Top 10 Hacker News posts, summarized
HN discussion
(584 points, 240 comments)
Adafruit received a demand letter from law firm Fenwick & West on behalf of Flux.ai on May 22, 2026. The letter accused Adafruit of making false and potentially defamatory claims about Flux's intellectual property, commercial traction, and user base, and threatened legal action under the Computer Fraud and Abuse Act (CFAA). Adafruit stated it accessed only publicly available information due to a server misconfiguration on Flux's part and characterized its actions as responsible disclosure on a matter of public security interest. Adafruit has paused its blog posts while considering its response.
The top comments reveal widespread skepticism toward Flux.ai, with several users sharing negative product experiences, criticizing its token-based pricing model, and questioning the quality of its AI-powered PCB design tools. Some commenters linked to Reddit discussions alleging deceptive billing practices and concerns about Flux's use of third-party IP, including Adafruit's own component footprints. Many viewed the legal threat as bullying, with some suggesting it was an attempt to suppress negative information. The discussion also included speculation that Flux's recent funding may have motivated the aggressive legal stance.
HN discussion
(435 points, 340 comments)
The article argues that global age verification measures for social media, ostensibly to protect children, actually serve as a foundation for government-controlled internet and mass surveillance. Instead of forcing platforms to block existing child users (which they already identify through data collection), governments mandate identity verification that eliminates online anonymity and enables censorship. This creates risks for activists and a chilling effect on free speech. The trend is accelerating worldwide, with countries like Australia, Brazil, and the UK implementing restrictions, and proposals emerging elsewhere. Age verification essentially functions as identity verification, exposing users to data breaches and state overreach. The EU's proposed app, while claiming anonymity, lacks robust Zero-Knowledge Proof (ZKP) safeguards and could be easily disabled by authorities. The article warns this path leads to broader controls on VPNs and operating systems, potentially culminating in state-mandated devices, and dismisses child safety as a pretext for surveillance.
HN comments critique the article's framing, noting that dismissing age verification proponents as "authoritarians" is unproductive and that privacy-preserving systems like ZKP can be implemented securely, akin to end-to-end encryption. Skepticism exists about the EU's ZKP implementation, with concerns it lacks proper safeguards despite claims of anonymity. Comments highlight that surveillance on major platforms like Facebook already occurs, making age verification an incremental rather than revolutionary threat. Alternative solutions are proposed, such as creating separate TLDs (e.g., .social) for adult content that parents can easily block. Some express cynicism about the current internet's state and call for a return to simpler, decentralized protocols like IRC. Others question whether social media was ever truly free and emphasize practical bypass strategies like VPNs and open-source systems, while noting the article's dystopian predictions (e.g., police searches for unauthorized phones) may be hyperbolic.
HN discussion
(481 points, 276 comments)
The author details their experience with Gmail's new AI features, which they found intrusive and disrespectful. The service automatically summarizes emails, generates draft replies, and repeatedly prompts the user to use AI writing assistance like "Help me write" and "Tab to improve." The author views these unsolicited features as a message that they are not capable of managing their own communication. After discovering that disabling these features is difficult and sometimes requires sacrificing useful functions, the author decided to leave the platform they had used for 16 years. They are migrating to Fastmail for greater control and a cleaner user experience, expressing frustration that Google's own stability contributed to making the break more impressive.
The Hacker News discussion echoed the author's sentiment of frustration with "enshittification" and user-hostile design, with many comparing the experience to Clippy but noting it's more aggressive. A key insight is that the core issue isn't the AI features themselves, but the inability to opt-out, which users describe as a "theft of consent" and a "hard nudge" or "shove." Commenters also highlighted a broader trend of companies like LinkedIn and Microsoft shoving unwanted AI integration into products. There was significant concern about the negative impact, including AI hallucinations causing real-world distress and the difficulty of finding a viable alternative due to the entrenched nature of Gmail's ecosystem.
HN discussion
(417 points, 222 comments)
The article enthusiastically advocates for the Janet programming language, a small Lisp dialect, highlighting its key advantages for side projects. Janet is praised for its extreme simplicity—only eight core instructions and a small standard library that can be learned quickly. Key features include effortless compilation into small native executables (under 1MB), powerful text parsing via PEGs instead of regex, an elegant subprocess DSL for shell scripting, easy embeddability as a C library, support for both mutable and immutable collections, and a powerful macro system. Notably, Janet allows passing values from compile-time to run-time, enabling unique capabilities like runtime code evaluation and asset embedding. The author emphasizes its modern, comfortable syntax and departure from traditional Lisp conventions.
The Hacker News discussion centers on Janet's practical appeal and lingering concerns. Key themes include praise for its macro system, compile-time features, and embeddability, with users sharing positive experiences replacing Python/Shell scripts. However, syntax remains a major friction point; many commenters criticize the parentheses as difficult to read and edit, referencing alternative syntax projects. Performance questions ("How fast is it?") and the lack of static typing were frequently raised. Community members noted comparisons to languages like Hy, Fennel, and TCL, highlighted specific projects built with Janet (bauble.studio, toodle.studio), and debated the advantages of PEGs over regex. Some expressed frustration with learning resources ("Janet for Mortals" not being accessible enough), while others enthusiastically shared adoption stories, emphasizing its fast startup time and utility for system scripting.
HN discussion
(359 points, 223 comments)
This article provides a walking tour guide of surveillance infrastructure in downtown Seattle, created as part of a Tech Equity Coalition workshop. The guide details various technologies and locations, including general surveillance cameras, Amazon Go stores, automated license plate readers (ALPRs), Acyclica Wi-Fi tracking devices, the Washington State Fusion Center, and an AT&T peering site. Each entry specifies the address, appearance, function, technical workings, and social implications of the surveillance technology, along with discussion prompts and references. The guide aims to help people identify and understand the layers of the "smart city" that collect and store data about their lives.
The Hacker News discussion focused on the accuracy and accessibility of the guide. Several commenters pointed out that the article is outdated, as it was published in 2019, with some examples, like the Amazon Go store, no longer being current. There was significant criticism of the guide's technical inaccuracies and its use of jargon, with one commenter calling it "po-mo art-school language" and another suggesting it lacked plain-language accessibility. Other top comments highlighted the complex trade-offs between surveillance and public safety, with one user expressing a willingness to trade privacy for safety in light of rising crime rates and another noting the public's general lack of concern or technical understanding regarding privacy trade-offs.
HN discussion
(323 points, 213 comments)
The article argues for replacing traditional cron jobs with systemd timers, asserting they offer a more robust, feature-rich, and predictable solution for scheduled tasks in Linux. It highlights key advantages of systemd timers, including better handling of environment variables (like $PATH), integrated logging via the journal, clearer and more flexible scheduling syntax (both calendar-based and relative time spans), and special features like system wake-up from suspension (`WakeSystem=`), randomized execution delays to prevent "thundering herd" problems (`RandomizedOffsetSec=`), and persistent execution to catch up on missed jobs (`Persistent=`). The author provides a practical example, demonstrating how to create a simple service and associate it with a timer, and concludes by recommending the use of systemd's command-line tools like `systemctl list-timers` for management.
The Hacker News discussion largely corroborates the article's praise for systemd timers, with many users sharing positive experiences of migrating from cron. Key benefits cited include better resilience to system reboots, seamless integration with the journal for logging, and superior handling of missed executions. However, significant criticism was aimed at the complexity and verbosity of systemd's syntax, with several users finding the unit file format "ugly" and less intuitive than cron. Other points of friction mentioned include the need to manage separate `.service` and `.timer` files and the lack of a simple way to force-trigger a timer for testing. The discussion also noted that systemd timers are not universally available on all Linux distributions, posing a challenge for cross-platform tooling.
HN discussion
(318 points, 146 comments)
Microsoft has launched MAI-Code-1-Flash, a new coding model integrated into GitHub Copilot for Visual Studio Code. Built on clean, licensed data by Microsoft, it emphasizes agentic coding in real development environments, adaptive response lengths (concise for simple tasks, deeper reasoning for complex ones), and strong instruction-following. The model is designed for production workflows rather than pure benchmark optimization, evaluated using telemetry-grounded tasks from actual GitHub Copilot usage. It benchmarked higher pass rates than Claude Haiku 4.5 across core coding evaluations (SWE-Bench Verified, Pro, Multilingual, Terminal Bench 2), including a 16-point lead on SWE-Bench Pro (51.2% vs. 35.2%), and achieved up to 60% fewer tokens for harder problems, improving efficiency and reducing latency.
Hacker News comments raised several points: skepticism about the "clean data" claim given LLM dataset pollution, criticism of benchmarking against an older Claude version (4.5 instead of 4.8), and concerns about the 51% SWE-Bench Pro pass rate being too low for reliable use. Users questioned the model's practical value compared to open models like Qwen3.6-35B-A3B (49.5%) and whether smaller models like this are truly useful for coding versus larger ones. Technical details emerged (137B parameters, A5B structure), and there was demand for tokens/second metrics and open-weight availability. Frustration was expressed over forced scrolling effects on Microsoft's website.
HN discussion
(147 points, 191 comments)
Anthropic is expanding Project Glasswing, an initiative to secure critical global software by providing access to its Claude Mythos Preview model. The program, which initially involved about 50 partners, has now been extended to approximately 150 new organizations across over 15 countries, including those in power, water, healthcare, and communications. These partners maintain vital infrastructure, and a successful attack on their systems could impact over 100 million people. The project aims to help the software industry adapt to the rise of powerful AI-driven cyber threats, such as increased and unpredictable attacks, by providing tools for vulnerability scanning, patching, and security improvements.
HN users expressed skepticism about Anthropic's motives, with many comments framing Project Glasswing as a marketing strategy or a precursor to an IPO. Critics questioned the program's transparency, noting the lack of a partner list and doubting the uniqueness of Mythos's capabilities compared to competitors like OpenAI. Some comments accused Anthropic of deliberately restricting access to create scarcity and drive enterprise sales, while others dismissed the hype as performative, citing examples where Mythos reportedly failed to find critical vulnerabilities. A user also highlighted environmental concerns, questioning Anthropic’s altruistic claims. Despite the skepticism, a few engaged in technical discussions about memory safety and open-source alternatives.
HN discussion
(128 points, 156 comments)
KDE Plasma will officially remove X11 support in version 6.8, scheduled for release in approximately five months. This includes eliminating the X11 login session option and removing all X11-specific code from Plasma Shell, System Settings, and device configuration. However, XWayland will remain supported for running legacy X11 applications, and KDE applications will continue to function in X11 on other desktop environments. The transition is driven by over 95% of Plasma 6.6 users already on Wayland, with minimal X11 development activity internally. KDE asserts this shift will enable future performance optimizations and new features by focusing solely on a Wayland codebase. Plasma 6.7 will be the last release with an X11 session, and users experiencing issues are encouraged to report them before the final removal.
The HN discussion reflects mixed reactions to KDE's X11 retirement. Many users praise KDE's Wayland implementation for its performance and responsiveness, though several commenters criticize the lack of specificity about upcoming features enabled by the transition. Significant concerns center on Wayland's limitations compared to X11, including reduced functionality for applications requiring window control (e.g., mpv embedding, Chrome PiP), accessibility tools like Talon being unsupported, and KRdp's immaturity for remote desktop use cases. Some users expressed frustration about forced workflow changes, while others indicated plans to switch desktop environments or abandon KDE entirely due to compatibility losses. Accessibility was a major theme, with warnings about legal risks under the Americans with Disabilities Act and critiques that KDE's Wayland stack lacks essential accessibility protocols present in GNOME's Mutter. A smaller subset of comments highlighted potential impacts on gaming and BSD support.
HN discussion
(148 points, 102 comments)
Unable to fetch article: HTTP 403
The Hacker News discussion centers on Trump's downsized AI executive order, noting it significantly reverses earlier drafts, particularly reducing the required pre-release government review period from 90 to 30 days after industry objections. While some view the order as a positive step toward necessary AI regulation and national security, mirroring past bipartisan efforts, others criticize it as vague, lacking substance, and potentially enabling government censorship of AI outputs. Key concerns raised include the enforceability of voluntary reviews, the risk of politically motivated content controls, and the potential to disadvantage open-source or international competitors, alongside skepticism about the government's capacity for meaningful evaluation. Reactions also highlight political timing speculation (e.g., Anthropic's IPO) and fundamental objections to government review of AI models as an overreach.
Generated with hn-summaries