Performance: dev-browser: 3m53s, $0.88, 100% success rate — beats MCP configs, Chrome extensions, 'browser skill' stacks.
令人惊讶的是：这种新技术不仅在功能上超越传统方法，在性能指标上也取得了显著优势，100%的成功率和相对较低的成本显示了其技术成熟度和实用性，这可能会使现有的浏览器自动化解决方案迅速过时。
11 Matching Annotations
- Last 7 days
-
-
-
www.xiaohu.ai www.xiaohu.ai
-
60 秒四路数据源并行采集，输出图文交错的研报。
令人惊讶的是，GLM-5V-Turbo集成的'股票分析师'Skill能在短短60秒内从四个不同数据源并行采集信息并生成图文交错的研报。这种速度和效率远超传统金融分析师，展示了AI在专业领域的惊人潜力。
Tags
Annotators
URL
-
-
www.technologyreview.com www.technologyreview.com
-
Where training a language model took 167 minutes on eight GPUs in 2020, it now takes under four minutes on equivalent modern hardware.
令人惊讶的是：AI训练效率的提升速度令人震惊。在短短6年内，语言模型的训练时间从167分钟缩短到不到4分钟，效率提升了40多倍。这种进步远超摩尔定律预测的5倍改进，展示了AI硬件和算法的飞速发展。
-
-
news.smol.ai news.smol.ai
-
Meta says its rebuilt pretraining stack can reach equivalent capability with >10× less compute than Llama 4 Maverick
令人惊讶的是，Meta声称他们重建的预训练栈只需要Llama 4 Maverick十分之一的计算量就能达到同等能力。这一效率提升是惊人的，表明AI模型训练可能正在经历一个范式转变，从单纯增加计算资源转向优化算法和架构。这可能会对整个AI行业的成本结构和竞争格局产生深远影响。
-
-
-
Brandon told the team on a Monday that OKRs were due Wednesday—a turnaround that would have been absurd without this agent.
令人惊讶的是：借助AI代理，原本需要数周才能完成的OKR规划流程可以在两天内完成，效率提升惊人。这展示了AI如何彻底改变传统企业规划流程，从冗长的手动过程转变为快速、智能的自动化系统。
-
-
-
we can reach the same capabilities with over an order of magnitude less compute than our previous model, Llama 4 Maverick.
令人惊讶的是：Meta声称他们的新模型Muse Spark在计算效率上取得了突破性进展，仅用前代模型Llama 4 Maverick十分之一的计算量就能达到相同能力。这种数量级的效率提升在AI领域极为罕见，可能代表着训练算法和架构设计的重大革新。
-
-
www.tomtunguz.com www.tomtunguz.com
-
Closed Loop + Finite Demand = Efficiency Plays. AI bookkeeping categorizes transactions, reconciles accounts, files returns. Deterministic rules applied to numbers.
令人惊讶的是：即使是有限需求领域，AI也能通过确定性规则实现显著效率提升。AI记账系统能够自动处理分类、对账和报税等任务，这表明即使在传统上需要人工判断的财务领域，AI也能通过标准化流程创造价值。
-
- Apr 2026
-
deepmind.google deepmind.google
-
E2B & E4B · A new level of intelligence for mobile and IoT devices
「手机和 IoT 设备的新智能层级」——这个定位本身就是宣战书。E2B 有效参数仅 2.3B，却能在不足 1.5GB 内存中运行，并支持 128K 上下文窗口。令人震惊的是，E4B 在多项指标上超越了 Gemma 3 27B——一个 4.5B 的边缘模型击败了 27B 的上一代旗舰。参数效率的边界正在被彻底重写。
-
-
www.tomtunguz.com www.tomtunguz.com
-
In 23 months, the same capability that needed 1.8 trillion parameters now fits in 4 billion parameters. A 450x compression.
大多数人认为AI模型性能提升主要依靠参数数量增加，但作者认为通过算法优化和人才聚集，AI模型可以实现450倍的参数压缩，这挑战了'更大参数等于更好性能'的行业共识。
-
-
-
By using SAM, the Alta team has been able to process more than 20 million images without incurring exorbitant costs, allowing them to focus on building the best possible product for their users.
大多数人可能认为初创公司需要依赖昂贵的第三方API来处理大量图像，但作者通过使用开源SAM模型，实现了大规模图像处理而不产生巨额成本。这一观点挑战了'高质量AI服务必须昂贵'的行业共识，展示了开源模型在成本效益方面的优势。
-
- Jun 2023
-
www.nngroup.com www.nngroup.com
-
the productivity and quality improvements are likely due to a switch in the business professionals’ time allocation: less time spent on cranking out initial draft text and more time spent polishing the final result.
This points to AI providing the best time savings in draft generation, which fits with the idea of having the AI generate the drafts based on the professional's queries.
For UX designers, this points to AI in a design tool being most useful when it generates drafts (sketches) that the designer then revises. Where UX deliverables don't compare easily to written deliverables is the contextual factors that influence the design, like style guides or design systems. Design too AI assistants don't yet factor those in, though it seems likely it will, if provided style guides and design systems in a format it can read.
Given a draft of sufficient quality that it doesn't require longer to revise than a draft the designer would create on their own, getting additional time to refine sounds great.
I'm not sure what to make of the reduced time to brainstorm when using AI. Without additional information, it's hard not to assume that the AI tool may be influencing the direction of brainstorming as professionals think through the queries they'll use to get the AI to generate the most useful draft possible.
-