VOL.2026.08.03 · 30 篇报道 · AI 日报
AI 日报 — 2026-08-03
星期一 · 30 篇报道 · 约 11 分钟读完
今天的AI领域,模型定价竞争日益激烈,阿里巴巴的Qwen3.8-Max以极具竞争力的价格挑战Kimi K3,而DeepSeek的V4-Flash则提供了更低的成本。这种价格压力,加上Qwen3.8-Max即将开源模型权重,预示着AI开发将走向更低成本和更易获取的时代。与此同时,欧盟AI法案的执法权已生效,带来了重要的监管审查和透明度要求,这将深刻影响这些日益强大且经济实惠的AI模型在欧洲的开发和部署。
- 01模型与开源阿里巴巴的Qwen3.8-Max定价远低于Kimi K3等竞争对手,输入每百万Token 2美元,输出每百万Token 6美元,预示着潜在的价格战,可能使高级大型语言模型对开发者而言更易获取且经济实惠。10
- 02Agent 与工具多个AI代理项目正在涌现,包括用于本地渗透测试的Nightcrawler和用于云端编码代理的Hoplite,这表明专业化、可部署的AI工具正日益普及,旨在自动化复杂任务。7
- 03应用落地OpenAI的ChatGPT占据了美国众议院办公室所有AI工具支出的约90%,凸显了其作为AI实际应用在政府领域的主导地位。1
- 04融资&商业据报道,OpenAI的超级政治行动委员会正在资助一个由AI生成新闻的网站The Wire by Acutus,以攻击行业批评者,引发了人们对AI在政治话语和企业影响力中伦理影响的担忧。4
- 05政策&风险欧盟AI法案的执法权已生效,允许欧盟在区域发布前评估和限制AI模型并处以罚款,为AI的开发和部署建立了强大的监管框架。2
- 06行业动态Visa计划以24亿美元收购BioCatch,凸显了AI在实时欺诈检测和安全领域的日益增长的价值,表明对关键金融基础设施中AI驱动解决方案的重大投资。6
01模型与开源10 篇
- LLMs reward expertise日榜第 3 名0 个来源热度 41
- Prevent cognitive debt by manually retyping LLM-generated code日榜第 4 名0 个来源热度 40
- How we built a realtime system for responsive voice AI in six months
OpenAI developed GPT-Live, a third-generation voice system, to enhance real-time voice AI responsiveness. Unlike previous turn-based systems that used turn detectors and often caused delays or interruptions, GPT-Live employs a full-duplex voice model, allowing it to listen and speak simultaneously. This eliminates the need for a separate detector, making conversations more natural and immediate. It can also consult frontier models like GPT-5.5 for deeper reasoning without disrupting the conversational flow, achieving an unprecedented combination of responsiveness and intelligence.
日榜第 22 名0 个来源热度 24 - DeepSeek-V4-Flash-0731: surpasses Fable-5, Sol & Kimi-K3 on Chess Benchmark
TensorSharp now supports DSpark on Deepseek v4 Flash 0731, demonstrating significant performance improvements. Benchmarks on 4x Nvidia A40 GPUs with CUDA 12.8 show DSpark enhancing generation speeds across various tasks. For instance, short generation (53 tokens) saw a 1.74x speedup, long generation (512 tokens) a 1.53x speedup, and 10K-token document processing a 2.03x speedup. TensorSharp is an open-source inference engine for GGUF LLMs, featuring CUDA, Vulkan, Metal, and OpenAI-compatible APIs.
日榜第 29 名0 个来源热度 23
02Agent 与工具7 篇
- Influencers draw backlash for attending OpenAI’s first luxury trip
OpenAI's inaugural influencer brand trip has ignited considerable online backlash amidst ongoing debates surrounding the ethical implications and widespread adoption of AI technologies. The luxury nature of the trip, coupled with the current societal tensions regarding AI, has fueled public criticism, highlighting the growing scrutiny faced by companies at the forefront of artificial intelligence development.
日榜第 15 名0 个来源热度 27
03应用落地1 篇
- Congress’ favorite AI tool? ChatGPT
During the year ending March 31, OpenAI's ChatGPT accounted for approximately 90% of all AI tool spending by House offices, committees, and institutional accounts. House disbursement records indicate that Congress spent around $100,580 on ChatGPT across 798 transactions, out of at least $113,740 in total AI spending. Anthropic’s Claude was the second most used tool, with $13,160 spent across 37 transactions.
日榜第 23 名0 个来源热度 23
04融资&商业4 篇
- OpenAI's super PAC is funding AI-generated news site attacking industry critics
OpenAI's super PAC is reportedly funding an AI-generated news site, The Wire by Acutus, which attacks critics of the AI industry. The site's reporting, exemplified by a piece attacking pharmacy benefit managers, shows overlap with PR work for Novus, a company that lists PhRMA as a client. This connection is highlighted by Acutus's internal source log referencing "PhRMA statements" in an article, despite no PhRMA figure being quoted, while PhRMA simultaneously lobbied for pharmacy benefit manager reform.
日榜第 6 名0 个来源热度 35 - Design Arena creators raise $7.9 million to bring taste to AI models
Design Arena, founded by Grace Li and college friends, has raised $7.9 million to integrate "taste" into AI models. The company originated from a challenge to make AI-generated games enjoyable, as initial models could create functional but not fun games. This funding aims to address the critical question of how to imbue AI with the ability to discern and create engaging experiences.
日榜第 19 名0 个来源热度 26
05政策&风险2 篇
- Europe’s AI labeling and transparency rules are now in effect
The European Union's new AI Act transparency obligations, effective August 2nd, require companies to disclose when users interact with AI models or when content is AI-generated or altered. These rules aim to help people identify chatbots and AI deepfakes online. Non-compliant companies face fines up to €15 million or 3% of global annual turnover. AI systems launched before August 2nd have a four-month grace period until December 2nd to comply.
日榜第 11 名0 个来源热度 27 - The EU's AI Act enforcement powers take effect, letting it evaluate AI models before regional release, restrict market access, fine model providers, and more (Kai Nicol-Schwarz/CNBC)
The EU's AI Act enforcement powers have taken effect, granting the European Union new authority to evaluate AI models before their regional release. This allows the EU to restrict market access for certain models and impose fines on model providers. These new powers aim to regulate AI technology within the EU.
日榜第 16 名0 个来源热度 27
06行业动态6 篇
- Ten advances in mathematics and theoretical computer science日榜第 1 名0 个来源热度 67
- YouTuber Hank Green says his AI usage is ‘not healthy’日榜第 9 名0 个来源热度 33