Smile//77

1. AI模型算力Claude&Gemini&GPT资源，云平台AWS，Azure，GCP资源 3. 各web3交易所最高返佣比例

[Smile//77 @smile77go]

It's been 72 hours since I placed a USDT withdrawal order by BSC. How much longer can I wait? What else can I do? I need a response.@BitMartExchange @sheldonbitmart @BitMart_Futures

[Smile//77 @smile77go]

@endrastaty @lisandro4 It's been 72 hours since I placed a USDT withdrawal order by BSC. How much longer can I wait? What else can I do? I need a response.

[Smile//77 @smile77go]

@DENA_tho @lisandro4 It's been 72 hours since I placed a USDT withdrawal order by BSC. How much longer can I wait? What else can I do? I need a response.

[Smile//77 @smile77go]

@chairbtc 经常刷总裁，互关一下呗

[Smile//77 @smile77go]

2025年即将到来，期待随着1月20号川普上任美国总统，若美国将比特币作为资产储备的提案通过，牛市能继续狂飞😍Like 2017s~ #BTC #ETH #SOL #ONDO

[Bruce J @BTCBruce1]

bull market has 3 stages:
stage1: small bump up
This part is just to get people’s attention. Some nervous players in the market might sell and leave.

stage2: big jump
Newbies start coming in and making money, and those who sold earlier come running back.

stage3: crazy jump
everyone crazy and all-in, and then bull market ends

Right now, we’re still in stage1
#Bitcoin #Comedian #Ban #Doge

[Smile//77 @smile77go]

@Morris_LT 短视频的段子都出来了哈哈

[Smile//77 @smile77go]

Token2049-SG的快图人山人海片段：
名副其实的加密货币和 Web3 领域盛会，愿更多的叙事和财富效应在第四季度如期涌现。
愿Web3的未来在日新月异的迭代中，继续在金融史的有限认知中抒发出无限的浓墨重彩。
接受一切正面和消极，我们只向前看~
#Token2049 #BTC #ETH #SOL #DePINAsia

“它们是好的而不是灾难的计划，给你一个未来和希望。” ---（耶利米书 29:11）

[Smile//77 @smile77go]

美联储降息50个基点（0.5%）基于以下目的：
1. 经济刺激 2.缓解经济衰退 3. 降低失业率 ， 除了会对整个传统金融市场产生广泛影响比如：股市上涨，债市变化，美元贬值等等。
对加密市场的也会产生潜在的影响：
1.资金流入加密市场： 降息可能导致传统金融市场的回报率降低，推动一些投资者寻求更高回报的资产，例如比特币和其他加密货币。这可能导致资金流入加密市场，推高其价格。
2.避险资产的吸引力： 当降息被视为应对经济不确定性的信号时，投资者可能会将加密货币视为避险资产，尤其是在对法币贬值的担忧加剧时。比特币等加密货币有时被称为“数字黄金”，因此在降息环境中，其避险属性可能增强。
3.去中心化金融（DeFi）和贷款： 降息可能使传统金融机构的利率极低，从而让加密货币领域的去中心化金融（DeFi）生态系统提供的贷款和收益率产品更具吸引力，吸引更多的投资者进入DeFi平台。
以上都是理论层面上的浅搬家探讨，而未来市场怎么运行和发展，以及接下来是否会受日本议息的影响 ，一切都交给时间去验证。在WEB3的圈子里，风云变幻，定数反转，就是它最大的魅力！#BTC #美联储降息 #牛市 #ETH #宏观经济

[Ondo Finance @Ondo]

The growth of RWA adoption is accelerating, led by tokenized US Treasuries per @BinanceResearch. 📈

The market value for tokenized non-stablecoin RWAs has surpassed $12B, "representing continued investor interest in blockchain-based tokenization of traditional assets".

Traditional financial institutions are increasingly viewing tokenization as an opportunity to leverage blockchain technology to access a new distribution channel.

📚 Read more: https://t.co/SzAVBI7L3W

[vitalik.eth @VitalikButerin]

I take this seriously. Starting next year, I plan to only publicly mention (in blogs, talks, etc) L2s that are stage 1+, with *maybe a short grace period* for new genuinely interesting projects.

It doesn't matter if I invested, or if you're my friend; stage 1 or bust.

Multiple ZK-rollup teams have told me they're on track to be stage 1 by year end. I'm excited to see that happen!

Of course we should not throw away training wheels become we're actually confident that the proof systems are secure; that would be irresponsible. But stage 1 (75% threshold on council to override the proof system, 26%+ of council must be outside the rollup team) is a very reasonable moderate milestone. The multisigs I'm in have not had a single liveness failure in years, let alone 26%.

The era of rollups being glorified multisigs is coming to an end. The era of cryptographic trust is upon us.

[Smile//77 @smile77go]

#AI 午报 /8月29日

1.Anthropic宣布面向所有 Claude 用户开放 Artifacts 功能，功能相当于创建一个动态工作区，用户和 Claude 对话让其生成文档、代码、矢量图，甚至是简单的游戏过程中，Artifacts 会出现聊天界面旁边，方便用户实时査看，选代和创建你的作品。

2.OpenAl正在开发"草莓“和“猎户座”(Oron)两种模型。草莓主要用于解决复杂推理任务，尤其是在数学和编程领域，也会极大增强大模型的语言理解能力。猎户座则是超越GPT-4的继任模型，但其预训练数据由草莓模型生成。

3.Magnifc Al和 Freepik 联合推出了 Freepik Mystic，宣称是目前最先进的 A图像生成器，也是唯一可以直接生成全高清图像的 A1 图像生成器。

https://t.co/NimQ4DiTyR Meet推出 AI 自动记笔记功能，这项功能与 Meet 的转录工具类似，不过它的工作方式有所不同，它不仅仅是自动记录每个人说的话，而是会对会议内容进行总结。

5.智谱AI宣布GLM-4-flash 大模型免费开放，通过智谱AI开放平台调用。GLM-4-Flash 具备多轮对话网页浏览、Function Call和长文本推理等功能，同时支持包括中文、英语日语、韩语、德语等在内的 26 种语言。

6.阿里智能信息事业群旗下夸克发布全新PC端，全面升级AI搜索、AI写作、AI PPT、AI文件总结等一系列功能。凭借“系统级全场景AI”能力，夸克支持一站式完成信息的检索、创作和总结。
#AI #OpenAl #MagnifcAl #FreepikMystic #GoogleMeet #GLM-4-flash #阿里AI

#Ai Afternoon post /August 29
1.Anthropic announced the opening of Artifacts function for all Claude users. The function is equivalent to creating a dynamic workspace. When users talk with Claude to generate documents, codes, vector drawings, and even simple game Artifacts will appear next to the chat interface. Easy for users to view, select and create your work in real time.

2.OpenAl is developing two models: Strawberry and Oron. Strawberries are mainly used to solve complex reasoning tasks, especially in mathematics and programming, and will also greatly enhance the language understanding of large models. Orion is the successor model to GPT-4, but its pre-training data was generated by the Strawberry model.

3.Magnifc Al and Freepik have jointly launched Freepik Mystic, which claims to be the most advanced A image generator available and the only A1 image generator that can directly generate full HD images.

https://t.co/NimQ4DiTyR Meet introduces an AI note-taking function, which is similar to Meet's transcription tool, but it works in a different way, it doesn't just automatically record what everyone says, but it summarizes the content of the meeting.

5. Zhipu AI announced that GLM-4-flash large model is free and open, which is called through Zhipu AI open platform. GLM-4-Flash features multi-round conversation web browsing, Function Call, and long text inference, and supports 26 languages including Chinese, English, Japanese, Korean, and German.

6. Quark, a unit of Ali Intelligent Information Business Group, released a new PC terminal, comprehensively upgrading a series of functions such as AI search, AI writing, AI PPT, and AI document summary. With "system-level full-scene AI" capabilities, Quark supports the retrieval, creation, and summary of information in one stop.

[Smile//77 @smile77go]

#币圈 大无语事件：
【Noones CEO：Binance已按照以色列国防军的要求扣押所有巴勒斯坦用户的资金】

金融通讯和P2P应用程序Noones首席执行官Ray Youssef表示，Binance已根据以色列国防军的要求扣押了所有巴勒斯坦用户的资金。他们拒绝归还资金，所有上诉均被驳回。根据其从多个渠道获取的消息，所有巴勒斯坦用户都会受到影响，而且从目前的情况来看，所有黎巴嫩用户和叙利亚用户都会受到同样的待遇。

当交易所听从政治摆布，用户的安全感何去何从？谁能保证自己未来不会是牺牲品之一呢？可谓是离谱至极……#BTC #ETH #币圈未来

[Smile//77 @smile77go]

#Ai 早报 /8月27日

1.xAI团队的Grok-2和Grok-Mini模型在LMSys聊天机器人Arena排行榜上脱颖而出，Grok-2以第二名的佳绩超越了OpenAI的GPT-4o，与Gemini并列。

2.Meta Reality 实验室最新推出了名为 Sapiens 的AI视觉模型，专门设计用于理解图片和视频中的人类动作。它支持多种任务，包括2D姿态估计、身体部位分割、深度估计和表面法线预测等。

3.苹果正式进军机器人市场，计划推出搭载生成式AI的桌面设备，预计在2026年或2027年上市。这一举措将为智能家居带来革命性变革，为用户提供更便捷的生活体验。

4.摩尔线程开源了音频理解大模型 -MooER(摩耳)，是业界首个基于国产全功能 GPU 进行训练和推理的大型开源语音模型。基于摩尔线程夸娥(KUAE)智算平台，MooER 大模型用 38 小时完成了 5000 小时音频数据和伪标签的训练。

5.西湖心辰推出的端到端语音大模型Lingo，具备实时打断、指令控制、超级拟人等能力，支持多种语音风格表达，提供高自然流畅度的交互体验。

6.云知声推出山海多模态大模型，该模型实现了多模态输入和输出，提供流畅的语音交互和个性化的视觉体验，支持文本、音频、图像等多种模态输入和任意组合输出，实现高效语音交互。#MetaReality
Ai Morning Post /August 27

1. The xAI team's Grok-2 and Grok-Mini models stand out on the LMSys chatbot Arena leaderboard, with Grok-2 taking second place ahead of OpenAI's GPT-4o and tied with Gemini.

2.Meta Reality Lab recently launched an AI vision model called Sapiens, specifically designed to understand human movements in pictures and videos. It supports a variety of tasks, including 2D pose estimation, body part segmentation, depth estimation, and surface normal prediction.

3. Apple is officially entering the robot market with plans to launch a desktop device equipped with generative AI, which is expected to be available in 2026 or 2027. This initiative will bring revolutionary changes to smart homes and provide users with a more convenient life experience.

4. Moore Thread open source audio understanding large model -MooER(Mol), is the industry's first large open source speech model based on domestic full-featured GPU training and reasoning. Based on the Moorle-threaded KUAE intelligent computing platform, the MooER large model completed 5,000 hours of audio data and pseudo-label training in 38 hours.

5. Lingo, an end-to-end speech large model launched by Xihu Xinchen, has the capabilities of real-time interruption, command control, super personification, etc. It supports a variety of speech styles and provides a high natural fluency interactive experience.

6. Yunzhishun launched a large multi-modal model, which realizes multi-modal input and output, provides smooth voice interaction and personalized visual experience, supports text, audio, image and other modes of input and any combination of output, to achieve efficient voice interaction.

[Smile//77 @smile77go]

#AI产品 #AIGC #AI音乐
昆仑万维最新打造了一个出海的AI音乐产品——Melodio，同时也支持网页版。除此之外昆仑万维还推出了一个AI音乐创作平台——Mureka。Melodio适合听AI音乐，Mureka是AI创作音乐。
Melodio目前已经在各大应用商店上架，只需下载就可以开始使用。它的使用方法极其简洁，打开并登录（或注册）之后，输入你想听的音乐的Prompt，然后点击“Let’s vibe”（感受一下）即可。
Mureka操作上的门槛也是非常低，即使你从来没有创作过音乐也是完全可以hold住的那种：只需两步。第一步，输入歌名歌词，在左侧对应的输入框中，把预先想好的歌名和歌词输入进；第二步，选择参考音乐，可以选择平台自带曲库中的音乐，也可以自行上传本地音乐，在点击“Create”。

[Smile//77 @smile77go]

#AI #OpenAI
Ai早报 /8月22日
1.OpenAI 推出了一项新功能，允许企业客户使用自己的数据来定制其最强大的 AI 模型 GPT-4o。通过微调，现有 AI 模型可以针对特定任务或领域进行优化。

2.Salesforce推出 xGen-MM 开源多模态AI模型，具有处理交错数据的能力，能同时理解和生成文本、图像等多种数据类型。

3.Recogni公司推出了一种名为"Pareto"的革命性AI计算方法，采用创新的对数方法，使AI芯片在运行大型模型时性能超越现有技术，同时大幅降低能耗和运行成本。

https://t.co/j1MTv2M6Yi开放 HeAR AI 模型 API，可以通过分析人的咳嗽和呼吸，诊断出疾病，对各种呼吸系统疾病进行低成本、无障碍的筛查，标志着声学健康研究向前迈出了重要一步。

5.微软最新发布了Phi-3.5-vision，一款轻量级、多模态的开源AI模型，它是Phi-3模型家族的新成员，专为需要同时处理文本和视觉输入的应用设计。

6.字节跳动旗下XR平台PICO推出首款MR混合现实一体机PICO4Ultra，硬件上搭载全新高通骁龙XR2Gen2计算平台，拥有12GB超大内存，GPU性能相比前代XR1提升2.5倍，AI性能比XR1提升8倍。

7.英伟达的研究团队开发出了一款名为 “StormCast” 的 AI 模型，能够在几公里的精度范围内进行雷暴天气的预测。这一技术突破对于气象预测领域具有重要意义，因为以往在这种精细尺度上捕捉大气复杂动态一直都非常具有挑战性。
#AI Ai Morning Post /August 22
1.OpenAI has launched a new feature that allows enterprise customers to use their own data to customize their most powerful AI model, GPT-4o. With fine-tuning, existing AI models can be optimized for specific tasks or domains.

2.Salesforce launched xGen-MM open source multi-modal AI model, which has the ability to process interleaving data, and can simultaneously understand and generate multiple data types such as text and images.

3.Recogni has introduced a revolutionary AI computing method called "Pareto", which uses an innovative logarithmic approach to enable AI chips to outperform existing technologies when running large-scale models, while significantly reducing energy consumption and operating costs.

4. Google opens the HeAR AI model API, which can diagnose diseases by analyzing people's cough and breathing, and carry out low-cost and barrier-free screening for various respiratory diseases, marking an important step forward in acoustic health research.

5. Microsoft recently released Phi-3.5-vision, a lightweight, multimodal open source AI model that is a new member of the Phi-3 model family, designed for applications that need to process both text and visual input.

6. Bytedance's XR platform PICO launched the first MR Mixed reality all-in-one PICO4Ultra, which is equipped with the new Qualcomm Snapdragon XR2Gen2 computing platform on the hardware, with 12GB large memory, GPU performance is 2.5 times higher than the previous generation XR1, AI performance is 8 times higher than XR1.

7. Nvidia's research team has developed an AI model called StormCast that can predict thunderstorms with an accuracy of several kilometers. This technological breakthrough is of great significance for the field of weather prediction, as capturing the complex dynamics of the atmosphere at such fine scales has been challenging in the past.