X
关注 AI 研究者、开发者与机构的动态
按账号或来源筛选(541)
@lifesinger@lifesingerAI 评分1616 @kimmonismus@kimmonismusAI 评分5555 
@ZHO_ZHO_ZHO@ZHO_ZHO_ZHOAI 评分1212 有那味了,笑死哈哈哈哈哈哈哈哈哈哈哈哈哈哈 https://t.co/TgP4uNqd9Z https://t.co/krr8iaXVeE
引用@ZHO_ZHO_ZHO@ZHO_ZHO_ZHO疯狂给朋友们做 OC,势必要把给 GPT 充的钱价值最大化哈哈哈哈哈哈哈哈 这效果都想整个剧集了哈哈哈哈 CharacterDesign|ZHGO|创意系列|GPT Images 2.5 https://t.co/1HXzdOMaIc https://t.co/vqRXNS2LCO
@thexpin@thexpinAI 评分6262 

@alibaba_cloud@alibaba_cloudAI 评分2222 
@testingcatalog@testingcatalogAI 评分3636 引用@testingcatalog@testingcatalogDAILY AI BRIEF 🗞 — Sept 10 APPLE 🔥: - iPhone Duo, first foldable. Starts at $1,999. Ships Oct 23. - iPhone 18 Pro and Pro Max announced. A20 Pro is built for on-device models. - New AirPods 5. New Apple Watch Series 12 and Ultra 4. - iPhone is now an intelligence hub: local models plus cloud. - Siri AI gets immersive voices running on local models. - Watch can capture nearby audio via Live Rewind and Siri Recap. DEEPSEEK 🔥: - V4.1 Flash is out. Smallest model in the new family at 552B params, with native vision. OPENAI 🔥: - ChatGPT Voice limits rose: 3h for Plus, 15h for Pro $100. Pro $200 stays unlimited. - Voice can now use GPT-5.6 Sol and GPT-6 Astra. Astra in voice is Pro-only. META 🔥: - Stilla is joining Meta. Secure multiplayer AI teammate for companies, next to Muse on the consumer side. GOOGLE 🔥: - Five new agentic Workspace features: Chat to slides, Drive to sheets, Docs to email, threads to briefs, proposals to decks. xAI 🔥: - Grok 4.7 is still aimed at ~Sept 12. * Used Grok to compose this brief, cherry-picking the news and doing some post-editing. ** Email version of this brief will be available soon.
@ZHO_ZHO_ZHO@ZHO_ZHO_ZHOAI 评分77 
@elonmusk@elonmuskAI 评分44 @ZHO_ZHO_ZHO@ZHO_ZHO_ZHOAI 评分2525
引用@ZHO_ZHO_ZHO@ZHO_ZHO_ZHO太 tm 酷了!!!我要给 GPT 加钱!!! CatMe 猫与我|ZHGO|创意系列|GPT Images 2.5 https://t.co/2oEW6QCabi
@Thom_Wolf@Thom_WolfAI 评分77 microducks 真是太便宜了 https://t.co/h6a3748LKy

@alibaba_cloud@alibaba_cloudAI 评分2828 @thexpin@thexpin精选AI 评分7171 
推荐理由:Anthropic 指控阿里、Moonshot 与 DeepSeek 未授权用 Claude 训练竞品模型,并给出各家调用量数字,可用于观察模型蒸馏争议的走向。
@foxshuo@foxshuoAI 评分5050 @rohanpaul_ai@rohanpaul_aiAI 评分55 @rohanpaul_ai@rohanpaul_aiAI 评分55 抱歉,您提供的主推文内容仅包含一个链接(https://t.co/OPBgvxE5is),没有可翻译的正文文本。请提供推文的实际文字内容,我将为您翻译。
@rohanpaul_ai@rohanpaul_aiAI 评分4040 
@kimmonismus@kimmonismusAI 评分2626 太疯狂了,以前用 AI 解决黎曼猜想还只是个段子,现在居然真的在被解决了。https://t.co/8RRdkoQ12G
@dongxi_nlp@dongxi_nlpAI 评分1414 @rohanpaul_ai@rohanpaul_aiAI 评分5656 引用@rohanpaul_ai@rohanpaul_aiOpen-weight models may undermine AI lab revenue while continuing to feed cloud demand. OpenRouter’s data suggests proprietary models’ share of routed queries fell from roughly 60% to 25% within months. Its routing system directs each request to the lowest-cost model expected to produce an adequate answer, letting open models gain usage without customers selecting them manually. e.g. AT&T is moving a rapidly growing share of its AI work from proprietary models to cheaper open models, and reports cost savings of up to 80%. If other large companies follow, OpenAI and Anthropic could lose much of the routine enterprise usage they currently charge for. FT published a piece
@rohanpaul_ai@rohanpaul_aiAI 评分6464 OpenRouter 的数据显示,专有模型在路由查询中的占比在数月内从约 60% 降至 25%。其路由系统会把每个请求分配给预计能给出足够答案的最低成本模型,开源模型因此无需用户手动选择就获得用量。
引用@rohanpaul_ai@rohanpaul_aiMark Zuckerberg: Most businesses will not own frontier AI in the way Meta or OpenAI does. But many will end up with something that feels like their own AI: a customized operational layer that reflects how that company actually works. He says, "OpenAI, Google, they're building an AI. But I think we're gonna have a lot of different AI systems, just like we're gonna have, we have a lot of different apps. I think in the future, every business, just like I have a website and a phone number and an email address, a social media account, is also going to have an AI that can interact with their customers to help them sell things, help them give support." --- What he is really describing that a company’s “own AI” will usually not be a frontier model trained from scratch, but a layer built on top of shared models, shaped by its products, policies, customer history, and way of working. Support, sales, and basic operations can be handled through a system that knows the business well enough to answer, route, recommend, and escalate without feeling generic. --- From 'Cleo Abram' YT channel (link in comment)
@rohanpaul_ai@rohanpaul_aiAI 评分44 @ZHO_ZHO_ZHO@ZHO_ZHO_ZHOAI 评分55 
@elonmusk@elonmuskAI 评分44 @elonmusk@elonmuskAI 评分1313 @rohanpaul_ai@rohanpaul_aiAI 评分55 @rohanpaul_ai@rohanpaul_aiAI 评分3838 
@foxshuo@foxshuoAI 评分55 太压抑啦…… https://t.co/dTKwH6S440

@rohanpaul_ai@rohanpaul_aiAI 评分5050 
@rohanpaul_ai@rohanpaul_aiAI 评分5858 斯坦福与 Together AI 发布论文 Intelligence per Watt,提出以单位功耗下的任务准确率衡量本地 AI 的能效。

@thsottiaux@thsottiauxAI 评分3636 下周我们将退役 GPT-5.3-Codex-Spark。你能相信我们居然发布了这么个名字的模型吗!! 它表现不错,也很有趣,但使用量一直在下降,而且我们现在有了明显更好的模型。是时候为未来腾出空间了。
@rohanpaul_ai@rohanpaul_aiAI 评分66 @rohanpaul_ai@rohanpaul_aiAI 评分6262 
@elonmusk@elonmuskAI 评分55 @ZHO_ZHO_ZHO@ZHO_ZHO_ZHOAI 评分1313 太 tm 酷了!!!我要给 GPT 加钱!!! CatMe 猫与我|ZHGO|创意系列|GPT Images 2.5 https://t.co/2oEW6QCabi

@bcherny@bchernyAI 评分2828 



@rohanpaul_ai@rohanpaul_aiAI 评分55 抱歉,您提供的主推文内容只有一个链接(https://t.co/3wmuI7qfCe),没有实际的推文文字可供翻译。请提供推文的完整文字内容,我将为您翻译并生成标题。
@rohanpaul_ai@rohanpaul_ai精选AI 评分7171 
推荐理由:Bloomberg 报道给出 OpenAI 暂停 RL 训练与监测算力的具体数字,可供判断安全测试对前沿研发节奏的影响。
@emollick@emollickAI 评分3030 
@lifesinger@lifesingerAI 评分99 @lifesinger@lifesingerAI 评分2828