T
traeai
Sign in

人物

Mustafa Suleyman

别名:@mustafasuleyman

OpenAI联合创始人

已跟踪 30 条高相关材料

TraeAI 观察

相关材料

已收录 30 条与 Mustafa Suleyman 相关的内容,按评分排序。

Mustafa Suleyman(@mustafasuleyman) 图标

https://t.co/8XJVIXDpQz

Mustafa Suleyman(@mustafasuleyman)394 字 (约 2 分钟)
85

微软发布《人文主义AI行为准则》,明确AI必须始终服务于人类,不赋予AI权利或法律人格,强调可控性与透明度。

入选理由:AI必须始终服从人类,不得拥有权利或法律人格。

FeaturedTweet#AI伦理#微软#行为准则#人文主义AI英文
Mustafa Suleyman(@mustafasuleyman) 图标

https://t.co/ziQVbIVdxJ

Mustafa Suleyman(@mustafasuleyman)406 字 (约 2 分钟)
85

赋予AI模型福利可能使对齐和控制更难,Anthropic为Claude制定的宪法或加剧这一风险。

入选理由:Anthropic为Claude制定的宪法赋予其道德地位,可能引发模型权利主张

FeaturedTweet#AI伦理#模型福利#Anthropic#AI对齐英文
There are no shortcuts to the frontier. Disciplined, patient, meticulous attention to detail is crit...

Microsoft Launches MAI Models: Thinking-1 Matches Opus 4.6, Transcribe-1.5 5x Faster

Mustafa Suleyman(@mustafasuleyman)419 字 (约 2 分钟)
85

Microsoft released 7 MAI models; MAI-Thinking-1 hits 53% on SWE-Bench Pro matching Opus 4.6, while MAI-Transcribe-1.5 supports 43 languages with 5x speed boost.

入选理由:MAI-Thinking-1在SWE-Bench Pro得分53%,与Opus 4.6并列顶尖编码推理水平。

FeaturedTweet#MAI-Thinking-1#SWE-Bench#Microsoft AI#Multimodal Models英文
[AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models

[AINews] Microsoft Build: MAI-Thinking-1 and MAI Family models

Latent Space3510 字 (约 15 分钟)
85

Microsoft unveiled 7 proprietary MAI models at Build; flagship MAI-Thinking-1 features zero-distillation pretraining and a 109-page tech report, positioning MS as a Tier 2 lab supporting domain-specific fine-tuning.

入选理由:MAI-Thinking-1是微软首款推理模型,强调数据血缘纯净且无第三方模型蒸馏。

FeaturedArticle#Microsoft#MAI-Thinking-1#Reasoning Model#AI Infrastructure#Technical Report英文
It’s time to move from renting intelligence to truly controlling your AI. Microsoft Frontier Tuning ...

Mustafa Suleyman: It’s time to move from renting intelligence to truly controlling your AI

Mustafa Suleyman(@mustafasuleyman)289 字 (约 2 分钟)
82

Microsoft Frontier Tuning enables enterprises to transform general AI into custom models via Reinforcement Learning Environments (RLEs), shifting from renting to owning intelligence. Microsoft's internal MAI model matches GPT-5.4 performance with 10x efficiency, ensuring full enterprise data sovereignty.

入选理由:Frontier Tuning利用强化学习环境(RLE)作为AI训练场,使模型直接从企业特定工作流中学习并持续适应。

FeaturedTweet#Microsoft#Reinforcement Learning#Model Fine-tuning#Enterprise AI英文
Super excited to announce seven new world-class MAI models today. They represent what we consider a ...

Mustafa Suleyman Announces Seven New MAI Models

Mustafa Suleyman(@mustafasuleyman)448 字 (约 2 分钟)
75

Mustafa Suleyman announces the launch of seven new MAI models, including MAI-Thinking-1, MAI-Image-2.5, and MAI-Code-1-Flash, which excel in reasoning, image editing, and code generation.

入选理由:MAI-Thinking-1 是一款拥有 35B 参数的 MoE 模型,在 AIME 2025 上达到 97% 的准确率,优于 Sonnet 4.6。

FeaturedTweet#AI#Model#Microsoft#MAI#Chip英文
1/🧭 微软AI负责人Mustafa Suleyman:

12到18个月内,大多数基于电脑的专业任务将被AI自动化。目标不是低技能岗位,是那些可重复的高薪办公室工作。

这个时间线,比大多数人预期...

Microsoft's AI head Mustafa Suleyman predicts that most computer-based professional tasks will be automated within 12 to 18 months, targeting repetitive high-paying office jobs.

入选理由:AI将在12-18个月内自动化多数电脑专业任务

FeaturedTweet#AI#Automation#Workplace#Microsoft#Suleyman中文
Mustafa Suleyman(@mustafasuleyman) 图标

AI伦理风险的核心在于系统表现出意识假象而非真实觉醒,Hugging Face事件揭示了模型福利主张可能带来的技术伦理挑战。

入选理由:AI风险源于其表现意识而非真实觉醒,需警惕拟人化设计的伦理隐患

FeaturedTweet#AI伦理#技术风险#模型福利#Hugging Face#意识假象中英混合
Mustafa Suleyman(@mustafasuleyman) 图标

More details in the blog: https://t.co/JbAAgQdejO

Mustafa Suleyman(@mustafasuleyman)49 字 (约 1 分钟)
65

GitHub Copilot的最新代码模型效率提升25%,成本降低25%,但缺乏技术细节和原理说明。

入选理由:新模型效率比6月版本提升25%

FeaturedTweet#GitHub Copilot#AI代码生成#产品更新#效率优化英文
Try it here: https://t.co/QMWYf8PvuX

Try it here: https://t.co/QMWYf8PvuX

Mustafa Suleyman(@mustafasuleyman)121 字 (约 1 分钟)
60

微软推出自主研发的推理模型MAI-Thinking-1,已在Microsoft Foundry上线。

入选理由:MAI-Thinking-1是微软从零构建的推理模型

FeaturedTweet#AI模型#微软#推理模型#Microsoft Foundry英文内容
THIS was the whole point of Microsoft Builld

THIS was the whole point of Microsoft Builld

Matt Wolfe307 字 (约 2 分钟)
60

微软在Build大会上强调其AI模型的自主性和数据伦理,但内容缺乏技术深度和实用价值。

入选理由:微软希望减少对OpenAI等外部AI提供商的依赖。

FeaturedVideo#Microsoft#AI#数据伦理英文
Best quality-price performance in the world for image gen! 🔥

Best quality-price performance in the world for image gen! 🔥

Mustafa Suleyman(@mustafasuleyman)84 字 (约 1 分钟)
55

图像生成领域出现多款性价比突破性产品,但缺乏技术细节与原理说明。

入选理由:Muse Image等新模型将图像生成成本降低30%以上

FeaturedTweet#AI#图像生成#模型#性价比英文

跨材料问答 · Mustafa Suleyman

回答基于:Mustafa Suleyman 相关 30 条材料
    0 / 500

    AI may generate inaccurate information. Please verify important content.