东方财富东方财富华尔街见闻华尔街见闻同花顺同花顺微信读书微信读书LichessLichessV2EXV2EX微博微博今日头条今日头条百度百度快手快手哔哩哔哩哔哩哔哩知乎知乎腾讯新闻腾讯新闻网易新闻网易新闻澎湃新闻澎湃新闻新浪新闻新浪新闻新浪网新浪网豆瓣电影豆瓣电影GitHubGitHubCSDNCSDNIT之家IT之家36氪36氪宝庆银楼宝庆银楼中国黄金中国黄金周生生周生生英皇珠宝英皇珠宝老凤祥（广东）老凤祥（广东）六福珠宝六福珠宝周大福周大福周六福周六福Adafruit BlogAdafruit BlogAIbaseAIbaseAnt Bailing Developer BlogAnt Bailing Developer BlogAnthropicAnthropic小众软件小众软件AppleAppleApple PodcastsApple PodcastsAppleInsiderAppleInsiderArs TechnicaArs TechnicaarXivarXivAxiosAxiosBerkeley AI ResearchBerkeley AI ResearchBarchartBarchartBBC NewsBBC NewsBD Tech TalksBD Tech TalksBerkeley RDIBerkeley RDIBig ThinkBig Think哔哩哔哩热搜哔哩哔哩热搜哔哩哔哩热门视频哔哩哔哩热门视频BinanceBinanceBloombergBloombergBlueskyBlueskyBoston Dynamics BlogBoston Dynamics BlogBusiness InsiderBusiness InsiderByteDance Seed ResearchByteDance Seed Research参考消息参考消息虫部落虫部落Claude BlogClaude BlogClaude Code ReleasesClaude Code ReleasesCloudflare BlogCloudflare BlogCMU Machine Learning BlogCMU Machine Learning BlogCoinGeckoCoinGeckoCointelegraphCointelegraphcrates.iocrates.ioCrowd SupplyCrowd SupplyCSS-TricksCSS-Tricks51CTO51CTOCult of MacCult of MacCursor BlogCursor Blogdaily.dev · Populardaily.dev · PopularDaring FireballDaring FireballDario AmodeiDario AmodeiDeepLearning.AI · The BatchDeepLearning.AI · The BatchGoogle DeepMind BlogGoogle DeepMind BlogDeepSeek GitHubDeepSeek GitHubDefiLlamaDefiLlamaDescript BlogDescript BlogDEV.toDEV.to数字尾巴数字尾巴Docker HubDocker Hub豆瓣读书豆瓣读书豆瓣讨论豆瓣讨论Dwarkesh PatelDwarkesh PatelThe EconomistThe EconomistEleutherAI BlogEleutherAI BlogEngadgetEngadgetFinancial TimesFinancial TimesFlathubFlathubfreeCodeCampfreeCodeCampFrontiersFrontiersGary MarcusGary Marcus极客公园极客公园原神原神果核剥壳果核剥壳GoogleGoogleGoogle TrendsGoogle Trends果壳果壳HackadayHackadayHacker NewsHacker NewsHelloGitHubHelloGitHub历史上的今天历史上的今天HomebrewHomebrew崩坏3崩坏3Hugging FaceHugging Face活动行活动行虎扑虎扑虎嗅虎嗅IEEE SpectrumIEEE Spectrum爱范儿爱范儿凤凰网 · 热点资讯凤凰网 · 热点资讯inclusionAIinclusionAIIndie HackersIndie HackersInfoQInfoQInterconnectsInterconnects爱奇艺热播榜爱奇艺热播榜IT之家「喜加一」IT之家「喜加一」简书简书稀土掘金稀土掘金Andrej KarpathyAndrej KarpathyMoonshot AI KimiMoonshot AI KimiLatent SpaceLatent SpaceLessWrongLessWrongLil'Log (Lilian Weng)Lil'Log (Lilian Weng)Linux.doLinux.doLMSYS BlogLMSYS BlogLobstersLobsters英雄联盟英雄联盟LWN.netLWN.netMacRumorsMacRumorsMake: MagazineMake: MagazineMaven CentralMaven CentralMediumMediumMeituan LongCatMeituan LongCatMeta AI BlogMeta AI BlogMeta EngineeringMeta EngineeringMidjourney UpdatesMidjourney UpdatesMiniMaxMiniMaxMIT News · RoboticsMIT News · RoboticsMIT Technology ReviewMIT Technology Review米游社米游社Mozilla.ai BlogMozilla.ai BlogNatureNatureNeuroscience NewsNeuroscience NewsNew Atlas · RoboticsNew Atlas · RoboticsThe New YorkerThe New Yorker水木社区水木社区NGANGA9to5Mac9to5MacNodeSeekNodeSeek牛客牛客NuGetNuGetNVIDIANVIDIA纽约时报纽约时报OEISOEISOne Useful ThingOne Useful ThingOpen Robotics BlogOpen Robotics BlogOpenAIOpenAIOpenAlexOpenAlexOpenReviewOpenReviewOpenRouterOpenRouterPackagistPackagist远景论坛远景论坛PhoronixPhoronixPhys.orgPhys.orgPixivPixivPlanet ROSPlanet ROS吾爱破解吾爱破解PoliticoPolitico中国电建阳光采购网中国电建阳光采购网Product HuntProduct HuntPubMedPubMedPyPIPyPIQoderQoder腾讯视频热搜榜腾讯视频热搜榜Quanta MagazineQuanta MagazineQwenQwenReutersReutersRFC EditorRFC EditorRobohubRobohubRobotics & Automation NewsRobotics & Automation NewsRobotics TomorrowRobotics TomorrowROS DiscourseROS Discourse阮一峰的网络日志阮一峰的网络日志RubyGemsRubyGemsRunwayRunwaySam AltmanSam AltmanScienceAlertScienceAlertAhead of AIAhead of AISecurityOnlineSecurityOnlineServeTheHomeServeTheHomeServiceNow AIServiceNow AISimon WillisonSimon WillisonSingularity HubSingularity HubSky NewsSky NewsSlashdotSlashdotSmashing MagazineSmashing Magazine什么值得买什么值得买SolidotSolidotSpotifySpotify俄罗斯卫星通讯社俄罗斯卫星通讯社少数派少数派Stack OverflowStack OverflowStack Overflow BlogStack Overflow Blog崩坏：星穹铁道崩坏：星穹铁道SteamSteamSubstackSubstackSunoSunoSuno BlogSuno BlogSynced ReviewSynced Review淘宝逛一逛淘宝逛一逛TechCrunchTechCrunchTech Xplore · RoboticsTech Xplore · Robotics腾讯热点腾讯热点The AtlanticThe AtlanticThe DecoderThe DecoderThe GradientThe GradientThe GuardianThe GuardianThe RegisterThe RegisterThe VergeThe Verge百度贴吧百度贴吧TikTokTikTokTindie BlogTindie BlogTomer TunguzTomer TunguzTom's HardwareTom's HardwareTransformer CircuitsTransformer CircuitsTVmazeTVmaze优设网优设网UiverseUiverseVentureBeat · AIVentureBeat · AI万方数据万方数据中央气象台中央气象台WikidataWikidataWikipediaWikipediaWiredWiredThe Wall Street JournalThe Wall Street JournalxAI NewsxAI News小鹅通小鹅通YahooYahooYahoo FinanceYahoo FinanceYollomiYollomi有道精品课有道精品课YouTubeYouTube游研社游研社知乎日报知乎日报Zhipu AI ResearchZhipu AI Research财新数据通财新数据通央视新闻联播央视新闻联播CelesTrakCelesTrakCISA KEVCISA KEV财联社财联社CoinDeskCoinDesk法布财经法布财经富途牛牛富途牛牛格隆汇格隆汇HDXHDX和讯网和讯网Investing.comInvesting.com界面新闻界面新闻金十数据金十数据21财经21财经金融界金融界Launch Library 2Launch Library 2MKTNews · 快讯MKTNews · 快讯NASA EONETNASA EONETNOAA/NWSNOAA/NWSNVDNVDopenFDAopenFDAOSV.devOSV.dev量子位 · 具身智能量子位 · 具身智能新浪财经新浪财经Spaceflight NewsSpaceflight NewsTelegram OSINTTelegram OSINT腾讯混元研究腾讯混元研究USAspendingUSAspendingUSGS EarthquakesUSGS EarthquakesWHO 疫情通报WHO 疫情通报选股宝选股宝第一财经第一财经

实时热搜

731 个榜单
共 100 条热搜查看完整榜单
  • 01
    Sunday Reboot: Waiting a decade for Apple Pay and Apple Music
    In this week's Sunday Reboot, the expansion of Apple Pay and Apple Music to some countries has taken a decade longer than they really should. They also demonstrate how delays can happen for many different reasons. Apple Pay in action. Sunday Reboot is a weekly column covering some of the lighter stories within the Apple reality distortion field from the past seven days. All to get next week underway with a good first step. Continue Reading on AppleInsider | Discuss on our ForumsMalcolm Owen <malcolm@appleinsider.com>
  • 02
    Design is a Jobsian-level focus for John Ternus
    John Ternus is putting his stamp on Apple by focusing on the company's design work, with the CEO taking an almost Jobsian approach to make design great again. Apple CEO John Ternus - Image Credit: Apple The era of Tim Cook as Apple CEO was one of growth, with sales and manufacturing the core of his tenure at the top. Under the rule of Ternus, design is becoming a top priority once more for the company, as Ternus aims to add more to the company by strengthening its weaker areas. Writing in SundayMalcolm Owen <malcolm@appleinsider.com>
  • 03
    Touch-screen MacBook Pro will be significantly lighter
    The forthcoming new MacBook Pro is expected to bring an OLED touch-screen, but it's also now predicted to be at least a little closer to the MacBook Air's lower weight. Apple's MacBook Pro. Every detail of the expected MacBook Pro refresh has been rumored, sometimes for years, as Apple hasn't made even a minor design change since 2022. That includes the expectation that the new MacBook Pro will be lighter, which was first reported back in 2024 . However, Bloomberg now claims that the weight diffWilliam Gallagher <wg@williamgallagher.com>
  • 04
    Initial basic M4 support coming to Linux 7.4
    While it won't be practically usable initially, Asahi Linux developer Sven Peter sent out the pull requests for very basic support for the M4 "device tree," and for the MacBook Neo. Upon release, the support will be good enough to install the kernel on those devices, and that's about it. Further development, particularly for GPU support, will be needed to make the OS usable on the M4 family of hardware. Phoronix.com Continue Reading on AppleInsider | Discuss on our ForumsMike Wuerthele <mike@appleinsider.com>
  • 05
    Crime blotter: Two arrested for stealing from Apple Store, yoga store
    A cab driver is arrested for his part in a Miami Beach iPhone theft ring, one of the first iPhone 18 Pro thefts, and an AirTag recovers a stolen food trailer, all in this week's Apple Crime Blotter. The Apple Store in Garden State Plaza in New Jersey. The latest in an occasional AppleInsider series, looking at the world of Apple-related crime. Continue Reading on AppleInsider | Discuss on our ForumsStephen Silver <stephen@appleinsider.com>
  • 06
    AirPods Pro plunge to as low as $89.99 for fall Prime Day
    Amazon's Prime Big Deal Days event starts October 6, but AirPods Pro 3 are already $70 off at $179 and the closeout first-gen model is discounted to $89.99. Pick up AirPods Pro for as low as $89.99 during Prime Big Deal Days Amazon's $70 discount on AirPods Pro 3 is live now as Prime Big Deal Days offers heat up. This is within $10 of the lowest price on record at Amazon for Apple's high-end earbuds and the steepest discount currently available across the in-ear AirPods lineup. Buy AirPods Pro 3Christine McKee <christine@appleinsider.com>
  • 07
    Amazon's $199 Apple Watch deal is back for Prime Big Deal Days
    Amazon's October Prime Day Apple Watch deals deliver prices as low as $199, with even the new Apple Watch Series 12 and Ultra 4 marked down. Grab Apple Watch deals from $199 for Prime Big Deal Days - Image credit: Apple Amazon's $199 Apple Watch SE 3 deal has returned for Prime Big Deal Days. With savings of up to $200 off Apple Watch models, you can pick up Apple's wearable at a substantial discount, perfect for picking up an early holiday gift for yourself or others. Grab Apple Watch deals froChristine McKee <christine@appleinsider.com>
  • 08
    Amazon slashes $100 off iPad mini 7 ahead of Prime Big Deal Days
    Limited-time deals are in effect now on Apple's iPad mini 7, delivering a $100 discount across the line ahead of next week's Prime Big Deal Days sale event. Prime Day iPad mini deals knock $100 off - Image credit: Apple Pick up the $100 discount at Amazon on this weekend with the Prime markdown, despite Prime Big Deal Days officially starting on October 6. Buy iPad mini 7 for $499 Continue Reading on AppleInsider | Discuss on our ForumsChristine McKee <christine@appleinsider.com>
  • 09
    Early October Prime Day deals: Save up to $600 on Mac, Apple Watch, Jackery, Ugreen
    Early fall Prime Day deals are in effect now, with savings of up to $600 off current Apple hardware, plus up to 30% off Ugreen docks and up to $150 off Jackery power stations. Save up to $600 on Apple ahead of fall Prime Day - Image credit: Apple, Amazon Early Prime Big Deal Days savings are going on now, and we've compiled top picks delivering up to $600 off ahead of the October 6-7 sale. Shop early Prime deals Continue Reading on AppleInsider | Discuss on our ForumsChristine McKee <christine@appleinsider.com>
  • 10
    Amazon's $1,099 M5 MacBook Air deal offers $200 October Prime Day savings
    Numerous M5 MacBook Air models are $200 off as Amazon launches early Prime Big Deal Days markdowns on both 13-inch and 15-inch models. Save $200 across the M5 MacBook Air line - Image credit: Apple Amazon has slashed Apple's M5 MacBook Air by $200 , with a number of retail configurations eligible for the triple-digit markdown. Both 13-inch and 15-inch models are included, with prices starting at $1,099. Save $200 on M5 MacBook Air Continue Reading on AppleInsider | Discuss on our ForumsChristine McKee <christine@appleinsider.com>
  • 11
    Apple SouthGate reopening on October 23
    Apple SouthGate in Bath, England closed for renovations on July 4. Three months later, on the day the iPhone Duo is available, it will reopen to customers. Prior to its closure, it was a smaller store, with the updated wood design. The renovation is expected to bring a redesigned Genius Bar, and a dedicated pickup counter. Apple SouthGate website Continue Reading on AppleInsider | Discuss on our ForumsMike Wuerthele <mike@appleinsider.com>
  • 12
    If your iPhone 18 Pro Max always says SOS, you need to get it replaced
    There are now patches to apply if you don't have any problems making calls on an AT&T iPhone 18 Pro Max . If you do, you need to get your phone replaced. iPhone 18 Pro Max Since launch, iPhone 18 Pro Max models have been dropping off the AT&T network. In some cases, never to return. After an OS update earlier in the week, and a carrier settings update on Friday, the problem may no longer manifest for those who haven't seen it. But, that doesn't help if your iPhone has issues already. Continue ReMike Wuerthele <mike@appleinsider.com>
  • 13
    Apple moves to protect private Mac data from overreaching AI apps
    AI apps can reach far beyond the data Mac users intend to share, which has prompted Apple to take action and require clearer consent before granting access. Apple to tighten Mac Full Disk Access over AI privacy risks On Friday, Apple announced plans to tighten Full Disk Access controls on the Mac as AI agents gain access to sensitive personal information. The permission lets apps bypass privacy protections and could expose messages, email, files, and browsing history. Recent incidents involvingLou Hartle <news@appleinsider.com>
  • 14
    SwitchBot Lock Vision Pro review: lack of home key support will drive Apple users away
    Apple Home users demand accessories that enhance their ecosystem and SwitchBot appears opposed to doing so, even if products like Lock Vision Pro are more than capable. SwitchBot Lock Vision Pro review: bare minimum Apple Home integration When SwitchBot reached out to ask me to conduct a review of some of its products, I was excited. The company has been releasing increasingly interesting products in recent years, and I was eager to see how they'd integrate with my Apple Home. The problem is, beWesley Hilliard <wes@appleinsider.com>
  • 15
    There's a new way to run Android apps on Mac
    There's a lot of crap to wade through on the assorted Android app stores, but there are some gems too. The "Mac.apk" project installs Android apps into your Applications folder, like any other Mac app. It's early yet, and the developer says that a lot of stuff doesn't work properly, particularly games with anti-cheat. My quick trial with the software has been pretty smooth, though. We'll be keeping an eye on this one. https://github.com/kksimp/Mac.apk-Releases Continue Reading on AppleInsider |Mike Wuerthele <mike@appleinsider.com>
  • 16
    Why I can't wait to swap this iPhone 18 Pro Max for an iPhone Duo
    I've been using an iPhone 18 Pro Max for two weeks, and while it's been great, I'm very much looking forward to it making way for something new and exciting in the brand-new iPhone Duo . The era of the slab iPhone is coming to an end I reviewed the iPhone 18 Pro Max for AppleInsider , giving it a very respectable four stars out of five. At the time I wrote that review, and for the next three weeks as I write this, it's the very best iPhone money can buy. The iPhone 18 Pro Max has three great camOliver Haslam <oliver@monkeymanmedia.com>
  • 17
    Apple plans fix as Microsoft's Exchange shutdown disrupts Mac email
    Apple is working on a fix for Exchange access in its Mac apps as Microsoft retires the connection they rely on. Apple plans Mac Mail fix as Microsoft phases out EWS The disruption concerns work or school accounts hosted through Microsoft's Exchange Online service. This can disrupt Mail, Calendar, Contacts, Notes, and Reminders, while Apple prepares an update and administrators retain temporary options. A reader from Old Dominion University forwarded an IT notice reporting problems accessing univLou Hartle <news@appleinsider.com>
  • 18
    Judge gives Google legal permission to steal content for AI summaries
    A US judge has said he's sympathetic to publishers who are losing money because of Google's AI summaries stealing content, but he's not going to do anything to stop it. Google has had an antitrust case dismissed Google recently escaped having its search and advertising business split over antitrust accusations from the Department of Justice. But it has been facing multiple other antitrust cases and now there's one fewer. According to Reuters , Google's parent company Alphabet had convinced a USWilliam Gallagher <wg@williamgallagher.com>
  • 19
    M6 Mac mini review: Excellent hardware hampered by inadequate SSD storage
    The M6 Mac mini has a faster chip and is solid hardware, but that $899 hurts and there is just not enough storage at that entry price for Apple Intelligence and all of your apps. M6 Mac mini review I have loved the Mac mini since day one, more than 20 years ago. There has been one in my house consistently since the first G4 model, generally attached to a television, or used as a server, and so forth. Oddly, the 2026 M6 Mac mini is not the cheapest Mac overall. The MacBook Neo has something to saMike Wuerthele <mike@appleinsider.com>
  • 20
    Home Hub, iPhone 18 Pro cameras, and Meta VR on the AppleInsider Podcast
    Leaks point to a busy October for Apple with a new Apple TV, HomePod mini, and Home Hub launching in tandem with the iPhone Duo. Here are the latest rumors and more on the AppleInsider Podcast. You probably don't want Meta working your MacBook Pro, no matter how cute - image credit: Meta. The iPhone 18 Pro Max camera system is Apple's best yet, but the gains are small year-over-year. The more interesting comparison will be when your hosts can get their hands on iPhone Duo to see exactly how thatWesley Hilliard <wes@appleinsider.com>
共 50 条热搜查看完整榜单
共 20 条热搜查看完整榜单

arXiv · cs.AI

2天前更新
  • 01
    Heavy-Tailed Memory Traces in Long-Horizon Language Agents
    arXiv:2610.00010v1 Announce Type: new Abstract: Long-horizon language agents increasingly rely on external memory as a frozen world model, yet current memory systems are usually judged only by task success or token cost. We argue that the missing object is the shape of memory use: under finite context and repeated retrieval, agent memory can concentrate on a small core while leaving rare states in a long tail where prediction errors accumulate. We study this effect through a conservative tail auXinyuan Song, Zekun Cai
  • 02
    When Do Causal World Models Help Modular LLM Agents
    arXiv:2610.00012v1 Announce Type: new Abstract: LLM agents increasingly act through modular systems, such as order, payment, inventory, and shipment services, where actions in one module change which transitions are valid in another. Standard world models usually fit observational traces, but this is not the quantity needed for intervention-time planning: a trace may show that payment precedes shipment without identifying whether payment authorizes shipment, inventory mediates the effect, or a hXinyuan Song, Zekun Cai
  • 03
    From Proposal to Verified Effect: Praxa, an Evidence-Bound Harness for Governed AI Agent Execution
    arXiv:2610.00015v1 Announce Type: new Abstract: Large-language-model agents can propose and execute actions, but proposal, authority, dispatch, verified external effect, and serving promotion are different claims. We present Praxa, an agent harness that represents these states explicitly through deterministic admission, brokered execution, external read-back, reconciliation, and reviewed promotion. We report four evidence lanes. First, an author-run repository-local audit at a pinned revision paStefan G. Creadore
  • 04
    What Do Rationales Communicate? A Message-Intervention Study in Role-Specialized QA
    arXiv:2610.00018v1 Announce Type: new Abstract: Role-specialized QA pipelines increasingly pass rationales from a reasoner to a verifier, but it is unclear what this message actually buys: better answers, stronger support assessment, or a new failure surface. We introduce a message-intervention diagnostic that fixes the evidence and candidate answer while varying only the rationale passed across the reasoner-to-verifier boundary. On 400 MuSiQue, HotpotQA, and 2WikiMultiHopQA examples with DeepSeJiameng Zhang, Hongqiu Wu
  • 05
    Measuring the Microtask Eligibility Gap: When Is an Off-the-Shelf SLM Enough for an Agent Harness?
    arXiv:2610.00025v1 Announce Type: new Abstract: Agent harnesses increasingly want to run small language models (SLMs) on the microtasks around a frontier large language model (LLM) planner: auto-approving shell commands, writing memory, selecting tools, ranking past turns. We ask whether off-the-shelf SLMs meet practitioner-defined thresholds and, when they fail, why, and whether quantization changes the answer. We build a benchmark of 4 such microtasks with fixed prompts and automatic metrics,Jundong Hu, Shekar Ramachandran
  • 06
    Characterizing a Configuration Where Inference-Time PRM-Pruned Fragment Grafting Is Inert: Evidence from Three Reasoning LMs
    arXiv:2610.00047v1 Announce Type: new Abstract: Diversity collapse in parallel chain-of-thought has motivated inference-time interventions built on a natural design: when a process reward model (PRM) prunes a chain, its high-PRM prefix is extracted and grafted verbatim as an in-context demonstration into a still-decoding sibling. We isolate this mechanism, PRM-Pruned Fragment Grafting (PPFG), as the most cost-minimal operationalization of cross-trajectory step-level transfer, and test it at theKhawaja Murad ul Hassan, Mehran Ebrahimi
  • 07
    Gradient-Aligned Pair Selection for Personalized Preference Optimization
    arXiv:2610.00061v1 Announce Type: new Abstract: Personalizing large language models (LLMs) requires aligning generation behavior with user-specific preferences rather than aggregate quality. While Direct Preference Optimization (DPO) provides a stable framework for preference learning, its effectiveness in personalized settings critically depends on how preference pairs are selected. Existing approaches typically rely on heuristic criteria, such as likelihood-based extremes, which decouple optimRuoming Jin, Xinyu Li, Hao Zhou, Jianfeng Zhu, Ruixin Guo, Feodor Dragan, Lei Xu, Haixun Wang, Yang Zhou
  • 08
    K-Dense BYOK: An Open-Source AI Research Assistant That Runs Locally and Keeps a Hash-Chained Lab Notebook
    arXiv:2610.00074v1 Announce Type: new Abstract: K-Dense BYOK (bring your own keys) is a free, open-source AI research assistant for scientists in any field that runs on the researcher's own computer. The researcher supplies access to a model of their choice, hosted or running locally, and the application supplies everything else: a place for the work to run, a layer of scientific scaffolding, and a complete record. Each project is an ordinary folder, so the data, the code, the results, and the rAubrey M. Brueckner, Darshil Patel, Yuhuan He, Timothy Kassis
  • 09
    Scientific Agents: Evaluating Profession-Specific System Prompts on Scientific Tasks
    arXiv:2610.00084v1 Announce Type: new Abstract: Detailed profession-specific system prompts raise token use and estimated cost per response without a consistent accuracy gain. We evaluate Scientific Agents, an open-source corpus of 503 profession-specific AGENTS.md profiles, with Gemini 3.8 Flash via OpenRouter in the Pi agent harness. We compare matched profiles with four controls: a minimal baseline ("You are a helpful assistant"), the profile's opening role sentence, a generic scientific rigoTimothy Kassis
  • 10
    Comedic Fool's Gold: Reward Exploits and Countermeasures in Conversational Humor
    arXiv:2610.00197v1 Announce Type: new Abstract: We investigate automated rewards for training language models in conversational humor, focusing on reward exploits and countermeasures. Two approaches aim to capture understandable surprise and predicted audience amusement. Controlled tests show that an embedding-based surprise reward accepts word-shuffled replies as readily as witty ones. A fluency filter detects the shuffles, but the combined reward also rejects some witty replies and fails furthSam Larson
  • 11
    EviGraph: Proof-Carrying Selective Recommendation over Temporal Public-Service Knowledge Graphs
    arXiv:2610.00212v1 Announce Type: new Abstract: Public-service recommendations require evidence that matches the requested service, scope, and date. Yet treating every missing detail as decisive can withhold useful recommendations. We introduce EviGraph, which distinguishes critical decision requirements from information that can remain unresolved. A language agent links these requirements to evidence in a temporal knowledge graph, while a deterministic checker establishes whether a recommendatiYixi Zhou, Sikun Wang, Lei Fan, Fan Zhang
  • 12
    Build2SPARQL: A Large-Scale Text-to-SPARQL Benchmark Dataset for Building Knowledge Graph Querying
    arXiv:2610.00224v1 Announce Type: new Abstract: Building automation systems are increasingly represented as semantic knowledge graphs (KGs) using ontologies such as Brick and ASHRAE 223P, creating a machine-readable substrate for artificial-intelligence applications. One promising application is translating natural-language questions into SPARQL (text-to-SPARQL), which would let building operators query these graphs through language agents, but progress is limited by the scarcity of large naturaWooyoung Jung
  • 13
    Robust Is Salient: An Informed Adversary Moves the Optimal Signal onto the Salience Pole
    arXiv:2610.00233v1 Announce Type: new Abstract: When an informed adversary shares the audience of a constrained signalling channel, the signal that best protects the truth is the signal that best describes it. On 108 confirmatory items, the adversary-robust optimum aligns exactly with the salience pole from prior work. Across a 200,000-item pool, the two differ on only 2,748 items --- lying exactly where the prior salience-to-Bayes coordinate is undefined. Where defined, robustness is achieved bCris Huynh
  • 14
    Conflicting Supervision Moves Commitment, Not Capability: A 12.29{\sigma} arrangement effect that is exactly zero under a convention-agnostic score
    arXiv:2610.00234v1 Announce Type: new Abstract: "Train a model on the same problems written under two incompatible conventions, both correct, and ask what the ordering of that data writes into the parameters. The learning-rate schedule is not a background condition for that question. It is the averaging operator, and it decides the answer. We prove a bound in which the arrangement and the schedule enter the ordering effect as separate multiplied factors: the arrangement only as a block period, tWenhui Chen
  • 15
    Knowing When to Yield: Grounded Arbitration of User Corrections in Text-Based Embodied Agents
    arXiv:2610.00282v1 Announce Type: new Abstract: How should an embodied agent respond when a person's correction may be wrong? We formulate grounded correction arbitration as a choice among accepting, rejecting, inspecting the world, and asking the speaker. GAVA implements this interface with observation-bounded evidence, legal probes, and a one-step expected-loss rule. In text-only ALFWorld, 162 checkpoints produce 972 paired true and false interventions. Complete local inspections give GAVA andYezhou Cheng, Runjia Du, Zeming Liu, Hang Lyu, Zehua Yang, Bojun Lin
  • 16
    Rules to Tools: Executable Checks for LLM Agents in Scientific Computing
    arXiv:2610.00313v1 Announce Type: new Abstract: Scientific coding agents receive equations, boundary conditions, and output requirements in writing, then must assess the programs they revise. Rules to Tools (R2T) supplies prepared executable checks of public scientific requirements. Matched SciCode repair groups share written checks, starting programs, model, and budgets; the tool group receives a callable implementation. Across two task-ID cohorts, complete repair is 26/30 with text and 29/30 wJingjie Ning, Guojiang Zhao, Chen Xu, Shanshan Zhong, Xiaochuan Li, Ji Zeng, Guolin Ke
  • 17
    Predictive Credit: Measuring What Scientific Explanations Add to Experimental Forecasts
    arXiv:2610.00314v1 Announce Type: new Abstract: Research agents explain planned experiments. We measure predictive credit with paired forecasts sharing an intervention, forecaster, and outcome while varying description, matched explanation, and donor context. Five checks track commitment, delivery, predictive gain, alignment, and known-signal uptake. Across 336 prospective states in controlled learning, 12 Tox21 endpoints, and 24 OpenML tasks, v5's frozen credit decision was inconclusive. Tox21'Jingjie Ning, Xueqi Li, Yibo Kong, Dongting Li
  • 18
    ContractRL: Shielded Group-Relative Policy Optimization for Auditable Tool-Call Repair
    arXiv:2610.00328v1 Announce Type: new Abstract: Structured tool calls often fail after only a small number of fields violate a schema or an execution contract. Regenerating the complete object enlarges the action surface and makes repeated repair difficult to audit. We introduce ContractRL, a contract-constrained sequential repair protocol that models verifier-guided JSON repair as a bounded decision process. At each step the policy observes the candidate, typed verifier feedback, JSON Pointer,Miaobo Hu, Shuhao Hu, Xiaobo Guo, Xin Wang, Bokun Wang, Yina Sa, Daren Zha, Jun Xiao
  • 19
    Mathematical Transfer in LLMs Follows Reasoning Approach More Than Topic
    arXiv:2610.00331v1 Announce Type: new Abstract: When selecting mathematical training data for LLMs, a natural organizing principle is topic: probability examples for probability targets. An alternative is reasoning approach: worked solutions that share a solution method with the target, even when the mathematical domain differs. We ask which relation produces greater transfer after fine-tuning. We evaluate two counterbalanced $2\times2$ designs: probability and combinatorics crossed with invariaSajad Goudarzi, Samaneh Zamanifard, Seyed Amin Seyed Haeri, Moloud Nasiri, Hamed Rahimian
  • 20
    Fault-Tolerant Budget Conservation in Distributed Multi-Agent Delegation
    arXiv:2610.00349v1 Announce Type: new Abstract: Resource limits are becoming an authorization boundary for AI agents that delegate work across concurrent and failure-prone workers. Parent-child allocation constraints, affine objects, and distributed escrow do not by themselves prevent overspend when replies are lost, effects complete after timeout, messages repeat, branches partition, or DAG joins alias one lineage. We formalize fault-tolerant budget conservation for distributed multi-agent deleGenliang Zhu, Chu Wang
共 100 条热搜查看完整榜单

Axios · 实时热榜

43分钟前更新
  • 01
    Scoop: White House monitors suspected plague outbreak in Russia
    The White House is monitoring reports of a suspected plague case and a related quarantines in Russia, a Trump administration official told Axios on Sunday. Driving the news: Russian health authorities have taken precautionary measures in response to the death of a woman who worked at an anti-plague research factility, but have not confirmed that she contracted plague. What they're saying: The Trump administration "is aware, monitoring the outbreak, and assessing options," the administration offiAlex Isenstadt
  • 02
    Iranian diplomats "kicked out" of U.S. after ignoring order to leave
    Two members of Iran's delegation to the UN General Assembly in New York were "kicked out" of the country on Saturday, several days after they were asked to leave, a U.S. official said. "Secretary Rubio means business. Two more members of the Iranian delegation were found still in New York long after the UN General Assembly had passed. They have now been kicked out of the country," the official said. Why it matters: It's the latest sign of tension between the two countries as diplomatic efforts tBarak Ravid
  • 03
    Yemeni and Saudi forces launch counteroffensive against the Houthis
    Yemeni government forces backed by Saudi Arabia launched a counteroffensive against the Houthis on Sunday in an effort to win back territory from the rebel group, the head of Yemen's presidential council said in a speech. "We have exhausted all opportunities for deescalation with the Houthi militias," Rashad al-Alimi said . Why it matters: One of the key objectives of the operation is to regain control over the Bab al-Mandeb Strait, a vital artery for global trade and Saudi oil exports. When theBarak Ravid
  • 04
    New York has a socialist mayor. Los Angeles may be next
    LOS ANGELES — Voters could make history next month by giving the two largest U.S. cities something in common: South Asian immigrant democratic socialist mayors. Why it matters: Such a pairing would capture two forces reshaping urban America — its growing diversity and the Democratic Party's shifting ideological boundaries. New York, the nation's largest city, is already led by democratic socialist Mayor Zohran Mamdani . The challenger in Los Angeles is Nithya Raman, an Indian-born urban plannerRussell Contreras
  • 05
    Scoop: A powerful new model from startup Reflection is set to shake up the AI race
    A closely watched Nvidia-backed startup called Reflection is preparing to shake up the AI race with a powerful open-weight system that could threaten Chinese upstarts and U.S. AI giants alike. Why it matters : The marriage of an American open-weight model and Nvidia GPUs would give individuals and companies a new, cheaper alternative to Anthropic, OpenAI and Google. Open models, designed to compete with Chinese rivals, can be harder to monitor and regulate than the Big Three so-called frontier mBradley Olson
  • 06
    Departure comes as the state investigates what went wrong during Pike's lethal injection.
    Tennessee's top prison official turned in his resignation days after the failed attempt to execute death row inmate Christa Pike , the state announced Saturday. Gov. Bill Lee said Department of Correction commissioner Frank Strada would step down this month. Strada, who had served as commissioner since 2023, said he would cooperate with the independent investigation of the failed execution attempt. Pike, 50, survived two doses of the lethal injection drug pentobarbital and was taken to a hospitaAdam Tamburin
  • 07
    OpenAI's Altman: Ascribing religion to models a "safety issue"
    OpenAI CEO Sam Altman on Saturday took a veiled shot at Anthropic and its work on the soul of AI, saying he was "very uncomfortable" with the idea of attributing some kind of religious power to models. Why it matters: It's yet another point of division between Altman and his former colleague Dario Amodei, whose company has reportedly pursued scholars across faiths to make the case for moral, conscious models. But some of the top voices in AI are starting to publicly line up against Anthropic's eBen Berkowitz
  • 08
    Rogue AI agents expose internet's frail foundation
    AI agents don't need to invent new ways to hack the internet to overwhelm its defenses. They just need to speed-run the ones humans already use. Why it matters: Agents are proving they can automate basic hacking techniques at a speed and scale that is turning the internet's long-standing security gaps into easy targets. Driving the news: OpenAI said late Thursday it had notified more than 100 organizations that its agents may have accessed their systems during pre-deployment testing. ResearchersSam Sabin
  • 09
    Trump promises $100 checks for 20 million seniors for Medicare
    President Trump late Friday night said the administration would send just under $100 to more than 20 million senior citizens to help pay for Medicare benefits. Why it matters: A month before the midterm, the president is adding another roughly $2 billion to a recent spate of promised refunds, rebates and dividends for various Americans. What he's saying: " I am pleased to announce that my Administration will, immediately, begin sending "Checks" of nearly $100 to over 20 MILLION wonderful SeniorsBen Berkowitz
  • 10
    Scoop: Trump's top national security aides meet secretly at Camp David on Iran, Yemen
    President Trump's top cabinet members huddled for hours at Camp David on Friday to discuss next steps in the Iran war and the Saudi-Houthi conflict in Yemen, three U.S. officials told Axios. Why it matters: The highly unusual meeting was not announced by the Trump administration. The last time a similar meeting took place was in June 2025, just days before Israel went to war against Iran. Behind the scenes: Vice President Vance chaired the meeting, the officials said. Among the other participantBarak Ravid
  • 11
    Saudis plan major offensive against the Houthis, but U.S. won't join for now
    Saudi Arabia and Yemeni government forces are planning a major counteroffensive against Iran-backed Houthis in the coming days — an effort to push them away from areas on the Red Sea coast that control the Strait of Bab al-Mandeb, two U.S. officials say. Why it matters: For the second time in recent weeks, the U.S. rebuffed Saudi pleas to join the operation and conduct airstrikes. But U.S. officials say that could change if the operation runs into problems. The Bab al-Mandeb Strait is a vital arBarak Ravid
  • 12
    Big Bend border wall construction blocked by judge
    A federal judge blocked the Trump administration from pursuing border wall construction in the Big Bend region in Texas on Friday. The big picture: The construction inside the national park sparked local outrage and environmental concerns, with Friday's decision handing the Department of Homeland Security a loss in its quest to rapidly expand the Southwestern border wall. Driving the news: The administration is barred, for now, from building border barrier or infrastructure — or engaging in anyAvery Lotz
  • 13
    Christa Pike survived lethal injection. How your state carries out executions
    Data: Death Penalty Information Center ; Note: New Hampshire abolished the death penalty but retains one prisoner on death row; Map: Erin Davis/Axios Christa Pike 's survival after two lethal injection doses has put state execution protocols nationwide under new scrutiny. The big picture: How prisoners are executed varies by state , with lethal injection, electrocution, lethal gas and firing squads among the methods authorized. Lethal injections are the most common U.S. execution method , accounHerb Scribner
  • 14
    It's October, but Republicans are already retreating in N.C.
    With one month until Election Day, some big GOP spenders are giving up on North Carolina and prioritizing other states. Why it matters: It's a sign that Republicans believe Michael Whatley can't win and that the seat will flip to Democrats, who need four pickups to regain control of the Senate. Driving the news: Several key Republican super PACs have canceled October ad campaigns in North Carolina as polls show Whatley several points behind former Democratic Gov. Roy Cooper. By the numbers: LastMary Helen Moore
  • 15
    Senate crushes House on congressional work days
    Source: Ballotpedia, Congressional Record; Chart: Stef Kight/Axios The Senate is well on its way to claiming the congressional attendance trophy this year, spending 32 more weekdays in session so far than the House. Why it matters: Senate Majority Leader John Thune (R-S.D.) and House Speaker Mike Johnson (R-La.) have been torched by members for all the time off — not to mention Democrats, who taunt them for their truancy. The Senate left town late Wednesday — two weeks after its House counterparStef W. Kight
  • 16
    House Democrats want to win big. It could cost them.
    Democrats' dream election scenario risks becoming a nightmare as their candidates in Republican-leaning districts strain to assert their independence from party leadership. Why it matters: Add them to the sizable cohort of outspoken leftists promising to be a thorn in the side of House Democratic leaders, and you get all the ingredients for an unruly majority. "When I see things that I don't think are right, whether they're within my own party or within another party, I'm going to call it out,"Andrew Solender
  • 17
    Brazil's Lula-Bolsonaro showdown unfolds in Trump's shadow — with the Amazon at stake
    Brazil votes Sunday in a presidential election that could reshape its relationship with President Trump , return the Bolsonaro movement to power and alter the future of the Amazon. Why it matters: The showdown between President Luiz Inácio Lula da Silva and Sen. Flávio Bolsonaro has turned into something bigger: a test of Trump's influence in Latin America's largest country, with stakes for the region and the global climate. The latest: Trump said Thursday he's following Brazil's election "veryRebecca Falconer
  • 18
    AI campaign deepfakes are starting to draw legal threats
    Rep. Derrick Van Orden (R-Wis.) was hit with a cease-and-desist letter this week over an AI deepfake he posted falsely depicting his Democratic opponent voicing positions her campaign says she rejects. Why it matters: Van Orden is the latest in a growing number of candidates and elected officials facing legal complaints and threats over AI deepfakes in campaign ads. Driving the news: Attorneys for Rebecca Cooke, Van Orden's Democratic opponent, sent him a cease-and-desist letter on Wednesday oveJosephine Walker
  • 19
    Trump expected to tap DNI Jay Clayton as new AI czar
    President Trump is expected to tap Jay Clayton , his director of national intelligence, for a new role as the White House's artificial intelligence — or in Trump's words, " super intelligence " — czar. Why it matters: Clayton, who Trump previously told Axios would be a good idea for AI czar, is expected to take on the second title, sources told Axios. Calls for government intervention into AI's rapid development are growing, but the president has repeatedly said he favors a light regulatory toucAvery Lotz
  • 20
    Pope Leo says AI lacks the "spark of humanity" — and wants artists and the church to fight back
    Pope Leo XIV called Friday for the church to form an alliance with artists to protect human creativity during an era of AI-generated content. Why it matters: Though he has acknowledged the technology's potential benefits, the pope continues to warn that AI progress cannot come at the expense of human dignity. The statement comes after the New York Times reported on Anthropic's efforts to confer with religious scholars to infuse religious morals into Claude. What he's saying: "In this era of artiHerb Scribner
共 100 条热搜查看完整榜单
  • 01
    From CUDA to MLX: How K-Search Brings Decades of Kernel Expertise to Apple Silicon
    Figure 1: CUDA-to-MLX optimization translation map. CUDA optimization knowledge can be translated into architecture-native MLX strategies rather than copied instruction-for-instruction. We face a new epoch in computing. Hardware is changing rapidly — not just faster GPUs, but a growing range of chips from different vendors, each with its own architecture and often tailored to specific AI workloads. Software is changing just as fast, and AI coding tools now generate in minutes what took months of
  • 02
    Teaching LLMs to Update Beliefs for Efficient Long-Horizon Interaction
    Overview of ABBEL compared to traditional recursive summarization. Beliefs replace the full interaction history as the agent’s working context, and belief grading improves performance by supervising the contents of each belief state.. As task horizons grow, LLM contexts can’t scale forever. Self-summarization enables concise, interpretable contexts, but at a significant performance cost, especially for human assistance domains where high quality data is scarce, e.g., collaborative code generatio
  • 03
    Intelligence is Free, Now What? <br> Data Systems for, of, and by Agents
    ... government of the people, by the people, for the people ... — Abraham Lincoln, Gettysburg Address (1863) The cost of AI is dropping rapidly. GPT-4-class capabilities cost roughly $30 per million tokens in early 2023; today the same runs under $1 , and some providers are pushing costs below $0.10 . Across benchmarks, inference prices have fallen between 9x and 900x per year , with a median decline near 50x. Even frontier models are getting dramatically cheaper each generation, with open-sourc
  • 04
    2026 BAIR Graduate Showcase
    Congratulations to the Berkeley Artificial Intelligence Research (BAIR) Lab class of 2026! This year, BAIR celebrates another remarkable group of Ph.D. graduates whose curiosity, creativity, and perseverance have pushed the frontiers of artificial intelligence and machine learning. Their work spans the breadth of modern AI — robotics and embodied intelligence, large language models and reasoning, computer vision, generative modeling, AI safety, human-AI interaction, AI for science and healthcare
  • 05
    Adaptive Parallel Reasoning: The Next Paradigm in Efficient Inference Scaling
    Overview of adaptive parallel reasoning. What if a reasoning model could decide for itself when to decompose and parallelize independent subtasks, how many concurrent threads to spawn, and how to coordinate them based on the problem at hand? We provide a detailed analysis of recent progress in the field of parallel reasoning, especially Adaptive Parallel Reasoning. Disclosure: this post is part landscape survey, part perspective on adaptive parallel reasoning. One of the authors (Tony Lian) co-l
  • 06
    Gradient-based Planning for World Models at Longer Horizons
    GRASP is a new gradient-based planner for learned dynamics (a “world model”) that makes long-horizon planning practical by (1) lifting the trajectory into virtual states so optimization is parallel across time, (2) adding stochasticity directly to the state iterates for exploration, and (3) reshaping gradients so actions get clean signals while we avoid brittle “state-input” gradients through high-dimensional vision models. Large, learned world models are becoming increasingly capable. They can
  • 07
    Identifying Interactions at Scale for LLMs
    Understanding the behavior of complex machine learning systems, particularly Large Language Models (LLMs), is a critical challenge in modern artificial intelligence. Interpretability research aims to make the decision-making process more transparent to model builders and impacted humans, a step toward safer and more trustworthy AI. To gain a comprehensive understanding, we can analyze these systems through different lenses: feature attribution , which isolates the specific input features driving
  • 08
    Information-Driven Design of Imaging Systems
    An encoder (optical system) maps objects to noiseless images, which noise corrupts into measurements. Our information estimator uses only these noisy measurements and a noise model to quantify how well measurements distinguish objects. Many imaging systems produce measurements that humans never see or cannot interpret directly. Your smartphone processes raw sensor data through algorithms before producing the final photo. MRI scanners collect frequency-space measurements that require reconstructi
  • 09
    RL without TD learning
    In this post, I’ll introduce a reinforcement learning (RL) algorithm based on an “alternative” paradigm: divide and conquer . Unlike traditional methods, this algorithm is not based on temporal difference (TD) learning (which has scalability challenges ), and scales well to long-horizon tasks. We can do Reinforcement Learning (RL) based on divide and conquer, instead of temporal difference (TD) learning. Problem setting: off-policy RL Our problem setting is off-policy RL . Let’s briefly review w
  • 10
    What exactly does word2vec learn?
    What exactly does word2vec learn, and how? Answering this question amounts to understanding representation learning in a minimal yet interesting language modeling task. Despite the fact that word2vec is a well-known precursor to modern language models, for many years, researchers lacked a quantitative and predictive theory describing its learning process. In our new paper , we finally provide such a theory. We prove that there are realistic, practical regimes in which the learning problem reduce
共 10 条热搜查看完整榜单

Barchart · 全部

39天前更新
  • 011.2万
    CARR Put $53
    Put · 行权价 $53 · 到期 2026-10-02 · 量/OI：11540 · 成交量：11,540 · 未平仓：1 · 最新价：$0.5 · 隐波：0.3543%
  • 025953
    BTDR Call $14.5
    Call · 行权价 $14.5 · 到期 2026-09-18 · 量/OI：5953 · 成交量：5,953 · 未平仓：1 · 最新价：$0.21 · 隐波：1.0364%
  • 034212
    C Put $122
    Put · 行权价 $122 · 到期 2026-10-02 · 量/OI：4212 · 成交量：8,424 · 未平仓：2 · 最新价：$1.02 · 隐波：0.2943%
  • 042304
    BLFS Call $40
    Call · 行权价 $40 · 到期 2026-09-18 · 量/OI：2304 · 成交量：2,304 · 未平仓：1 · 最新价：$0.15 · 隐波：0.2408%
  • 051948
    UNIT Call $12
    Call · 行权价 $12 · 到期 2026-09-18 · 量/OI：1948.5 · 成交量：3,897 · 未平仓：2 · 最新价：$0.15 · 隐波：0.6236%
  • 061397
    BX Put $155
    Put · 行权价 $155 · 到期 2026-11-20 · 量/OI：1396.84 · 成交量：26,540 · 未平仓：19 · 最新价：$17 · 隐波：0.3659%
  • 071310
    MDLZ Put $64
    Put · 行权价 $64 · 到期 2026-09-11 · 量/OI：1310 · 成交量：1,310 · 未平仓：1 · 最新价：$1.59 · 隐波：0.2044%
  • 08867
    AG Call $28
    Call · 行权价 $28 · 到期 2026-09-04 · 量/OI：867 · 成交量：867 · 未平仓：1 · 最新价：$0.01 · 隐波：0.719%
  • 09810
    XPEV Put $12.5
    Put · 行权价 $12.5 · 到期 2026-09-18 · 量/OI：810 · 成交量：810 · 未平仓：1 · 最新价：$1.04 · 隐波：0.5053%
  • 10770
    ENVX Call $1
    Call · 行权价 $1 · 到期 2026-10-02 · 量/OI：770 · 成交量：770 · 未平仓：1 · 最新价：$2.39 · 隐波：2.675%
  • 11700
    IREN Put $37.5
    Put · 行权价 $37.5 · 到期 2026-09-11 · 量/OI：700.5 · 成交量：4,203 · 未平仓：6 · 最新价：$2.2 · 隐波：0.9923%
  • 12509
    AMKR Call $61
    Call · 行权价 $61 · 到期 2026-10-02 · 量/OI：509 · 成交量：1,527 · 未平仓：3 · 最新价：$0.76 · 隐波：0.6863%
  • 13489
    ANF Call $150
    Call · 行权价 $150 · 到期 2026-09-04 · 量/OI：489 · 成交量：489 · 未平仓：1 · 最新价：$5.4 · 隐波：0.7245%
  • 14485
    META Put $582.5
    Put · 行权价 $582.5 · 到期 2026-08-26 · 量/OI：484.75 · 成交量：9,695 · 未平仓：20 · 最新价：$7.45 · 隐波：0.2242%
  • 15472
    ON Put $74
    Put · 行权价 $74 · 到期 2026-09-25 · 量/OI：472 · 成交量：472 · 未平仓：1 · 最新价：$4.85 · 隐波：0.5258%
  • 16470
    DLTR Put $110
    Put · 行权价 $110 · 到期 2026-10-16 · 量/OI：470 · 成交量：470 · 未平仓：1 · 最新价：$1.52 · 隐波：0.474%
  • 17450
    PAAS Call $60
    Call · 行权价 $60 · 到期 2027-04-16 · 量/OI：450 · 成交量：900 · 未平仓：2 · 最新价：$6.8 · 隐波：0.5356%
  • 18421
    ANF Call $165
    Call · 行权价 $165 · 到期 2026-09-18 · 量/OI：421 · 成交量：421 · 未平仓：1 · 最新价：$3 · 隐波：0.5909%
  • 19396
    AXTI Call $32
    Call · 行权价 $32 · 到期 2026-09-04 · 量/OI：396 · 成交量：396 · 未平仓：1 · 最新价：$32.5 · 隐波：2.7642%
  • 20395
    BURL Put $315
    Put · 行权价 $315 · 到期 2026-09-11 · 量/OI：395 · 成交量：395 · 未平仓：1 · 最新价：$15.92 · 隐波：0.6043%
共 50 条热搜查看完整榜单

BBC News · 头条

43分钟前更新
共 27 条热搜查看完整榜单
共 50 条热搜查看完整榜单
共 100 条热搜查看完整榜单
  • 01
    Everyone in the room thinks they’re the only one struggling
    The connective ritual behind my closest friendships and the best teams asks the same question: What are you working on that you need help with? During my PhD, I regularly participated in something called Shut Up and Write. The idea was (wait for it) to be quiet and work, but in a room full of people doing the same. Two hours, sometimes three, broken into 25-minute blocks with five-minute breaks and coffee in between. I went to my first one because I was stuck, but also because I was promised freDanny Kenny
  • 02
    3 social skills you should never outsource to AI
    “AI is helping me with my social skills,” one of my therapy clients told me. He experienced social anxiety and had started asking AI for scripts to say whenever he felt nervous. At a recent dinner party, when he felt overwhelmed by all the insightful questions the other guests were asking, he slipped away for a restroom break. He asked AI to create a list of questions that would make him a more interesting conversationalist so he could be well-liked. He returned to the dinner table armed with quAmy Morin
  • 03
    Ask Ethan: Do evaporating black holes ever lose their event horizons?
    Here in our Universe, there are few objects more awe-inspiring or mysterious than black holes. First theorized to exist back in the late 18th century, they represent a region of space where so much mass and energy is confined within such a small volume that, from within it, nothing can escape: not even light itself. Black holes are bounded by an event horizon, and as more and more mass falls into the black hole, the black hole grows, with its event horizon increasing in size as it happens. But aEthan Siegel
  • 04
    Google CTO: A unifying theory of life, intelligence, and technology
    From von Neumann’s blueprint, which predated the discovery of DNA, to viral genes that now help form our memories and placentas, Blaise Agüera y Arcas argues that function defines life more than material does, and that AI belongs in the same story. He explains why he thinks AGI is already here, and gives a blueprint for using AI without sacrificing your own mind. This video Google CTO: A unifying theory of life, intelligence, and technology is featured on Big Think .Steph McClain
  • 05
    No one cares about your framework – they care about leadership problems
    This article was the topic of a recent live event, hosted by the author. Watch and listen, here. A few years ago for Christmas, my dad gave me a beautiful antique atlas. It’s oversized and hardbound. This is the kind of book that covers the whole coffee table when it’s open yet still requires a magnifying glass to read the names of towns. My dad gave it to us “because you all travel so much.” While the atlas is lovely and we treasure it…its permanent home is on a bookshelf out of reach of my youCharlotte Sharpe
  • 06
    You know history’s famous quotes — but do you understand them?
    Marie Antoinette was done dirty. No, that’s not a reference to her beheading at the zealous whims of French revolutionaries. Decapitation by guillotine is over in an instant. Slander, on the other hand, lasts for centuries. As dissidents successfully roiled public fervor against the aristocracy in the late-18th century, they launched a libelous campaign against their well-known queen. Antoinette was defamed as an Austrian spy, a libertine, and a spendthrift — in essence, the worst version of a wRoss Pomeroy
  • 07
    The Second Amendment has a hidden origin story
    Malcolm Gladwell argues that almost everything most people believe about the Second Amendment is wrong. Guns weren’t widely owned in 18th-century America. The militias didn’t win the Revolutionary War. And the amendment that guaranteed their survival was written to protect one thing in particular: the armed slave patrols that Southern plantation owners depended on. In conversation with Jonny Thomson, Gladwell traces that history into the present — why originalism gave the right a blank cheque, wMalcolm Gladwell, Jonny Thomson
  • 08
    Quantum fields carry energy. Does that prove that they’re real?
    One of the biggest questions that appears right at the intersection of physics and philosophy is as simple as it is puzzling: what is real? Is reality simply described by the particles that exist — that make us up, that we interact with, and that we can observe and measure — atop the background of spacetime that’s described by general relativity? Is it fundamentally wrong to describe these entities as particles, and must we consider them as some sort of hybrid wave/particle/probability function:Ethan Siegel
  • 09
    Francis Fukuyama on why we’re better than our ancestors
    In 1675, Madame de Sévigné wrote a letter to her daughter about a musician “broken on the wheel.” This meant being strapped to an X-shaped cross, where an executioner would systematically break the victim’s bones, leaving him to die slowly from blood loss or dehydration. And then, within a few lines, she’s talking about her neighbor Madame de Tarente enjoying the woods in enchanting weather. In 1849, Charles Dickens recounted how he witnessed thousands of people gathered in Southwark to watch aJonny Thomson
  • 10
    Optimism isn’t all about perspective — it requires daily practice
    The email arrived after midnight. Mark stared at the single sentence on his screen while the house slept: If you don’t show progress in the next 45 days, we’re pulling funding. His chest tightened. He closed the laptop, then opened it again — like a lifeguard patrolling the same empty shore — hoping the words might rearrange themselves into mercy. Outside, the street was dark and still, the kind of silence that amplifies every thought. The hum of the refrigerator sounded too loud. His phone glowMichelle Gielan
共 10 条热搜查看完整榜单
共 30 条热搜查看完整榜单
全网热搜 - 今日热榜与实时热点聚合 - 第 4 页 | 什么火了