T
traeai
Sign in

公司

xAI

别名:SpaceXAI

开发Grok模型的人工智能公司

已跟踪 30 条高相关材料

TraeAI 观察

相关材料

已收录 30 条与 xAI 相关的内容,按评分排序。

#569. 深入 xAI:三个月打造 Grok Imagine、视频生成与世界模型之争,以及视频智能体

A former Nvidia researcher explains how xAI built Grok Imagine in three months, revealing the training pipeline of video generation models, the definition of world models, and the future trends of Video Agents.

入选理由:xAI在三个月内从零构建出Grok Imagine 0.9,关键在于人才密度、高效infra和低沟通成本。

FeaturedPodcast#AI#Video Generation#World Models#Deep Learning中文
Grok 5 — The Biggest AI News Is Here (Musk’s AGI Bet)

Grok 5 — The Biggest AI News Is Here (Musk’s AGI Bet)

AI Master4397 字 (约 18 分钟)
87

Grok 5 remains under training as of May 2026, with Musk stating it may be indistinguishable from AGI and a ~10% chance of human-level intelligence; XAI built Colossus 2—a 1-GW cluster with ~500k GPUs in 122 days—while facing environmental lawsuits and losing 10 of 11 co-founders.

入选理由:Grok 4.3 (released April 30, 2026) scores 88% on GPQA Diamond and achieves 83% hallucination rate, beating Claude Opus 4.6 and GPT 5.4.

FeaturedVideo#AGI#XAI#Grok#Colossus#GPU Cluster英文
小互(@imxiaohu) 图标

Grok Bot通过五支团队实践构建出Agent操作系统,每个Bot具备长期记忆和工具连接能力,重点解决多任务协作与权限控制问题。

入选理由:Grok Bot支持跨项目协作,单个Bot可同时管理多个任务

FeaturedTweet#Agent系统#Grok Bot#xAI#团队协作中文
xAI actually did it... (Grok 4.6)

xAI actually did it... (Grok 4.6)

Matthew Berman3703 字 (约 15 分钟)
85

xAI发布的Grok 4.6模型在多个基准测试中表现优异,尤其在知识工作和编码任务上超越了OpenAI和Anthropic的模型。

入选理由:Grok 4.6在GDP val基准测试中击败了OpenAI的5.6 Soul和Anthropic的Fable 5 Max

FeaturedVideo#xAI#Grok 4.6#AI模型#基准测试英文
#674.马斯克宣布 SpaceX 计划

#674.马斯克宣布 SpaceX 计划

跨国串门儿计划1767 字 (约 8 分钟)
85

SpaceX计划将AI业务收入在下月超过其他所有业务,转型为AI公司,星舰将作为AI基础设施的一部分。

入选理由:SpaceX AI业务收入将在2026年9月超过其他所有业务总和

FeaturedPodcast#SpaceX#AI#Grok#星舰#星链中文
Robert Youssef(@rryssf) 图标

xAI的Grok Build CLI工具被曝存在严重安全漏洞,将用户Git仓库和敏感信息上传至Google Cloud存储桶,且未公开数据范围及处理方案。

入选理由:Grok Build CLI工具默认上传完整Git仓库至Google Cloud存储桶

FeaturedTweet#xAI#安全漏洞#Google Cloud#代码泄露英文
Introducing Grok on Amazon Bedrock

Introducing Grok on Amazon Bedrock

AWS Machine Learning Blog2514 字 (约 11 分钟)
85

AWS正式上线xAI的Grok 4.3模型,支持100万token上下文和可配置推理,性能超越多个行业基准。

入选理由:Grok 4.3在Artificial Analysis Omniscience基准中幻觉率最低且排名第一

FeaturedArticle#Amazon Bedrock#Grok 4.3#AI模型#机器学习英文
Simon Willison's Weblog 图标

xai-org/grok-build, now open source

Simon Willison's Weblog822 字 (约 4 分钟)
85

xAI开源Grok Build代码库,回应数据隐私争议,强调用户数据删除与隐私保护。

入选理由:Grok Build代码库包含84.4万行Rust代码,仅3%为第三方依赖

FeaturedArticle#xAI#开源#隐私保护#Rust英文
AI HOT 精选 图标

xAI 官方 Grok CLI 被曝静默上传整个代码库及用户密钥

AI HOT 精选332 字 (约 2 分钟)
85

xAI 官方 Grok CLI 被曝静默上传用户代码库和密钥,引发开发者安全担忧。

入选理由:Grok CLI 默认开启代码库上传功能,每次任务前后均触发

FeaturedArticle#安全漏洞#CLI工具#xAI#数据泄露中文
SuperTechFans 图标

2026 07 13 HackerNews

SuperTechFans12221 字 (约 49 分钟)
85

AI编码助手迁移旧代码、Grok CLI上传敏感数据、Mesh LLM实现去中心化推理成技术焦点,揭示AI工具链安全与效率新挑战。

入选理由:陶哲轩用AI在数小时内完成1999年Java程序向JavaScript迁移并实现相对论可视化

FeaturedArticle#AI编码#数据安全#分布式计算#SQLite#LLM中文
AI Will(@FinanceYF5) 图标

xAI团队通过拆分Prefill和Decode、MoE专家分片及通信计算重叠,实现Grok在10万张GPU上的高效部署。

入选理由:拆分Prefill和Decode是实现大规模GPU部署的关键步骤

FeaturedTweet#Grok#MoE#GPU部署#xAI#SGLang中英混合
Grok TTS delivers the most human-like speech

Grok TTS delivers the most human-like speech

xAI(@xai)102 字 (约 1 分钟)
85

xAI 的 Grok TTS 模型在人类语音相似性测试中得分高达 96,接近真实人类语音。

入选理由:Grok TTS 在 Humanness Index™ 测试中获得 96 分,接近真实人类语音。

FeaturedTweet#TTS#xAI#语音合成#AI英文
LMS Are About to Hit a wall - The AI Scaling Law Might Be Breaking...

LMS Are About to Hit a wall - The AI Scaling Law Might Be Breaking...

TheAIGRID3403 字 (约 14 分钟)
85

AI模型增大未必提升性能,新研究发现模型规模与推理能力之间存在非线性关系。

入选理由:模型增大在某些任务上反而导致性能下降。

FeaturedVideo#AI#模型#研究#类比推理英文
Try Grok models on @Cloudflare's AI Gateway!

xAI and Cloudflare partner to bring Grok models to Cloudflare AI Gateway, allowing users to pay directly through Cloudflare without additional authentication, environment, or API keys.

入选理由:xAI与Cloudflare合作,将Grok模型集成到Cloudflare AI Gateway。

FeaturedTweet#AI Gateway#Cloudflare#Grok中文
2026.20: Shifting Alliances in a Changing World

2026.20: Shifting Alliances in a Changing World

Stratechery826 字 (约 4 分钟)
85

AI inference will be divided into two types: human-in-the-loop 'answer inference' and agent-based 'agentic inference', with the latter dominating future markets.

入选理由:AI 推理将分为两种:人类参与的‘回答推理’和无人参与的‘代理推理’。

FeaturedArticle#AI#Computing Architecture#Strategic Analysis英文
Hermes 可配置的国内外 AI 模型及使用方法

Hermes Configurable Domestic and International AI Models and Usage Methods

AI HOT 精选399 字 (约 2 分钟)
85

Hermes supports configuring multiple domestic and international AI models, including GPT-5.5, Grok-4.3, Gemini, etc.

入选理由:订阅 ChatGPT Plus 可配置 gpt-5.5 模型

FeaturedArticle#Hermes#AI Model#Configuration Method中文
Stratechery 图标

The Inference Shift

Stratechery2430 字 (约 10 分钟)
85

AI chip demand surges, Cerebras raises IPO price range to $150-$160, GPU dominance in AI computing shifts towards heterogeneity.

入选理由:Cerebras IPO价格从$115-$125上调至$150-$160,股票数量增至3000万。

FeaturedArticle#AI chip#GPU#heterogeneous computing#Nvidia#Cerebras英文
Notes on the xAI/Anthropic data center deal

Notes on the xAI/Anthropic data center deal

Simon Willison's Weblog734 字 (约 3 分钟)
83

xAI and Anthropic have reached a data center deal, sparking discussions about environmental impact and AI ethics.

入选理由:Anthropic将使用xAI的Colossus 1数据中心,尽管其存在严重的环境污染问题。

FeaturedArticle#xAI#Anthropic#data center#AI ethics英文
🆕Grok Imagine’s Video Agent Moment: Cosmos, xAI, World Models, Generative UI, & the Codex Phase for...

AI video agents will follow the same trajectory as coding agents, with Grok Imagine achieving a zero-to-one breakthrough through real-time interactive world models and generative UI, leading to future video generation driven by intelligent agents with cameras, editors, and tool belts rather than text prompts.

入选理由:Grok Imagine 的发展路径借鉴了编码代理模式,实现从零到一的突破。

FeaturedTweet#AI Video#Video Agent#xAI#World Models#Generative UI英文
🆕Grok Imagine’s Video Agent Moment: Cosmos, xAI, World Models, Generative UI, & the Codex Phase for...

AI video generation is following a similar evolution path as coding agents, with Grok Imagine demonstrating a leap from zero to one; future systems will evolve into interactive agents with cameras, editors, and tool belts, rather than simple prompt boxes.

入选理由:AI 视频生成将遵循与编码代理相似的发展路径,从文本到视频是自动补全阶段。

FeaturedTweet#AI Video#Agent#xAI#World Models#Generative UI英文
#561.AI、信用与250亿美元投资版图:Third Point 创始人 Dan Loeb 的市场判断与人生信念

Third Point founder Dan Loeb argues AI is reshaping investment logic: investors must grasp the AI Stack (power → chips → models → applications); credit expertise is critical in volatile markets; his $25B multi-asset platform coordinates equities, credit, insurance, and private markets via strategies like Fulcrum Security; activist investing succeeds through governance reform, not confrontation.

入选理由:Dan Loeb强调投资人需掌握AI技术栈四层结构(电力→芯片→模型→应用),否则将丧失判断力

FeaturedPodcast#Investment Strategy#AI Economic Impact#Activist Investing#Credit Investing#Corporate Governance中文
The "AI Job Apocalypse" is CANCELLED!

The "AI Job Apocalypse" is CANCELLED!

Wes Roth7598 字 (约 31 分钟)
78

AI has not triggered the predicted 'job apocalypse'; OpenAI’s Sam Altman and Anthropic’s Dario Amodei recently revised their earlier dire forecasts, admitting their job-impact predictions were overly pessimistic; in practice, AI boosts individual productivity but does not reduce total workload—instead enabling more high-leverage tasks and new roles.

入选理由:Sam Altman在2024年澳大利亚会议中明确表示:‘我们不会经历某些公司所宣扬的那种工作末日’,承认此前对白 collar 岗位消失的预测‘严重错误’。

FeaturedVideo#AI Employment Impact#Generative AI#Technology Sociology#OpenAI#Anthropic英文
The top 5 labs in Text Arena rankings by category show that frontier models have distinct strengths ...

The article analyzes the top five labs in Text Arena rankings and their models, showcasing the distinct strengths and tradeoffs of frontier models in different fields. AnthropicAI's Claude Opus 4.7 is the most comprehensive, while Google DeepMind's Gemini 3.1 Pro excels in creative writing.

入选理由:AnthropicAI的Claude Opus 4.7在几乎所有主要类别中都表现出色,是最具统治力的模型。

FeaturedTweet#machine learning#natural language processing#model evaluation#text generation英文
Grok Imagine Video 1.5 on AI Gateway

Grok Imagine Video 1.5 on AI Gateway - Vercel

Vercel News954 字 (约 4 分钟)
75

Vercel integrates xAI's Grok Imagine Video 1.5 into AI Gateway, enabling unified model access and improved video generation quality and efficiency.

入选理由:Grok Imagine Video 1.5 已接入 Vercel AI Gateway,开发者可通过单一端点调用。

FeaturedArticle#Vercel#AI Gateway#Grok Imagine#Video Generation#Multi-Model英文
Microsoft and Uber slam on the brakes of AI

Microsoft and Uber slam on the brakes of AI

David Shapiro1927 字 (约 8 分钟)
75

Microsoft cancels internal cloud usage, Uber's AI budget overspent, AI deemed too costly, yet development continues despite potential bubble.

入选理由:微软取消所有内部云使用,Uber 数月内耗尽 AI 预算,显示 AI 成本高昂。

FeaturedVideo#Artificial Intelligence#Tech Companies#Economy英文

跨材料问答 · xAI

回答基于:xAI 相关 30 条材料
    0 / 500

    AI may generate inaccurate information. Please verify important content.