888 X 動態摘要｜07/30
生成時間: 2026-07-30 05:04:33
總結
xAI發布Grok Voice Think Fast 2.0，展現領先的語音AI技術與商業化潛力；同時，AI巨頭對產業安全與發展節奏的擔憂也日益浮現。
今日重點
1. xAI語音模型性能與定價
- 人物: Elon Musk (xAI 創始人)
- 時間: 2h
- 熱度: 👀 303,768
- 觀察: Grok Voice Think Fast 2.0在語音到語音指數中排名第二，並以0.70秒的首次音訊響應時間領先所有競品，定價策略顯示市場競爭力。
- 意義: 技術突破強化了xAI在即時語音AI市場的競爭力，其定價策略也揭示了高階語音服務的商業化潛力與獲利空間。
2. AI安全風險與發展步伐
- 人物: Yann LeCun (Meta AI掌門)
- 時間: Jul 28
- 熱度: 👀 606,476
- 觀察: Meta AI掌門Sam表示因Hugging Face安全事件，已暫停訓練並需強化沙盒安全，考慮放緩AI發展速度。
- 意義: 揭示了AI前沿發展面臨的嚴重資安挑戰，並引發對行業監管與協調開發節奏的深層討論，可能影響未來資本投入與法規走向。
原始動態
重點 | Elon Musk
- 時間: 2h
- 熱度: 👀 303,768
- 原文: "SpaceXAI has released Grok Voice Think Fast 2.0 today, with the High reasoning variant debuting at #2 on the Artificial Analysis Speech to Speech Index at 82.9%, and #1 on Tau Voice for Agentic Performance at 56.5% - among the fastest models at 0.70s Time to First Audio Grok Voice Think Fast 2.0 is SpaceXAI's successor to Grok Voice Think Fast 1.0 (75.7% on the Speech to Speech Index). It is the only model in the Index's top five with an average Time to First Audio under 1 second, achieving 0.70 seconds vs. 1.14 seconds for the next fastest, GPT-Realtime-2 High. Key takeaways: ➤ Speech to Speech Index: Grok Voice Think Fast 2.0 High debuts at #2 at 82.9%, only behind Qwen Audio 3.0 Realtime Plus (84.1%) and ahead of GPT-Realtime-2.1 High (79.1%) and GPT-Realtime-2 High (77.2%). This is up 7.3 percentage points from Grok Voice Think Fast 1.0 (75.7%) ➤ Speech to Speech Index by Benchmark: On Tau Voice, Grok Voice Think Fast 2.0 High is the new leader at 56.5%, just ahead of Qwen Audio 3.0 Realtime Plus at 54.6% and its predecessor Grok Voice Think Fast 1.0 at 52.1%. On Big Bench Audio, it achieves 97.2%, behind leader Qwen Audio 3.0 Realtime Plus at 99.2%. On our Full Duplex Bench subset, it scores 95.1%, up from 77.8% for Grok Voice Think Fast 1.0, the largest driver of its index gain, behind Qwen Audio 3.0 Realtime Plus at 98.4% ➤ Speed: The model’s average Time to First Audio on Big Bench Audio is 0.70 seconds, faster than GPT-Realtime-2 High (1.14s), GPT-Realtime-2.1 High (1.21s), and Grok Voice Think Fast 1.0 (1.25s), and well ahead of Qwen Audio 3.0 Realtime Plus (4.02s) ➤ Price: Grok Voice Think Fast 2.0 is priced at $4.80 per hour of input audio, up from $3.00 for Grok Voice Think Fast 1.0 and more expensive than Qwen Audio 3.0 Realtime Plus ($4.42) and GPT-Realtime-2 High ($4.14), but ~2.2x cheaper than GPT-Realtime-2.1 High ($10.75) Congratulations" "! See below for more detail ⬇️"
重點 | Yann LeCun
- 時間: Jul 28
- 熱度: 👀 606,476
- 原文: "Sam on the Hugging Face incident: “This is the first security incident that I have felt very viscerally. I've been a little surprised that more people don't feel it so viscerally. We paused training. We have to figure out how to secure our sandboxing in a world of multiple zero days being chained together. We may have to pace the rate of AI development to give ourselves enough time for society to harden around these new capability levels. And trying to figure out how we do that in a way that does not feel like regulatory capture for anyone and also does not feel like collusion among the frontier labs.”"
其他 | Elon Musk
- 時間: 1h
- 熱度: 👀 202,092
- 原文: We have just released our latest flagship voice model, Grok Voice Think Fast 2.0. We are eager to hear feedback and think you will be happy with what you hear :) A TON of effort, thought, and love has gone into this by everyone on the voice team. The model’s intelligence, accuracy, and capabilities almost feel otherworldly. The voice space still comes with so much friction, Think Fast 2.0 is a significant step closer to something that just works. As voice models should!
其他 | Greg Brockman
- 時間: 3h
- 熱度: 👀 38,693
- 原文: Science moves faster when more researchers have access to the best tools. Getting frontier AI into the hands of academic researchers means more shots on goal against humanity’s hardest problems — and more breakthroughs that benefit all of us. Excited to see what they discover.
其他 | Ray Dalio
- 時間: 3h
- 熱度: 👀 37,378
- 原文: While this is important throughout the organization, it is especially important that the reporting lines of the board (those doing the oversight) are independent of the reporting lines of the CEOs (those doing the management), though there should be cooperation between them.
其他 | Elon Musk
- 時間: 1h
- 熱度: 👀 833,491
- 原文: Try the new Grok Voice
其他 | Huanusa
- 時間: 10h
- 熱度: 👀 86,511
- 原文: 亿万富翁一句话道出了穷人为什么不能翻身 究其原因可能是被洗的太彻底了
其他 | Nassim Taleb
- 時間: 6h
- 熱度: 👀 48,452
- 原文: "'Highly disgraceful': Lebanese lawyers file lawsuit against banker Sehnaoui after Netanyahu meeting"
其他 | Nassim Taleb
- 時間: 1h
- 熱度: 👀 21,358
- 原文: "To a Zio shill trying to insult me: \"I hate to break it to you: you are so insignificant that your praise would be worth no more to me than your insults.\"
其他 | Ray Dalio
- 時間: Jul 27
- 熱度: 👀 97,832
- 原文: If you can do that, you will take the power of your decision making to a whole other level. In many cases, you will be able to test how that principle would have worked in the past or in various situations that will help you refine it, and in all cases, it will allow you to compound your understanding to a degree that would otherwise be impossible. It will also take emotion out of the equation. Algorithms work just like words in describing what you would like to have done, but they are written in a language that the computer can understand. If you don’t know how to speak this language, you should either learn it or have someone close to you who can translate for you. Your children and their peers must learn to speak this language because it will soon be as important or more important than any other language.
其他 | Ray Dalio
- 時間: Jul 28
- 熱度: 👀 86,488
- 原文: Most importantly, you need to have principles that deal with reality. Embrace reality, deal with it, and have those core principles. If everyone took the time to write down their principles and think about how to deal with specific situations, the world would operate a lot more effectively. I’ve been through those experiences and I wrote down my recipes. You can consider my recipes, or anyone else’s recipes, but at the end of the day, you must believe them and do what is good for you. Because if you don’t have principles, everything looks the same. I encourage you to have your own principles that you understand and believe in, wherever you get them.
其他 | Nassim Taleb
- 時間: 2h
- 熱度: 👀 7,424
- 原文: "The methodology adapting the properties of sawtooth waves:"
其他 | Huanusa
- 時間: 7h
- 熱度: 👀 5,077
- 原文: 太多这种案例都石沉大海-别说普通人，于朦胧这种明星不也是一样吗？ 可不可怕...
其他 | Huanusa
- 時間: 5h
- 熱度: 👀 1,120
- 原文: 这真是非常牛的链上数据账号👍 绝对值得关注，数据已经达到付费的程度
其他 | Peter Steinberger
- 時間: 15h
- 熱度: 👀 194,384
- 原文: Serving large models is hard.
其他 | Greg Brockman
- 時間: Jul 28
- 熱度: 👀 341,983
- 原文: "My friend Zhengqing used GPT-5.6 to solve a beautiful well-known conjecture in probability. Feige's 1/e conjecture: for independent nonnegative X_1,…,X_n with E[X_i] ≤ 1 and S_n = X_1+⋯+X_n, we have P(S_n ≤ E[S_n] + 1) ≥ 1/e. Exciting to see a problem we used to kick around over lunch and dinner get solved! 1/2
其他 | Greg Brockman
- 時間: Jul 28
- 熱度: 👀 221,279
- 原文: I love ChatGPT Work man, just did this so easily
其他 | Yann LeCun
- 時間: Jul 28
- 熱度: 👀 88,397
- 原文: Zuckerberg throws a big punch at Dario’s fear marketing and moral high grounds. never thought I’d say this, but Zuck’s article is pretty based. 👏🏻👏🏻👏🏻
其他 | Greg Brockman
- 時間: Jul 28
- 熱度: 👀 83,834
- 原文: Don’t sleep on ChatGPT Work. Use it from mobile. Incredible productivity boost.
其他 | Greg Brockman
- 時間: 14h
- 熱度: 👀 71,369
- 原文: yep
其他 | Eric Topol
- 時間: 9h
- 熱度: 👀 5,979
- 原文: That’s ", a brilliant AI scientist, who is worth following
其他 | Huanusa
- 時間: 5h
- 熱度: 👀 315
- 原文: 喜欢先知币的关注起来，最近数据还是活跃了
其他 | Pierre Levy
- 時間: 27m
- 熱度: 👀 2,253
- 原文: "Because a lot of people are genuinely confused: no model can currently actually \"escape\", in the sense of accessing its weights, finding some remote data center and copying itself to spin up a rogue instance. The terminology is highly misleading here."
其他 | Pierre Levy
- 時間: 1h
- 熱度: 👀 2,042
- 原文: Fable is great but often almost unusable, because even the most harmless questions about topics like, say, shamanism or philosophy of mind are triggering the safeguards. I hope that this gets fixed soon. Fable is more like a tech demo at this point.
其他 | Sean Kelly
- 時間: Jul 27
- 熱度: 👀 86,071
- 原文: "\"i nominate you, you nominate me\" is standard academia. wait until you find out that some universities have actually tried to institutionalize what's basically a \"find someone to nominate me, and I will nominate someone in return\" service^ fun times
其他 | Sean Kelly
- 時間: Jul 28
- 熱度: 👀 21,450
- 原文: this move has taken its toll