早在 1980 年代初期，当 8-bit PC 硬件受到极其严重的限制时，制作流畅、可玩的电子游戏是一项极其耗费精力的优化工作。作者的一位朋友当时正在开发一款图形密集的横向卷轴游戏，游戏几乎已经完成，但却深受帧率过低而无法运行的问题困扰。尽管程序员已经应用了诸如循环展开（loop unrolling）等标准优化手段，但游戏速度依然缓慢。当时的挑战在于硬件缺乏诸如 blitters 或 texture mappers 等现代特性，这要求程序员必须手动逐字节地移动数据以将图像绘制到屏幕上。

作者在研究代码后意识到，瓶颈在于负责将 background tiles 复制到 screen buffer 的那段程序。机器使用的 6809 处理器提供了一些特定且强大的寻址模式，但朴素的复制例程的周期计数依旧很高。通过进一步展开循环，并手动计算内存访问的偏移量，他们实现了 60% 的速度提升。然而，这仍不足以满足游戏的性能要求，于是作者开始寻求更非传统、近乎"疯狂"的解决方案。

在查阅 6809 微处理器手册后，作者发现处理器的 stack 指令可以被改造成用于快速数据传输。通过征用硬件 stack 寄存器，即 S 和 U，作者开发出一种方法：利用包括 DP 寄存器在内的所有可用寄存器，以单次突发（single burst）的方式把数据从 tile 中取出并推送到 screen buffer 。这种方案效率极高，但也带来了副作用——tiles 会被绘制得上下颠倒，且左右两半被交换。

为了解决这一问题，两人意识到这种"搅乱"过程其实是一个 involution 。于是他们在游戏运行前就对 tiles 做了预处理，提前把内部结构反转。这样一来，等到把被破坏的数据绘制到屏幕后，图像在到达屏幕时会被自动完美地还原。该方案带来了 70% 的速度提升，基本有效地解决了性能问题。但当他们为声音加入 hardware-clocked interrupts 时，最后一道难题出现了。由于该例程需要征用 stack pointer，如果在绘制 tile 的过程中触发中断，就会覆盖保存 reference tiles 的内存。

他们通过巧妙地交换 stack 指针来化解这次最终的技术危机，确保中断造成的任何破坏都会落到 screen buffer 上，而不是原始的 tile data 。为了避免这种短暂的屏幕异常被玩家看到，他们调整了绘制顺序：从高内存往低内存绘制。这样，任何意外的损坏都会在下一张 tile 被绘制时被覆盖，从而保证在 buffer 送往显示器之前问题不会显现出来。

回想当初，作者把这段经历视作早期计算时代所必需的创造力的证明。面对极其苛刻的硬件限制，程序员不得不反复尝试、失败并迭代，才能找到既巧妙又有时不那么正统的解决方案。这个故事也提醒人们：在这样的技术挑战中，同样能收获乐趣，并鼓励他人把自己的经验记录下来，免得这些具有专业性与历史意义的见解随着时间流逝而被遗忘。 Back in the early 1980s, when 8-bit PC hardware was incredibly limited, creating smooth, playable video games was a grueling exercise in optimization. A friend of the author was developing a graphically intensive side-scroller that was nearly complete but suffered from unplayable frame rates. Despite the programmer having already applied standard optimizations like loop unrolling, the game remained too slow. The challenge was that the hardware lacked modern features like blitters or texture mappers, requiring programmers to manually move data byte by byte to draw images onto the screen.

The author, after studying the code, realized that the bottleneck was the routine responsible for copying background tiles to the screen buffer. The 6809 processor used in the machine offered specific, powerful addressing modes, but the cycle counts for the naive copying routine remained high. By further unrolling the loops and manually calculating offsets for memory access, they achieved a 60 percent speed increase. However, this was still insufficient to meet the performance requirements of the game, leading the author to seek even more unconventional, "crazy" solutions.

After consulting a 6809 microprocessor manual, the author discovered that the processor's stack instructions could be repurposed for rapid data movement. By commandeering both the hardware stack registers, S and U, the author developed a method to pull data from a tile and push it to the screen buffer using every available register—including the DP register—to move blocks of data in a single burst. This approach was remarkably efficient, but it introduced a side effect: the process drew the tiles upside down and with their halves swapped.

To overcome this, the pair realized that the mangling process was actually an involution. By pre-processing the tiles to reverse their internal structure before the game ran, the act of drawing the corrupted data would perfectly restore the images once they reached the screen. This solution provided a 70 percent speed improvement, effectively solving the performance issues. However, a final hurdle emerged when they implemented hardware-clocked interrupts for sound. Because the routine relied on commandeering the stack pointer, an interrupt occurring during a tile draw would overwrite the memory where the reference tiles were stored.

They addressed this final technical crisis by cleverly swapping the stack pointers, ensuring that any corruption caused by an interrupt would target the screen buffer rather than the original tile data. To prevent this temporary screen glitch from being visible to the player, they altered the drawing order to proceed from high memory to low memory. This ensured that any accidental corruption would be overwritten by the next tile being drawn before the buffer was sent to the display.

Looking back, the author highlights this experience as a testament to the ingenuity required during the early era of computing. Faced with severe hardware limitations, programmers had to constantly experiment, fail, and iterate to find elegant, if sometimes unorthodox, solutions. The story serves as a reminder of the joy found in such technical struggles and an encouragement for others to document their own experiences before these specialized, historic insights are lost to time.

在 Anthropic 的 Project Glasswing 启动后的七个月里，这项计划在软件安全领域真正带来的影响仍遭到质疑。尽管项目一开始就宣称能够识别数千个高严重性与关键严重性漏洞，但实际效果远没有达到预期。追踪数据显示，归因于 Anthropic 的漏洞仅占今年已发布 CVE 总量的 0.5%，而其中真正可能在实际环境中被利用的证据少之又少。

项目内部的"账本"难以继续推进，暴露出发现与修复之间的差距越来越大。自 9 月以来，Anthropic 报告的发现数量几乎增加了 12,000 项，但已确认的修复数量却基本停滞不前。此外，被标记为重复或被撤回的发现数量，明显超过了实际完成的修复数量，这也让人对项目所声称的高真阳性率产生了疑虑。

鉴于人工分诊无法有效扩展，Anthropic 近期调整了策略，推出了自动化的 OSS Scanner 。该工具把验证压力从 Anthropic 内部团队转移给开源维护者：维护者现在需要处理未经核实、由模型生成的报告。 Anthropic 也承认这些发现是在没有人工审查的情况下生成的，这引发了对低质量甚至错误提交风险的担忧。

对这套新扫描器进行的一次实际测试以 curl 项目为对象，进一步凸显了维护者的困境。在 curl 处理的 34 份报告中，超过一半被判定为无效，只有少数被确认属于低严重性问题。这种结果与 Anthropic 所宣称的高成功率形成了鲜明反差，也表明项目产出的严重性或准确性往往达不到最初承诺。

归根结底，尽管 AI 驱动的漏洞发现具备潜力，但 Project Glasswing 的相关数据反映出目前的落地效果未能兑现其宣传叙事。让开源维护者承担大量潜在误报的处理工作，不仅没有减轻负担，反而给原本希望帮助的社区增加了额外工作。随着项目主张与维护者所观察到的现实之间的偏差持续存在，这项倡议的整体有效性与必要性仍将受到持续质疑。 Seven months into Anthropic's Project Glasswing, the initiative continues to face skepticism regarding its actual impact on software security. While the project launched with ambitious claims about identifying thousands of high- and critical-severity vulnerabilities, the reality has proven to be less significant. Tracking data indicates that vulnerabilities attributable to Anthropic account for a mere 0.5% of all CVEs published this year, with an extremely small fraction showing evidence of being exploited in the wild.

The project's internal ledger has struggled to maintain momentum, revealing a growing disconnect between discovery and resolution. Since September, the number of findings reported by Anthropic has surged by nearly 12,000, yet the number of confirmed fixes has remained largely stagnant. Furthermore, the volume of findings marked as duplicates or withdrawn has significantly outpaced the number of actual fixes, casting doubt on the project's reported high true-positive rates.

Recognizing that manual triage was not scaling, Anthropic recently shifted its approach by launching an automated OSS Scanner. This tool effectively offloads the burden of validation from Anthropic's internal teams to open source maintainers, who must now process unverified, model-generated reports. Anthropic acknowledges that these findings are produced without human review, raising concerns about the potential for low-quality or incorrect submissions.

A practical test of this new scanner, involving the curl project, highlighted the challenges maintainers face. Out of 34 reports processed by curl, more than half were determined to be invalid, while only a small minority were confirmed as low-severity issues. This experience stands in stark contrast to the high success rates claimed by Anthropic and suggests that the project's output often lacks the severity or accuracy initially promised.

Ultimately, while AI-driven vulnerability discovery holds promise as a tool, the data from Project Glasswing suggests that the current execution is falling short of its marketing narrative. The reliance on open source maintainers to handle a high volume of potential false positives has created an additional workload for the very communities the project aimed to assist. As the discrepancies between the project's claims and the reality observed by maintainers persist, the overall effectiveness and necessity of the initiative remain subject to ongoing critique.

Apple 正在开发一款名为 Harbor 的新应用，旨在简化在 Mac 硬件上本地运行人工智能模型的流程。通过让用户把电脑变成私有 AI 推理服务器，该工具确保数据始终留在用户设备上，而不会被传输到云端。此举也显示出 Apple 对支持基于 Silicon 架构设备上的本地 AI 工作负载的更深承诺，包括 Mac Studio 和 Mac mini 。

Harbor 最引人注目的功能之一，是它能通过 Thunderbolt 5 将多台 Mac 电脑连接起来，从而组建一个集群。借助这种连接方式，应用可以在不同设备之间汇聚统一内存，使用户能够运行通常会超出单台机器容量的更大、更复杂的模型。同时，这种集群配置还能提升并发能力，以便同时处理多个请求。

Harbor 的用户体验设计为简洁、上手容易，几乎不需要专业的技术能力。软件会自动检测通过 Thunderbolt 5 连接的设备以组建集群，无需复杂的手动设置或命令行操作。此外，应用内置模型管理功能，用户可以直接在界面中浏览、下载并加载开源权重模型。

目前，Harbor 仍处于 Apple 旗下 Claris 品牌的私有预览阶段。虽然它还未对公众开放，但其开发进展表明，未来很可能会推出正式版本，让更多 Mac 用户更容易获得强大的私有 AI 计算能力。 Apple is developing a new application called Harbor, designed to simplify the process of running artificial intelligence models locally on Mac hardware. By enabling users to turn their computers into private AI inference servers, the tool ensures that data remains on the user's device rather than being transmitted to the cloud. This initiative signals a deeper commitment from Apple to support local AI workloads on silicon-based machines, including the Mac Studio and Mac mini.

One of the most notable features of Harbor is its ability to create a cluster by connecting multiple Mac computers over Thunderbolt 5. This connectivity allows the application to pool unified memory across devices, which effectively enables users to run larger, more complex models that would typically exceed the capacity of a single machine. Additionally, this cluster setup supports higher concurrency for handling multiple requests simultaneously.

The user experience in Harbor is intended to be streamlined and accessible, requiring minimal technical expertise. The software automatically detects Thunderbolt 5-connected devices to form a cluster without the need for complex manual configurations or command-line inputs. Furthermore, the app includes integrated model management, allowing users to browse, download, and load open-weight models directly within the interface.

Currently, Harbor is in a private preview phase under Apple's Claris brand. While it is not yet available to the general public, its development suggests a forthcoming release that will likely make powerful, private AI computing more attainable for a wider range of Mac users in the near future.

在 Terence Tao 的著作《 Six Math Essentials 》中，这位 Fields Medalist 将浩瀚的数学领域提炼为六个基本概念，它们自古至今一直是思想者们的指路明灯。通过聚焦这些核心支柱，本书为理解定量推理如何帮助我们驾驭并解读当代世界的复杂性提供了一份清晰路线图。

这六个要素中的第一个，是"数"的概念。 Tao 将其视为通往定量思维的主要入口。随后，他转向代数，把它描述为实现抽象化、从而概括各种关系的关键工具。几何并不仅仅被当作研究形状的学科来引入，而是一种用于计算与测量那些超出人类视觉边界的现象的实用方法。

为应对生活中固有的不可预测性，本书引入概率，作为一种用严谨、合乎逻辑的思维来处理不确定性的框架。与此同时，分析学提供了驾驭极端情形所必需的数学语言，无论面对的是极其宏大的尺度，还是无穷小的存在。最后，本书以动力学收束，Tao 将其呈现为理解"变化本身"的决定性数学。

这本书篇幅短小却精炼优雅，在不让人感到负担的前提下，覆盖了足够广阔的思想空间。它既是希望弄清塑造我们社会的科学与数据背后机制的读者的便捷入门，也充分体现了作者作为顶尖数学家的深厚造诣。 In his book Six Math Essentials, Fields Medalist Terence Tao distills the vast field of mathematics into six fundamental concepts that have served as guideposts for thinkers from antiquity to the modern day. By focusing on these core pillars, the book offers a roadmap for understanding how quantitative reasoning allows us to navigate and interpret the complexities of our contemporary world.

The first of these essentials is the concept of numbers, which Tao frames as the primary gateway to quantitative thinking. From there, he moves into algebra, describing it as the essential tool for abstraction that allows us to generalize relationships. Geometry is introduced not just as a study of shapes, but as a practical method for calculating and measuring phenomena that exist well beyond the limits of human vision.

To address the inherent unpredictability of life, the book explores probability as a framework for navigating uncertainty with rigorous, logical thinking. These tools are complemented by the study of analysis, which provides the mathematical language necessary to tame extremes, whether they involve the incredibly large or the infinitesimally small. Finally, the book rounds out the collection with dynamics, which Tao presents as the definitive mathematics of change itself.

Clocking in at a slim, elegant length, the book manages to cover a vast amount of intellectual ground without feeling overwhelming. It serves as an accessible entry point for those looking to understand the mechanics behind the science and data that shape our society, while also reflecting the author's deep expertise as a preeminent mathematician.

在经历了多年销量下滑并于 2016 年被 Toyota 停产后，Scion 品牌正在迎来一场意外回归。该品牌最初于 2003 年推出，瞄准年轻人群，主打价格亲民、风格古怪的车型，并在巅峰时期曾以汽车行业里最低的平均购车者年龄而闻名。如今，它以"Scion by Toyota"的身份回归，把自己重新定位为"freedom lab"，致力于通过创新设计与实验性概念来挑战传统的出行方式。

此次复兴品牌旗下的首款量产车型是 KEI-01——一款基于现有 Toyota Pixis 打造的超紧凑型多用途卡车，而 Toyota Pixis 本身又是 Daihatsu Hijet Jumbo Cab 的换标车型。 KEI-01 在 Toyota 位于 Indiana 的体系内生产，作为真正的 kei 卡车，它搭载 660cc 发动机，并配备 CVT 变速箱。由于该车获得认证并符合美国低速车辆的安全标准，其行驶范围被限制在限速 35 mph 以下的道路上，因此更适合作为农业、仓储作业以及城市配送的专用工具，而不是普通的日常通勤车。

尽管车身尺寸大约只有五英尺宽、十一英尺长，KEI-01 却被打造为真正的干活利器。它提供 600 磅的有效载荷能力，配备可灵活折叠的三向货箱，并采用四轮驱动以应对不平整路面。车内同样具备加热与空调等现代化舒适配置。不过，Toyota 计划把首批上市限定在一个面向车队运营的利基试点项目中，时间从 2027 年底开始。至于未来是否向普通消费者开放，仍取决于首批车队推广后的实测数据与用户反馈。

除了计划量产的 KEI-01 外，"Scion by Toyota"阵容还包含两款实验性概念车，用以展示品牌新的探索方向。 SxS Concept 是一款偏硬派的越野 side-by-side 车型，搭载来自 Toyota Tacoma 的 300 马力涡轮增压发动机，并特别加入了"Silent Mode"，可实现纯电动模式的全工况运行。与之相对的 BKE Concept 则是一款为户外爱好者打造的个人出行设备。它小巧的车身尺寸和扎实耐用的轮胎，旨在连接 KEI-01 的实用属性与 SxS 的高性能能力，为穿行于狭窄的非铺装越野小路提供更便携的解决方案。 After being retired by Toyota in 2016 following years of declining sales, the Scion brand is making an unexpected comeback. Originally launched in 2003 to target a younger demographic with budget-friendly, quirky vehicles, the marque famously achieved the lowest average buyer age in the automotive industry during its peak. Now returning as "Scion by Toyota," the brand is repositioning itself as a "freedom lab" dedicated to challenging conventional mobility through innovative design and experimental concepts.

The first production vehicle under this revived identity is the KEI-01, an ultra-compact utility truck based on the existing Toyota Pixis, which is itself a rebadged Daihatsu Hijet Jumbo Cab. Manufactured using Toyota's infrastructure in Indiana, the KEI-01 is a genuine kei truck featuring a 660cc engine and a CVT transmission. Because it is certified to meet American low-speed vehicle safety standards, it is restricted to roads with speed limits under 35 mph, making it a specialized tool for agriculture, warehouse operations, and urban delivery rather than a conventional daily driver.

Despite its compact size of roughly five feet wide and eleven feet long, the KEI-01 is designed as a functional workhorse. It boasts a 600-pound payload capacity, a versatile three-way fold-down cargo bed, and four-wheel drive for navigating uneven terrain. While the vehicle includes modern comforts like heat and air conditioning, Toyota plans to limit its initial release to a niche pilot program focused on fleet usage starting in late 2027. Future consumer availability remains uncertain and will depend on performance data and feedback from the initial fleet rollout.

In addition to the production-bound KEI-01, the Scion by Toyota lineup features two experimental concepts that illustrate the brand's new focus on exploration. The SxS Concept is a rugged, off-road side-by-side vehicle powered by a 300-horsepower turbocharged engine from the Toyota Tacoma, which uniquely includes a "Silent Mode" that allows for fully electric operation. Complementing this is the BKE Concept, a personal mobility device designed for outdoor enthusiasts. Its small form factor and rugged tires are intended to bridge the gap between the utilitarian nature of the KEI-01 and the high-performance capabilities of the SxS, offering a portable solution for navigating narrow, off-road paths.

Starlink 已正式推出 Mini 2，这是其便携式卫星互联网硬件的下一代产品。该版本专为需要在偏远或恶劣环境中保持可靠连接的用户打造，采用了完全重新设计的机身外壳，比上一代更薄也更耐用。机身厚度仅 23.1mm，产品定位强调便携性，宣称可以和笔记本电脑一起装进包里，面向追求高机动性的用户。

Mini 2 的一项重要升级是内置了可快速更换的电池，可实现最长约 4.5 小时的独立运行。摆脱必须连接专用外部电源的限制后，用户可在露营、徒步或靠近水域工作时实现真正不受束缚的上网。电池额定容量为 71Wh，可通过标准 60W USB-C 接口充电。若需要更长时间使用，用户既可以更换备用电池，也可以在设备仍处于充电状态时继续保持连接。

Mini 2 的机身结构也体现了其面向户外的定位，具备 IP67 防护等级。该等级可确保设备在多种环境挑战下依然可用，包括路途中遇到的雨淋、在水上遭遇的风浪，以及海滩环境中的沙尘。为便于移动场景搭建，设备还提供了实用设计，例如内置支架，以及标准 1/4-20 相机螺纹接口，便于安装到多种三脚架配置上。

在连接能力方面，Mini 2 支持 Wi‑Fi 6，覆盖范围最高可达 1,800 平方英尺。尽管内部硬件有所调整，设置流程仍然简单：用户只需将设备对准天空，约 60 秒即可完成联网。整套产品包含 Starlink Mini 2 主机、可拆卸电池、电源适配器，以及一条 3 米 USB-C 线缆。

目前，Starlink 正在邀请美国境内的部分客户进行预订。尽管初期可用范围有限，但公司计划在不久的将来扩展到更多客户和更多国家。此次发布意味着 Starlink Mini 系列的关键进化，朝着更灵活、且更自成体系的使用生态迈进，为那些希望保持移动性的用户提供在不牺牲高速卫星连接前提下的解决方案。 Starlink has officially introduced the Mini 2, the next generation of its portable satellite internet hardware. This iteration is specifically designed for users who need reliable connectivity in remote or rugged environments, featuring a completely redesigned enclosure that is both thinner and more durable than its predecessor. With a sleek profile of just 23.1mm, the device is marketed as being compact enough to fit alongside a laptop in a bag, making it a highly mobile solution for adventurers.

A major upgrade in the Mini 2 is the inclusion of a quick-swap, built-in battery that provides up to 4.5 hours of independent operation. This shift away from requiring a dedicated external power source allows for true untethered internet access, whether you are camping, hiking, or working near water. The battery itself is rated at 71 Wh and can be charged via a standard 60W USB-C connection. If longer sessions are required, users have the flexibility to swap in a spare battery or remain connected while the unit is actively charging.

The physical construction of the Mini 2 reflects its focus on outdoor use, boasting an IP67 environmental rating. This level of protection ensures that the unit remains operational through various challenges, including rain on the trail, wind on the water, or sand at the beach. The hardware includes convenient touches for mobile setups, such as a built-in kickstand and a standard 1/4-20 camera thread mount, which allows for easy installation on a variety of tripod configurations.

Connectivity performance remains a priority, with the device supporting Wi-Fi 6 and covering an area of up to 1,800 square feet. Despite the internal hardware changes, the setup process remains straightforward, requiring users to simply point the device at the sky to get online in about 60 seconds. The complete package ships with the Starlink Mini 2 unit, the detachable battery, a power supply, and a 3-meter USB-C cable.

Currently, Starlink is opening preorders by invitation for select customers within the United States. While availability is initially limited, the company plans to expand access to more customers and countries in the near future. This launch marks a significant evolution for the Starlink Mini series, moving toward a more versatile and self-contained ecosystem for users who prioritize mobility without sacrificing high-speed satellite connectivity.

Mike Lothian 是一名开发人员，也是一位 Linux 用户。近日，他因为试图提升 DisplayLink 扩展坞在 Linux 上的性能，遭到来自 Synaptics 的法律威胁。由于对制造商提供的性能低下、且仅支持 x86 的专有驱动程序感到不满，Lothian 借助 LLM 辅助逆向工程，破解了专有的 DisplayLink DL3 协议。他开发了基于 Rust 的驱动程序 Vino，并将成果提交到 Linux kernel 邮件列表，希望能为 Linux 用户带来更好的功能对等性和稳定性。

然而，Lothian 没有得到技术层面的合作，反而接到了 Synaptics 法务部门的联系。对方声称他的工作侵犯了他们的专利组合。 Synaptics 于 2022 年收购了 DisplayLink，并坚持认为其技术并非开放标准，且受多项专利保护。公司随后发出正式的停止侵权通知，要求 Lothian 移除公开代码、停止开发，并提交签署的合规承诺；同时警告他若不照做，将面临高额的经济处罚。

Lothian 向 Open Invention Network（OIN）寻求支持，指出 Synaptics 是该组织成员，而 OIN 的宗旨正是保护开源开发者免受专利诉讼。然而，OIN 认定所涉及的具体技术——尤其是 DisplayLink 设备中使用的视频编解码器——不在其 Linux System Definition 的覆盖范围内。由于 OIN 的保护并未延伸到这一具体实现，Lothian 发现自己未能获得期望中的机构背书。

在面临一场他根本负担不起的昂贵且旷日持久的诉讼风险时，Lothian 最终选择遵从停止侵权要求：下架其 GitHub 代码库，并停止对内核的公开贡献。尽管他仍会私下运行该驱动程序供自己使用，但在诉讼威胁的阴影下，推动更稳定的开源方案并合入主线的努力已经基本被迫中止。

这一事件再次敲响警钟：尽管开源社区不断借助新工具与新的开发方式持续前进，软件专利仍然是开发者绕不开的巨大障碍。即便贡献者用现代技术弥补硬件支持的空白，企业依然可能通过知识产权法律来压制创新、维持专有控制，让当 Linux 用户遇到硬件支持不足时只能面临有限选择。 Mike Lothian, a developer and Linux user, recently faced legal threats from Synaptics after attempting to improve the performance of his DisplayLink docking station on Linux. Frustrated by the subpar performance and proprietary, x86-only driver provided by the manufacturer, Lothian used an LLM to help him reverse-engineer the proprietary DisplayLink DL3 protocol. He developed a Rust-based driver called Vino and submitted his work to the Linux kernel mailing list, hoping to achieve better feature parity and stability for Linux users.

Instead of receiving technical collaboration, Lothian was contacted by Synaptics's legal department, which asserted that his work infringed upon their patent portfolio. Synaptics, which acquired DisplayLink in 2022, maintains that their technology is not an open standard and is protected by numerous patents. The company issued a formal cease-and-desist notice, demanding that Lothian remove his public code, cease development, and provide a signed undertaking to comply, threatening him with heavy monetary penalties if he failed to do so.

Lothian reached out to the Open Invention Network (OIN) for support, noting that Synaptics is a member of the organization, which is designed to protect open-source developers from patent litigation. However, OIN concluded that the specific technology involved, particularly the video codecs used in DisplayLink devices, falls outside the scope of its Linux System Definition. Because the OIN's protections did not extend to this specific implementation, Lothian found himself without the institutional backing he had hoped for.

Facing the prospect of an expensive and protracted legal battle he could not afford, Lothian ultimately complied with the cease-and-desist demand by taking down his GitHub repository and halting his public contributions to the kernel. While he continues to run the driver privately for his own use, the effort to upstream a more stable, open-source solution for the Linux community has been effectively silenced by the threat of litigation.

This case serves as a stark reminder that while the open-source community continues to evolve with new tools and development methods, software patents remain a significant obstacle for developers. Even as contributors utilize modern technologies to bridge gaps in hardware support, companies can wield intellectual property law to stifle innovation and maintain proprietary control, leaving Linux users with limited options when hardware support falls short.

Modern Web Guidance 是一项旨在为代码代理（coding agents）提供最新 Web 平台专业知识与最佳实践的计划。由于许多大型语言模型都在包含大量 legacy code 的海量数据集上训练，它们往往会默认沿用过时的模式，从而在如今可由高性能的原生浏览器能力直接完成的任务上生成臃肿的 JavaScript 。该工具充当桥梁：将专家精心策划的指南直接注入到代理的上下文窗口中，确保生成的代码是现代、高效且易于访问的。

该体系覆盖数百项现代 Web 功能，从 container queries 和 subgrid 这类 CSS 布局方案，到用于性能、安全性以及内置 AI 的复杂 JavaScript API 。指南本身也经过 token-efficient 设计，会剔除模型已能很好理解的冗余内容，并将重点放在高影响领域，如性能优化、可访问性，以及对平台级回退（platform-level fallbacks）的负责任使用。借助该工具，开发者可以让 AI 生成的代码优先利用性能与原生浏览器能力，而非依赖臃肿的第三方库。

实施 Modern Web Guidance 十分便捷：提供命令行界面，允许代理在本地执行语义搜索，并通过 ID 获取特定指南，无需网络调用或 API keys 。系统还提供可选的配置文件，便于团队管理实验性特性或易变 API 的设置。该工具兼容多种开发环境与代理类型，包括 GitHub Copilot 、 Claude Code 以及各类基于 CLI 的工作流。

为检验这些指南的效果，该项目采用一套严格的评估基准（evaluation harness），对比有无指南时代码代理的表现。项目通过闭环的模拟开发任务、基于浏览器的断言（assertions）以及 Playwright 的评测（graders），持续校准其建议。结果表明，多种模型在任务通过率上都有显著提升，这为整合这些现代化指南能够产出更高质量、可直接投入生产的代码提供了实证依据。 Modern Web Guidance is an initiative designed to provide coding agents with the latest web platform expertise and best practices. Because many large language models are trained on massive datasets containing legacy code, they often default to outdated patterns that result in bloated JavaScript for tasks now handled by high-performance native browser features. This tool acts as a bridge, delivering expert-curated guidelines directly into an agent's context window to ensure the code produced is modern, efficient, and accessible.

The system covers hundreds of modern web features, ranging from CSS layout systems like container queries and subgrid to complex JavaScript APIs for performance, security, and built-in AI. These guides are specifically designed to be token-efficient, pruning redundant information that models already understand well while focusing on high-impact areas like performance optimization, accessibility, and responsible use of platform-level fallbacks. By using this tool, developers can ensure their AI-generated code prioritizes performance and native browser capabilities over heavy third-party libraries.

Implementing Modern Web Guidance is straightforward, with a command-line interface that allows agents to perform local semantic searches and retrieve specific guides by ID without requiring network calls or API keys. The system includes an optional configuration file that allows teams to manage settings for experimental features or volatile APIs. The tool is compatible with a wide array of development environments and agents, including GitHub Copilot, Claude Code, and various CLI-based workflows.

To verify the effectiveness of these guidelines, the project utilizes a rigorous evaluation harness that compares coding agent performance with and without the guidance. Using a closed-loop pipeline of simulated developer tasks, browser-based assertions, and Playwright graders, the project continuously calibrates its advice. Results have shown significant uplift in task pass rates for various models, providing empirical proof that integrating these modern guidelines leads to higher-quality, production-ready code.

Anthropic 推出了 Claude Whiteboards 和 Claude Sheets，目前已进入测试版；同时将 Claude Dashboards 更名为 Claude Data，现已全面可用。这些更新的目标是让工作流程始终留在 Claude 环境中，避免反复把输出搬到外部程序。通过让团队能够在平台内直接完成头脑风暴、构建与数据分析，用户现在可以在不丢失工作背景的前提下保持连续性，并实现实时协作。

Claude Whiteboards 提供了团队头脑风暴的共享空间。用户可以和 Claude 一起梳理初步方案，随后 Claude 会把想法整理成结构化看板，包含时间线、负责人和具体日期。该工具是可交互的：团队成员可以添加备注，Claude 再根据主题对备注进行分组，或将其转化为可执行事项。会议结束后，这些看板可导出到 Lucid 、 Miro 或 Vani by Zoho 等外部工具，以继续推进协作。

Claude Sheets 可根据简单的自然语言提示生成带有公式、格式和下拉选项的电子表格。它面向团队协作而设计，支持用户与 Claude 之间进行共同编辑。当前功能已覆盖常见需求；若用户需要更接近 Excel 的强大能力，可将内容导出为 .xlsx 或 CSV 格式，或迁移到 Airtable 、 Asana 和 Smartsheet 等合作平台以继续扩展。

Claude Data（原 Claude Dashboards）是一套扩展的工具，用于与来自 Amazon Redshift 、 BigQuery 和 ClickHouse 等数据源的公司数据进行交互。本次更新包含经认证的指标定义，用于避免口径冲突；支持创作数据驱动的故事；并提供实时共享，让同事无需直接访问底层连接器也能查看最新数字。此外，平台还加入了设计系统，改进缓存以加快加载速度，并记录变更以支持版本控制。

在这些特定工具之外，Claude 还支持交互式连接器：连接器会以完整应用的形式在对话旁边打开，用户可直接在 Claude 界面中使用 Adobe 、 Notion 或 Atlassian 等工具。这些功能适用于所有付费计划，并为团队与企业管理员提供相应设置以管理访问权限。通过这一系列更新，平台力求带来更紧密的一体化体验，让用户能够以"所想即所得"的效率完成各类专业任务。 Anthropic has introduced Claude Whiteboards and Claude Sheets, now available in beta, while rebranding Claude Dashboards as Claude Data, which is now generally available. These updates aim to keep workflows within the Claude environment, eliminating the need to constantly move outputs to external programs. By allowing teams to brainstorm, build, and analyze data directly inside the platform, users can now maintain continuity and real-time collaboration without losing the context of their work.

Claude Whiteboards functions as a shared space for team brainstorming. Users can walk through rough plans with Claude, which then organizes ideas into structured boards featuring timelines, owners, and specific dates. The tool is interactive, enabling team members to add notes that Claude can subsequently group by theme or turn into actionable items. Once a session is complete, these boards can be exported to external tools such as Lucid, Miro, or Vani by Zoho to continue the collaboration process.

Claude Sheets enables the generation of spreadsheets equipped with formulas, formatting, and drop-downs based on simple natural language prompts. Designed for team use, these sheets allow for collaborative editing between users and Claude. While the current functionality covers standard needs, users can export their work to .xlsx or CSV formats if they require the more robust capabilities of Excel, or they can transition into partner platforms like Airtable, Asana, and Smartsheet for further development.

Claude Data, formerly known as Claude Dashboards, serves as an expanded suite for interacting with company data from sources like Amazon Redshift, BigQuery, and ClickHouse. This update includes features such as certified metric definitions to prevent conflicting data, the ability to craft data-driven stories, and live sharing so colleagues can view up-to-date numbers without needing direct access to underlying connectors. The platform also now supports design systems, includes improved caching to speed up loading times, and tracks changes to allow for version control.

Beyond these specific tools, Claude now supports interactive connectors that open as full applications alongside conversations, allowing users to work in tools like Adobe, Notion, or Atlassian directly from the Claude interface. These features are available on all paid plans, with specific settings for team and enterprise administrators to manage access. Through these updates, the platform aims to provide a more integrated experience that allows users to work at the speed of thought across various professional tasks.

Val Town 已正式宣布这一艰难决定：停止运营并关闭该平台。 Steve Krouse 代表团队撰文，感谢社区的信任与创造力，同时也坦言在竞争激烈的云计算市场中要脱颖而出并不容易。

为帮助用户顺利完成过渡，公司已立即暂停计费；所有现有订阅将继续免费提供服务。对于已预付 annual plans 的用户，公司将发放部分退款以确保公平。团队还强调，由于 Val Town 基于常规技术搭建，迁移到 Cloudflare 、 Vercel 、 Netlify 或 Render 等其他提供商整体上很直接，借助当下的现代 large language models 更是如此。

关闭流程将按既定时间表推进，让用户有足够时间导出数据并迁移项目。自即日起已关闭新注册与计划升级功能。到 2026 年 12 月 10 日，平台将对大多数服务进入只读模式，但 sqlite 和 blob storage 仍会保留一段更长时间。其余所有服务计划于 2027 年 1 月 10 日停止运行。

最后，团队计划在 2027 年 2 月 1 日将平台开源，作为对社区的收官回馈。 Krouse 同时感谢过去四年中为项目付出的敬业团队成员，也感谢在整个发展过程中一直支持该项目的投资者。若用户在迁移过程中需要特定帮助，建议联系 support 或加入社区 Discord 。 Val Town has officially announced the difficult decision to cease operations and wind down the platform. Steve Krouse, writing on behalf of the team, expressed gratitude to the community for their trust and creativity, while acknowledging the challenge of competing in the crowded cloud computing market.

To assist users during this transition, the company has immediately paused billing, and all existing subscriptions will continue to be serviced for free. For those who pre-paid for annual plans, the company is issuing partial refunds to ensure fairness. The team emphasized that because Val Town was built upon standard technologies, migrating to other providers like Cloudflare, Vercel, Netlify, or Render is a straightforward process, especially with the help of modern large language models.

The wind-down process follows a specific timeline to allow users time to extract their data and migrate their projects. New signups and plan upgrades have been disabled effective immediately. By December 10, 2026, the platform will enter a read-only mode for most services, though sqlite and blob storage will remain available slightly longer. All remaining services are scheduled to shut down on January 10, 2027.

Finally, the team intends to open-source the platform on February 1, 2027, as a concluding act for the community. Krouse concluded by thanking the dedicated team members who contributed over the past four years, as well as the investors who supported the venture throughout its lifespan. Users are encouraged to reach out to support or the community Discord if they require specific assistance with their migration process.