探测器计划于 11 月 4 日抵达双小行星系统:Didymos 和 Dimorphos。 2022 年,NASA 的 DART 探测器撞向 Dimorphos,测试人类能否改变小行星的运行轨道。 Hera 将近距离研究撞击后的情况,帮助科学家了解类似技术未来如何用于保护地球。 #欧洲航天局 #NASA #小行星 #太空
16
JUSTIN: A European spacecraft is approaching an asteroid that humanity deliberately hit. ESA’s Hera mission is scheduled to arrive at the Didymos and Dimorphos asteroid system on November 4. In 2022, NASA’s DART spacecraft crashed into Dimorphos to test whether an asteroid’s path could be changed. Hera will examine the aftermath up close—helping scientists understand how a similar technique might one day protect Earth. #ESA #NASA #Asteroid #Space
15
最新消息:日本发射了一艘探测器,前往研究火星最神秘的两颗卫星。 日本 MMX 探测器于 10 月 20 日升空,目前正在前往火星的途中。 它的任务是研究火卫一和火卫二,并尝试从火卫一采集样本,最终带回地球。 科学家希望借此回答一个长期存在的问题:火星的卫星究竟是被捕获的小行星,还是一次巨大撞击后形成的碎片? #日本 #火星 #太空
11
JUSTIN: Japan has launched a spacecraft to investigate one of Mars’ most mysterious moons. The MMX mission lifted off on October 20 and is now on its way toward Mars. Its goal: study the Martian moons Phobos and Deimos, collect material from Phobos, and eventually bring a sample back to Earth. Scientists hope the mission will help answer a long-standing question: Are these moons captured asteroids, or were they formed from debris after a giant impact? #Japan #Mars #Space
8
最新消息:OpenAI 发布了数百篇由 AI 生成的数学研究稿件,但接下来还有关键的一步。 10 月 6 日,OpenAI 公布了 722 篇数学稿件,涉及 372 组相关研究结果。 该公司还公开了部分可由计算机检查的证明,并支持数学家审查和进一步研究这些成果。 接下来的关键问题是:哪些结果能够经得起独立验证? #OpenAI #人工智能 #数学
10
JUSTIN: OpenAI has released hundreds of AI-generated mathematics manuscripts—but there’s an important next step. On October 6, the company published 722 manuscripts covering 372 groups of related results. OpenAI also released computer-checkable versions of some proofs and is supporting mathematicians who will examine and build on the work. The key question now: which results will stand up to independent scrutiny? #OpenAI #AI #Mathematics
5
AI 到底是不是一个烧钱的买卖? 从科技巨头最新财报来看,AI 竞赛的投入规模已经非常惊人。 #谷歌母公司Alphabet $GOOGL 2026 年第三季度,Alphabet 用于房产和设备的支出超过 510 亿美元。尽管公司当季盈利,自由现金流仍约为负 91 亿美元。 #微软 $MSFT 微软最近一个季度对计算基础设施的投资达到 452 亿美元,包括融资租赁。但在现有业务和云业务增长的支持下,微软仍产生了 186 亿美元的正自由现金流。 #Meta $META Meta 第三季度的资本支出,包括融资租赁,达到 349 亿美元。当季自由现金流约为负 21 亿美元。 #亚马逊 $AMZN 亚马逊预计,2026 年资本支出约为 2,200 亿美元。与此同时,其云业务 AWS 第三季度营收达到 460 亿美元,同比增长 39%。 那么,AI 到底是不是一个烧钱的买卖? 目前来看,建设 AI 基础设施确实需要巨额前期投入。 公司需要购买芯片、建设数据中心、支付电力成本,而这些投资最终能带来多少回报,还需要时间验证。 但 AI 也正在通过云服务和其他产品带来收入。关键问题是:这些收入和最终产生的利润,能否支撑如此庞大的投入。 需要注意的是,上述资本支出也包括云计算和其他基础设施,并非全部用于 AI。自由现金流为负,也不一定意味着公司亏损。 #Google #Microsoft #Amazon #AI
2
2
49
IS AI A MONEY-BURNING BUSINESS? The latest earnings from Big Tech show just how expensive the AI race has become. #Alphabet / #Google $GOOGL In the third quarter of 2026, Alphabet spent more than $51 billion on property and equipment. Its free cash flow was approximately negative $9.1 billion, even though the company reported a profit. #Microsoft $MSFT Microsoft invested $45.2 billion in computing infrastructure during its latest quarter, including finance leases. But it still generated $18.6 billion in positive free cash flow, supported by its existing businesses and cloud growth. #Meta $META Meta’s third-quarter capital expenditure, including finance leases, reached $34.9 billion. Its free cash flow was approximately negative $2.1 billion. #Amazon $AMZN Amazon expects to spend approximately $220 billion on capital expenditure in 2026. Its cloud business, AWS, reported third-quarter revenue of $46 billion, up 39% year over year. So, is AI a money-burning business? Right now, building AI infrastructure requires enormous upfront spending. Companies are buying chips, building data centers and paying for power before they know exactly how much return those investments will produce. But AI is also generating revenue through cloud services and other products. The key question is whether that revenue—and the profit it produces—can eventually justify the spending. One important distinction: these capital spending figures include cloud and other infrastructure, not just AI. And negative free cash flow does not necessarily mean a company is losing money.
1
50
AI是个烧钱的买卖
1
15
@grok, AI 对各大公司而言,是不是很烧钱?
1
12
最新消息:#福特 第三季度核心利润达到 27 亿美元,并将全年调整后利润预测的下限提高至 105 亿美元。 福特还将全年大宗商品成本预估提高至近 25 亿美元,涉及铜、钢铁、燃料和芯片等成本上涨。$F #Ford
1
16
最新消息:#沃尔玛 将在未来数年投入约 60 亿美元,扩大与仓储自动化公司 #Symbotic 的合作。 这笔投资将用于在沃尔玛配送网络中部署更多供应链自动化技术。$WMT $SYM #Walmart
31