Skip to content

Instantly share code, notes, and snippets.

@puf

puf/lobsters-digest.xml

Last active March 13, 2026 17:16
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save puf/b3df69776c3f7a506e278c62b03cf4fa to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save puf/b3df69776c3f7a506e278c62b03cf4fa to your computer and use it in GitHub Desktop.
Download ZIP
Lobsters Digest Feed
Raw
lobsters-digest.xml
<?xml version='1.0' encoding='utf-8'?>
<rss xmlns:ns0="http://www.w3.org/2005/Atom" version="2.0">
<channel>
<title>Lobsters</title>
<link>https://lobste.rs/</link>
<ns0:link href="https://lobste.rs/rss" rel="self" />
<description />
<pubDate>Fri, 13 Mar 2026 06:41:55 -0500</pubDate>
<ttl>120</ttl>
<lastBuildDate>Fri, 13 Mar 2026 09:08:27 GMT</lastBuildDate><item><title>Well Being in Times of Algorithms</title><link>https://www.ssp.sh/blog/well-being-algorithms/</link><guid>https://lobste.rs/s/3elkxf</guid><author>ssp.sh by sspaeti</author><pubDate>Fri, 13 Mar 2026 04:29:23 -0500</pubDate><comments>https://lobste.rs/s/3elkxf/well_being_times_algorithms</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The essay "Well Being in Times of Algorithms" critiques the detrimental effects of algorithm-driven social media on human attention spans and overall well-being. It highlights the importance of health, family, and self-contentment as foundational pillars for a fulfilling life, contrasting them with the attention-seeking and status-driven nature of current digital platforms. The author advocates for a return to an "open web" with direct connections, promoting solutions like the AT Protocol-based Bluesky and personal blogs as alternatives to counteract the negative influence of big tech and AI-generated content.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>browsers</category><category>mobile</category><category>web</category><category>ai</category></item><item><title>OpenClaw and the Dream of Free Labour</title><link>https://entropytown.com/articles/2026-03-12-openclaw-sandbox/</link><guid>https://lobste.rs/s/dm1j3a</guid><author>entropytown.com via chaosprint</author><pubDate>Fri, 13 Mar 2026 10:49:15 -0500</pubDate><comments>https://lobste.rs/s/dm1j3a/openclaw_dream_free_labour</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; This article analyzes OpenClaw as the latest embodiment of the belief that software can function as "free labor," comparing it to historical industrial shifts and the internet's "earn while you sleep" promises. It explores how OpenClaw packages existing AI components to create an impression of autonomous staff, highlighting significant security vulnerabilities and the industry's subsequent move towards sandboxing. The author argues that while sandboxing addresses some immediate risks, it doesn't solve the core issue of AI models' inherent unreliability or the broader security implications of granting extensive permissions. The piece concludes that the appeal of 24/7 automation often conflates continuous runtime with continuous judgment, driven by a fear of missing out on cheap labor.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>ai</category></item><item><title>The Plumbing of Everyday Magic</title><link>https://plumbing-of-everyday-magic.hyperclay.com</link><guid>https://lobste.rs/s/ehohsb</guid><author>plumbing-of-everyday-magic.hyperclay.com via hannahilea</author><pubDate>Fri, 13 Mar 2026 10:17:02 -0500</pubDate><comments>https://lobste.rs/s/ehohsb/plumbing_everyday_magic</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; "The Plumbing of Everyday Magic" critiques modern web development, highlighting the "trap doors" that divert creators from their core work into complex infrastructure management. The author uses the metaphor of plumbing to describe the hidden complexities of state sync, schema lock-in, and deployment. The essay advocates for "surface-level artifacts" and "malleable documents" that integrate creation and deployment, allowing for direct shaping and sharing of experiences without getting entangled in backend operations, envisioning a web where personal software can be easily created and remixed.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>web</category></item><item><title>Bucketsquatting is (Finally) Dead</title><link>https://onecloudplease.com/blog/bucketsquatting-is-finally-dead</link><guid>https://lobste.rs/s/bspgu4</guid><author>onecloudplease.com via caius</author><pubDate>Fri, 13 Mar 2026 06:06:24 -0500</pubDate><comments>https://lobste.rs/s/bspgu4/bucketsquatting_is_finally_dead</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; AWS has introduced a new namespace protection mechanism for S3 buckets to finally resolve the long-standing problem of bucketsquatting, where attackers could register names of previously deleted buckets. The new recommended naming convention, `&lt;yourprefix&gt;-&lt;accountid&gt;-&lt;region&gt;-an`, ensures that only the owning account can create a bucket with that name, preventing malicious takeovers for new buckets. This can be enforced via SCP policies, though existing buckets require migration for protection. The article also briefly compares how Google Cloud Storage and Azure Blob Storage address similar bucket naming security concerns.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Discussion revolved around the practical implementation of AWS's new namespace. &lt;a href="https://lobste.rs/c/uemgmy"&gt;eduard&lt;/a&gt; raised a concern about how AWS would handle "namespace collisions" with existing buckets that might coincidentally match the new format, to which &lt;a href="https://lobste.rs/c/r0xqmj"&gt;hc&lt;/a&gt; suggested AWS could potentially check all existing buckets.&lt;/p&gt;</description><category>devops</category></item><item><title>Reinventing Python's AsyncIO</title><link>https://blog.baro.dev/p/reinventing-pythons-asyncio</link><guid>https://lobste.rs/s/jadk6d</guid><author>blog.baro.dev by gi0baro</author><pubDate>Fri, 13 Mar 2026 10:17:52 -0500</pubDate><comments>https://lobste.rs/s/jadk6d/reinventing_python_s_asyncio</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; In "Reinventing Python's AsyncIO," gi0baro criticizes the complexity and limitations of Python's built-in AsyncIO, drawing inspiration from projects like tinyio. The author introduces TonIO, a new async runtime primarily developed in Rust, designed for simplicity around the concepts of "events" and "waiters" for managing coroutine suspension and resumption. TonIO aims to provide a multi-threaded async experience similar to Rust's Tokio, optimized for free-threaded Python, and demonstrates significant performance improvements (2-3.5x faster) compared to AsyncIO in benchmarks. The article advocates for an "unreasonable" approach to development, encouraging engineers to rethink existing solutions and build novel, well-designed code to improve the status quo.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>python</category></item><item><title>How to use storytelling to fit inline assembly into Rust</title><link>https://www.ralfj.de/blog/2026/03/13/inline-asm.html</link><guid>https://lobste.rs/s/hfypk8</guid><author>ralfj.de via ohrv</author><pubDate>Fri, 13 Mar 2026 09:08:08 -0500</pubDate><comments>https://lobste.rs/s/hfypk8/how_use_storytelling_fit_inline_assembly</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Ralfj.de explores how inline assembly and FFI integrate with the Rust Abstract Machine, proposing a "storytelling" approach. This method requires each inline assembly block to have a corresponding Rust code "story" that describes its effects on the Abstract Machine state, which is then used for compiler reasoning and ensuring soundness. The article demonstrates this principle with examples like pure instructions and page table manipulations, arguing that it prevents Undefined Behavior and allows safe, principled use of low-level code within Rust.
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.</description><category>assembly</category><category>rust</category></item><item><title>What are you doing this weekend?</title><link>https://lobste.rs/s/jwdr0q/what_are_you_doing_this_weekend</link><guid>https://lobste.rs/s/jwdr0q</guid><author> by caius</author><pubDate>Fri, 13 Mar 2026 05:25:04 -0500</pubDate><comments>https://lobste.rs/s/jwdr0q/what_are_you_doing_this_weekend</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; This article is an open invitation for readers to share their weekend plans, offering a space to discuss activities, seek feedback, or simply state their intention to relax. It explicitly notes that doing nothing is a perfectly valid option.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters shared a diverse range of weekend plans. Many are looking forward to a break from technology, such as &lt;a href="https://lobste.rs/c/kigwmv"&gt;mtset&lt;/a&gt;, who just finished a difficult writing task, and &lt;a href="https://lobste.rs/c/9jqqnj"&gt;alper&lt;/a&gt;, who plans on cycling. Others are engaging in practical tasks like taxes (&lt;a href="https://lobste.rs/c/hqjmlp"&gt;icefox&lt;/a&gt;), moving homes (lthms, apazzolini), or recovering from surgery (&lt;a href="https://lobste.rs/c/foonvo"&gt;nemin&lt;/a&gt;). Tech-related activities include &lt;a href="https://lobste.rs/c/l9jdes"&gt;invlpg&lt;/a&gt; brainstorming FreeBSD/OpenBSD integration with Zig, &lt;a href="https://lobste.rs/c/xmbloa"&gt;mcherm&lt;/a&gt; working on a Simplenote replacement, and &lt;a href="https://lobste.rs/c/neqbyy"&gt;mt&lt;/a&gt; setting up a consultancy website and exploring FreeBSD jails. Several are enjoying family time (halper, apazzolini), taking trips (dhruvp), or dealing with phone issues (toastal). A detailed discussion by chrismorgan covered migrating a personal server to Oracle Cloud's free tier in India.&lt;/p&gt;</description><category>programming</category><category>ask</category></item><item><title>Computing in freedom with GNU Emacs</title><link>https://protesilaos.com/codelog/2026-03-13-computing-in-freedom-with-gnu-emacs/</link><guid>https://lobste.rs/s/czj5p4</guid><author>protesilaos.com via achyudh</author><pubDate>Fri, 13 Mar 2026 08:34:02 -0500</pubDate><comments>https://lobste.rs/s/czj5p4/computing_freedom_with_gnu_emacs</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Protesilaos Stavrou presents a holistic introduction to GNU Emacs, highlighting its utility and adherence to free software principles. The article details Emacs' capabilities as an extensible text editor with features like Unicode support, multiple fonts, and graphics display. It emphasizes how Emacs Lisp allows for deep customization, fostering a strong community, and grants users significant control over their computing experience, leading to a consistent workflow, enhanced productivity, and true software freedom.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion features a comment from &lt;a href="https://lobste.rs/c/tykup3"&gt;trevorflowers&lt;/a&gt;, who shares their recurring experience of adopting and then abandoning Emacs due to the continuous effort required for customization and learning, which often sidetracked them from other projects. They express a renewed interest in trying Emacs again, particularly in retirement, hoping to dedicate enough time to build a comfortable and productive environment within it.&lt;/p&gt;</description><category>lisp</category><category>emacs</category><category>editors</category></item><item><title>Windows 11 after two decades of macOS: okay, but also awful</title><link>https://rakhim.exotext.com/windows-11-experience</link><guid>https://lobste.rs/s/zcmoab</guid><author>rakhim.exotext.com by freetonik</author><pubDate>Fri, 13 Mar 2026 10:15:03 -0500</pubDate><comments>https://lobste.rs/s/zcmoab/windows_11_after_two_decades_macos_okay</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; An author recounts their experience switching from macOS to Windows 11 after two decades, prompted by a dying iMac and a desire to use a beefy PC for programming, music production, and gaming. While finding Windows 11 "not as bad as I thought" and even more productive in some areas, the author ultimately prefers Mac. Key struggles include the absence of the "Things 3" app and, more significantly, adapting to Windows' keyboard bindings, leading to extensive customization attempts that felt like "wasted time." Positive aspects noted include Windows Explorer, Task Bar, the Winget package manager, powerful third-party software like File Pilot and Everything, native screen scaling, superior window management, better Unreal Engine performance, smooth gaming, Phone Link, and the utility of WSL for development.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>windows</category></item><item><title>I traced $2 billion in nonprofit grants and 45 states of lobbying records to figure out who's behind the age verification bills</title><link>https://www.reddit.com/r/linux/s/WqwsCnFE8E</link><guid>https://lobste.rs/s/ddrcpa</guid><author>reddit.com via rjzak</author><pubDate>Fri, 13 Mar 2026 06:41:55 -0500</pubDate><comments>https://lobste.rs/s/ddrcpa/i_traced_2_billion_nonprofit_grants_45</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The article traces $2 billion in nonprofit grants and lobbying records across 45 states to identify the entities behind age verification bills, revealing a company that profits from data collection is actively involved in shaping laws designed to acquire more personal data.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion largely focuses on the moderation status of the original Reddit post, with users like &lt;a href="https://lobste.rs/c/vbmfcl"&gt;raymii&lt;/a&gt; and &lt;a href="https://lobste.rs/c/buirqo"&gt;3digitdev&lt;/a&gt; reporting that it was awaiting moderator approval, leading to speculation that Meta-associated actors were mass reporting it (&lt;a href="https://lobste.rs/c/evkwu0"&gt;jamesgecko&lt;/a&gt;). On the content itself, &lt;a href="https://lobste.rs/c/d2otwt"&gt;ngp&lt;/a&gt; observes that Colorado and California's age verification laws appear to incorporate privacy considerations, while &lt;a href="https://lobste.rs/c/b5hfqg"&gt;nyarmith&lt;/a&gt; conveys exhaustion with corporate influence on legislation and suggests contacting the EFF and FSF. Additionally, &lt;a href="https://lobste.rs/c/rwg4nq"&gt;dpc_pw&lt;/a&gt; recommends NextDNS for parental control.&lt;/p&gt;</description><category>privacy</category><category>law</category></item><item><title>This Is Not The Computer For You</title><link>https://samhenri.gold/blog/20260312-this-is-not-the-computer-for-you/</link><guid>https://lobste.rs/s/zyt5uz</guid><author>samhenri.gold via msangi</author><pubDate>Thu, 12 Mar 2026 11:43:27 -0500</pubDate><comments>https://lobste.rs/s/zyt5uz/this_is_not_computer_for_you</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Sam Henri Gold's article "This Is Not The Computer For You" critiques computer reviews that dictate who a product is for, arguing that true learning and obsession often come from pushing inadequate hardware to its limits. Referencing his own childhood experience with a low-spec iMac and the new MacBook Neo, he asserts that even with limited resources, the Neo offers the full macOS and teaches fundamental lessons about computing constraints, unlike Chromebooks. The author concludes that this computer is for the aspiring user who will discover their path by stretching the machine's capabilities.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion largely revolves around the philosophy of optimizing software versus hardware. &lt;a href="https://lobste.rs/c/alwour"&gt;jcs&lt;/a&gt; advocates for using slow computers to encourage software optimization, preventing the creation of bloated applications. &lt;a href="https://lobste.rs/c/4g3ybj"&gt;Qyriad&lt;/a&gt; suggests a middle ground of compiling on fast machines but testing on slower ones. &lt;a href="https://lobste.rs/c/u6w18r"&gt;miro&lt;/a&gt; and &lt;a href="https://lobste.rs/c/jklayt"&gt;op&lt;/a&gt; share experiences of daily driving low-memory machines, reinforcing the idea that it highlights the inefficiency of modern software. &lt;a href="https://lobste.rs/c/xc4uzu"&gt;vpr&lt;/a&gt; introduces Ken Thompson's perspective, questioning the value of extreme optimization given Moore's Law. Other comments touch on performance relative to production environments and applying the "slow machine" principle to server applications.&lt;/p&gt;</description><category>hardware</category><category>mac</category></item><item><title>How do you manage SSH keys?</title><link>https://lobste.rs/s/zcoz8h/how_do_you_manage_ssh_keys</link><guid>https://lobste.rs/s/zcoz8h</guid><author> by mt</author><pubDate>Thu, 12 Mar 2026 14:19:34 -0500</pubDate><comments>https://lobste.rs/s/zcoz8h/how_do_you_manage_ssh_keys</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The article solicits user practices for managing SSH keys, covering aspects such as the number of keys employed (e.g., a single key for all uses, one per host, or per use case), preferred storage locations like the &lt;code&gt;~/.ssh&lt;/code&gt; directory or password managers, and usage methodologies like &lt;code&gt;ssh-agent&lt;/code&gt; or direct specification with the &lt;code&gt;-i&lt;/code&gt; flag. The author specifically invites sharing of any unconventional yet effective approaches to SSH key management.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion highlights several distinct strategies for SSH key management. Many users emphasize the use of hardware tokens, such as Yubikeys, frequently paired with &lt;code&gt;gpg-agent&lt;/code&gt; or leveraged as a certificate authority, as detailed by &lt;a href="https://lobste.rs/c/krqdwe"&gt;jamesog&lt;/a&gt; and &lt;a href="https://lobste.rs/c/f7e029"&gt;sarcasticadmin&lt;/a&gt;. Another common approach involves utilizing secure enclaves or macOS-specific applications like Secretive, discussed by &lt;a href="https://lobste.rs/c/x7p2tu"&gt;z0mbix&lt;/a&gt; and &lt;a href="https://lobste.rs/c/jihyuf"&gt;heckie&lt;/a&gt;. Password managers, including 1Password and Bitwarden, offering integrated SSH agent support, are also popular for key storage, noted by &lt;a href="https://lobste.rs/c/begbgd"&gt;gcollazo&lt;/a&gt;. Advanced users like &lt;a href="https://lobste.rs/c/r9yv5u"&gt;Foxboron&lt;/a&gt; explore short-lived SSH certificates, often involving TPMs or custom certificate authorities. Practical configuration techniques, such as those within &lt;code&gt;~/.ssh/config&lt;/code&gt; or &lt;code&gt;git config&lt;/code&gt; for managing multiple keys per host or context, are shared by &lt;a href="https://lobste.rs/c/m9p127"&gt;skobes&lt;/a&gt;. The convenience of Tailscale SSH is also mentioned by &lt;a href="https://lobste.rs/c/w0r033"&gt;Diti&lt;/a&gt;. A recurring theme across the comments is the crucial balance between ease of use and security, alongside the importance of defining one's threat model when selecting a management strategy, as underscored by &lt;a href="https://lobste.rs/c/g3q85r"&gt;vpr&lt;/a&gt; and &lt;a href="https://lobste.rs/c/d5q9z2"&gt;viraptor&lt;/a&gt;.&lt;/p&gt;</description><category>practices</category><category>ask</category></item><item><title>Plan 9's Acme: The Un-Terminal and Text-Based GUIs</title><link>https://www.danielmoch.com/posts/2025/01/acme/</link><guid>https://lobste.rs/s/x9znhg</guid><author>danielmoch.com via runxiyu</author><pubDate>Thu, 12 Mar 2026 22:54:26 -0500</pubDate><comments>https://lobste.rs/s/x9znhg/plan_9_s_acme_un_terminal_text_based_guis</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Daniel Moch explores Plan 9's Acme editor as an "un-terminal" and a prime example of a text-based graphical user interface (GUI). He argues that while many developers prefer terminal-based UIs, the true advantage lies in highly-regular, text-based interfaces with clear utility integration, rather than the terminal itself. Modern GUIs, by contrast, often create overwhelming user experiences due to each application being built from scratch without shared conventions. Acme's power comes from its seamless integration with command-line tools and its use of the 9P protocol for "plugins." Its simplicity—lacking extensive configuration, color themes, or syntax highlighting—is presented as a core strength that has allowed it to age gracefully.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters largely resonated with the article's critique of modern GUI complexity. &lt;a href="https://lobste.rs/c/xyajgf"&gt;kevinc&lt;/a&gt; agreed that building GUIs from scratch makes computers overwhelming, noting the loss of core GUI principles. &lt;a href="https://lobste.rs/c/kopcym"&gt;duncan_bayne&lt;/a&gt; recounted a moment in 2004 when a lack of commitment to UI standards for XAML foreshadowed this decline. &lt;a href="https://lobste.rs/c/p8wyct"&gt;pgeorgi&lt;/a&gt; discussed the tension between uniformity and branding, citing Google's Material Design evolution. While &lt;a href="https://lobste.rs/c/orj41e"&gt;bitshift&lt;/a&gt; appreciated Acme's powerful text-to-hyperlink integration demonstrated in a Russ Cox screencast, &lt;a href="https://lobste.rs/c/do5byj"&gt;mhd&lt;/a&gt; challenged the idea that terminal-based programs are immune to standardization issues, arguing that TUIs also vary widely and require guidelines for uniformity, similar to pixel-based GUIs.&lt;/p&gt;</description><category>editors</category></item><item><title>Parametricity, or Comptime is Bonkers</title><link>https://noelwelsh.com/posts/comptime-is-bonkers/</link><guid>https://lobste.rs/s/l3rkdl</guid><author>noelwelsh.com via soareschen</author><pubDate>Thu, 12 Mar 2026 01:48:59 -0500</pubDate><comments>https://lobste.rs/s/l3rkdl/parametricity_comptime_is_bonkers</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Noel Welsh explores parametricity, a programming language property where a generic function's type signature strictly dictates its implementation (e.g., in Rust, `fn mystery&lt;T&gt;(a: T) -&gt; T` must return `a`). This contrasts with Zig's `comptime`, which allows function bodies to inspect and branch on compile-time types, breaking parametricity and leading to varied, signature-unpredictable behavior. The author argues that parametricity enhances code comprehension and modularity, and while `comptime` offers powerful staging capabilities, it's a less suitable mechanism for generic programming compared to type classes or traits found in languages like Haskell and Rust.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion largely revolves around whether Zig's `comptime` inherently breaks parametricity or if it's the lack of higher-level abstractions like traits. &lt;a href="https://lobste.rs/c/apcxdg"&gt;dcreager&lt;/a&gt; (41 pts) suggests it's Zig's imperative type manipulation and absence of trait-like constructs that lead to the loss of parametricity, not `comptime` itself. &lt;a href="https://lobste.rs/c/u7ymcj"&gt;jamii&lt;/a&gt; argues against calling Zig's type system parametric due to its allowance of arbitrary code in functions that operate on types, enabling non-parametric behavior. Other comments touch upon whether `comptime` aligns with dependent type systems (&lt;a href="https://lobste.rs/c/twyitt"&gt;gf0&lt;/a&gt; believes it doesn't quite yet), comparisons to other languages like Odin (&lt;a href="https://lobste.rs/c/u1azdb"&gt;dfawcus&lt;/a&gt;), and wishes for combined `comptime` with traits (&lt;a href="https://lobste.rs/c/oizgnq"&gt;zesterer&lt;/a&gt;). &lt;a href="https://lobste.rs/c/grlg2s"&gt;invlpg&lt;/a&gt; (35 pts) provides a "facetious" Rust example demonstrating how one could force similar non-parametric behavior using unsafe operations.&lt;/p&gt;</description><category>rust</category><category>haskell</category><category>plt</category><category>zig</category></item><item><title>How to make your own static site generator</title><link>https://gaultier.github.io/blog/how_to_make_your_own_static_site_generator.html</link><guid>https://lobste.rs/s/qnnok5</guid><author>gaultier.github.io via abhin4v</author><pubDate>Thu, 12 Mar 2026 23:54:39 -0500</pubDate><comments>https://lobste.rs/s/qnnok5/how_make_your_own_static_site_generator</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Philippe Gaultier details the process and lessons learned from building his custom static site generator, which evolved from a simple Makefile to a highly optimized tool for his blog. Key aspects covered include efficiently listing articles using Git for timestamps, parsing content into an Abstract Syntax Tree (AST) for precise control over HTML generation and linting, and generating components like the table of contents, RSS feed, and home page. He also discusses client-side search implementation, an unconventional approach to drafts using Git branches, and integrating live reloading for a fast development feedback loop.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion delves into various aspects of static site generation. &lt;a href="https://lobste.rs/c/ontuw1"&gt;chrismorgan&lt;/a&gt; questions the use of Git for publication dates and UUIDs for feed IDs, suggesting frontmatter and URLs respectively, while advocating for Atom over RSS. &lt;a href="https://lobste.rs/c/dvcjvv"&gt;veqq&lt;/a&gt; explores architectural patterns for small sites, envisioning pipelines with middleware and contemplating more complex ideas like storing code in a database. &lt;a href="https://lobste.rs/c/nwj1tk"&gt;matklad&lt;/a&gt;, supported by &lt;a href="https://lobste.rs/c/udlayj"&gt;zimpenfish&lt;/a&gt; and &lt;a href="https://lobste.rs/c/gtl2ct"&gt;polywolf&lt;/a&gt;, argues that live-reloading should be an orthogonal feature provided by an external static file server rather than being built into the generator itself, highlighting tools like `live-server`.&lt;/p&gt;</description><category>programming</category></item><item><title>Dreaming of a ten-year computer</title><link>https://alexwlchan.net/2026/ten-year-computer/</link><guid>https://lobste.rs/s/eng6mr</guid><author>alexwlchan.net via msangi</author><pubDate>Thu, 12 Mar 2026 11:17:40 -0500</pubDate><comments>https://lobste.rs/s/eng6mr/dreaming_ten_year_computer</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Alexwlchan details their ambition to make their current Mac mini last a decade, driven by unchanging personal computing needs, environmental concerns, and supply chain anxieties. They describe strategies for longevity, including limiting background processes, disabling JavaScript, using static websites, and writing custom tools. The author also discusses the hardware choice – an M4 Pro Mac mini with expanded external storage – and the expectation of macOS software support covering most of the ten-year period, while acknowledging the role of the desktop form factor in device endurance.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters largely concur with the article's premise, sharing numerous examples of using computers for ten or more years, particularly with Linux. &lt;a href="https://lobste.rs/c/bp3tc3"&gt;jmiven&lt;/a&gt; notes this is common in the PC world, while &lt;a href="https://lobste.rs/c/wvuokt"&gt;nicoco&lt;/a&gt; finds it "extremely standard" for high-end machines outside of 4K gaming. &lt;a href="https://lobste.rs/c/0geioa"&gt;gnafuthegreat&lt;/a&gt; emphasizes proprietary software as the main threat to hardware longevity and extols open-source solutions. The discussion also touches on Moore's Law, with &lt;a href="https://lobste.rs/c/yfjvuh"&gt;veqq&lt;/a&gt; asserting it stopped for CPUs some time ago and that we've had 10-year computers for a decade, and &lt;a href="https://lobste.rs/c/gjfzg9"&gt;veqq&lt;/a&gt; further elaborates on the diminishing returns of increased compute power on daily user experience.&lt;/p&gt;</description><category>hardware</category></item><item><title>Guix System - One Month Later</title><link>https://nemin.hu/guix-one-month-later.html</link><guid>https://lobste.rs/s/koymkj</guid><author>nemin.hu via Kratacoa</author><pubDate>Thu, 12 Mar 2026 15:12:06 -0500</pubDate><comments>https://lobste.rs/s/koymkj/guix_system_one_month_later</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The author, after one month with Guix System, shares positive and negative experiences, ultimately deciding not to continue using it. Key positives include a welcoming and active IRC community and the "go-getter" attitude for packaging, allowing users to contribute readily, unlike Nix's maintainer model. However, the author found mailing lists daunting, experienced slow review times for package contributions, and criticized the debugging process due to unhelpful error messages related to missing parentheses, non-existent variables, and circular dependencies.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters discussed the author's hardware struggles with Icedove/Thunderbird dependencies. &lt;a href="https://lobste.rs/c/dplxgf"&gt;jaredkrinke&lt;/a&gt; questioned why a 12-core Ryzen CPU and 32 GB of RAM failed to build certain dependencies without crashing, suggesting a potential cooling issue. &lt;a href="https://lobste.rs/c/nrz6br"&gt;nemin&lt;/a&gt;, the author, responded that heat-related problems were unlikely. &lt;a href="https://lobste.rs/c/fgirho"&gt;untrusem&lt;/a&gt; thanked the author and expressed hope for their return to the Guix world.&lt;/p&gt;</description><category>lisp</category><category>nix</category></item><item><title>On Making</title><link>https://beej.us/blog/data/ai-making/</link><guid>https://lobste.rs/s/8xccbo</guid><author>beej.us via mysticmode</author><pubDate>Thu, 12 Mar 2026 13:29:15 -0500</pubDate><comments>https://lobste.rs/s/8xccbo/on_making</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; In "On Making," the author Beej explores the profound personal fulfillment derived from creating things oneself, contrasting it with the detachment felt when delegating tasks, particularly to AI. Acknowledging the "AI dev schism," he discusses the loss of craft and joy from low-level problem-solving. He illustrates his point by presenting various "creations"—a woodcut, a rebuilt deck, and roguelike code—only to reveal that most were AI-generated (except the deck, done by craftsmen), highlighting his discomfort with claiming credit for work not personally executed. The article questions the blurry line between using tools like compilers and asking an AI to "make" something, concluding that while prompting requires skill, it feels more like management than genuine making, diminishing the sense of pride from direct creation.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion largely resonates with the author's sentiment, with &lt;a href="https://lobste.rs/c/3fzbgz"&gt;zetashift&lt;/a&gt; expressing a similar lack of emotional connection to LLM-generated code and a disillusionment with the software industry. &lt;a href="https://lobste.rs/c/zjkphb"&gt;ocramz&lt;/a&gt; ponders if the perceived humanity of AI contributes to this feeling and suggests actively engaging with generated output by studying and modifying it. &lt;a href="https://lobste.rs/c/mpbjc2"&gt;nickmonad&lt;/a&gt; clarifies that the core issue is the reversal of the thought process; building code manually fosters deep understanding, whereas LLM-generated code often leads to a "distrusting investigation." &lt;a href="https://lobste.rs/c/ylljzb"&gt;sunflowerseastar&lt;/a&gt; elaborates on the "AI-powered hammer" analogy, distinguishing between beneficial assistance and skill-sapping dependency. &lt;a href="https://lobste.rs/c/pnqmnb"&gt;Johz&lt;/a&gt; further questions whether LLMs are power tools that assist a creator or automation that restricts output to predetermined forms.&lt;/p&gt;</description><category>vibecoding</category></item><item><title>My PostgreSQL database got nuked lol</title><link>https://akselmo.dev/posts/they-broke-my-server/</link><guid>https://lobste.rs/s/vb7ipx</guid><author>akselmo.dev by Aks</author><pubDate>Wed, 11 Mar 2026 13:06:57 -0500</pubDate><comments>https://lobste.rs/s/vb7ipx/my_postgresql_database_got_nuked_lol</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The author recounts how their PostgreSQL database, running in a Docker container for a personal project, was unexpectedly compromised twice. They discovered that Docker's default behavior exposed the database port publicly, and their VPS lacked a firewall, allowing attackers to exploit default `postgres:postgres` credentials to delete data and leave a ransom note. The issue was resolved by explicitly binding Docker ports to `127.0.0.1` and installing/configuring UFW to restrict public access to only necessary web ports, highlighting critical lessons in server security.
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; The discussion largely centers on Docker and firewall configurations. &lt;a href="https://lobste.rs/c/89bbj0"&gt;cyberia&lt;/a&gt; suggests Podman as an alternative that doesn't interfere with host firewalls, though &lt;a href="https://lobste.rs/c/prwkv7"&gt;yorickpeterse&lt;/a&gt; clarifies that Podman can still bypass `firewalld` rules with `--publish`, offering a detailed workaround. For internal services like databases, &lt;a href="https://lobste.rs/c/3aaddw"&gt;stephank&lt;/a&gt; recommends skipping port forwarding by connecting directly to container IPs, and &lt;a href="https://lobste.rs/c/1dckjw"&gt;valpackett&lt;/a&gt; proposes using Unix domain sockets or fd inheritance. Many users, including &lt;a href="https://lobste.rs/c/wmbdvt"&gt;hoistbypetard&lt;/a&gt; and &lt;a href="https://lobste.rs/c/r4reis"&gt;bitshift&lt;/a&gt;, empathize with the author's learning experience regarding Docker's surprising default network behavior. &lt;a href="https://lobste.rs/c/efbeny"&gt;technomancy&lt;/a&gt; draws a parallel to historical PostgreSQL configuration difficulties, noting how development-friendly defaults can become production liabilities.</description><category>security</category><category>web</category></item><item><title>How Many Times Can a DVD±RW Be Rewritten?</title><link>https://goughlui.com/2026/03/07/tested-how-many-times-can-a-dvd%c2%b1rw-be-rewritten-part-2-methodology-results/</link><guid>https://lobste.rs/s/jow6ea</guid><author>goughlui.com via susam</author><pubDate>Thu, 12 Mar 2026 22:09:23 -0500</pubDate><comments>https://lobste.rs/s/jow6ea/how_many_times_can_dvd_rw_be_rewritten</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; This article details the methodology and results of an experiment to determine how many times a DVD±RW disc can be rewritten. The author used `Opti Drive Control` for writing, verifying, transfer rate testing, and quality scanning, automating the repetitive process with a `pyautogui` script. To accelerate testing, a second `Lite-On iHAS120 6` drive was used in parallel. The article outlines the setup, the challenges encountered (like Windows updates), and the image processing techniques used to analyze the results, while also noting limitations such as the variability of quality scan results and the specificity of findings to the tested hardware.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters express surprise at the high rewrite counts, with &lt;a href="https://lobste.rs/c/abdftz"&gt;m_eiman&lt;/a&gt; admitting their initial guess of 5-10 cycles was significantly underestimated. &lt;a href="https://lobste.rs/c/vnaysq"&gt;zod000&lt;/a&gt; shares a similar sentiment, finding the extensive testing impressive but lamenting its timing as "15 years too late" for their own practical use.&lt;/p&gt;</description><category>hardware</category></item><item><title>Grief and the AI Split</title><link>https://blog.lmorchard.com/2026/03/11/grief-and-the-ai-split/</link><guid>https://lobste.rs/s/wssz9m</guid><author>blog.lmorchard.com via splitbrain</author><pubDate>Thu, 12 Mar 2026 00:17:51 -0500</pubDate><comments>https://lobste.rs/s/wssz9m/grief_ai_split</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The article explores the "AI split" among developers, revealing a fundamental division between those who value the craft of coding and those focused on achieving results. The author, who identifies with the latter, sees AI assistance as a natural progression in making computers "do things." He contrasts his own grief—stemming from changes in the web's ecosystem, the economy, and career landscape—with the grief of others who mourn the loss of the coding craft itself, arguing both perspectives are valid but his is more actionable.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters largely resonated with the idea of a developer split. &lt;a href="https://lobste.rs/c/buomun"&gt;fleebee&lt;/a&gt; and &lt;a href="https://lobste.rs/c/klfroi"&gt;chai&lt;/a&gt; expressed similar sentiments to the "craft-lovers," highlighting their satisfaction in understanding and molding code, and their discomfort with AI disconnecting them from that process, noting that many companies prioritize efficiency over craft. A discussion arose around the next generation's ability to gain experience: &lt;a href="https://lobste.rs/c/l5vhp7"&gt;darth-cheney&lt;/a&gt; questioned how they would acquire decades of judgment, while &lt;a href="https://lobste.rs/c/6rqjoe"&gt;simonw&lt;/a&gt; was optimistic, believing they will adapt and develop necessary skills. &lt;a href="https://lobste.rs/c/6o83ib"&gt;alexjurkiewicz&lt;/a&gt; and &lt;a href="https://lobste.rs/c/luweqy"&gt;viraptor&lt;/a&gt; drew parallels to the impact of calculators on math skills or the niche nature of assembly language, suggesting core coding skills might erode but not vanish.&lt;/p&gt;</description><category>vibecoding</category></item><item>
<title>Lobsters Interview with ngoldbaum</title>
<link>https://alexalejandre.com/programming/interview-with-ngoldbaum/</link>
<guid>https://lobste.rs/s/6lqnhh</guid>
<author>alexalejandre.com by veqq</author>
<pubDate>Wed, 11 Mar 2026 17:16:55 -0500</pubDate>
<comments>https://lobste.rs/s/6lqnhh/lobsters_interview_with_ngoldbaum</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Alex Alejandre interviews Nathan Goldbaum, who is instrumental in enabling free-threading for Python. Goldbaum discusses his path from astrophysics to open-source Python development (NumPy, PyO3), driven by the satisfaction of resolving specific technical problems compared to the open-ended nature of academic research. He advocates for open science, reproducible workflows, and community collaboration, noting how Python's ecosystem is enhancing interoperability through standardization like the Array API Standard.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters applaud Goldbaum's work and his advocacy for open science, relating to the academic "secret sauce" issue in various fields (&lt;a href="https://lobste.rs/c/ptwckx"&gt;rebeca&lt;/a&gt;, &lt;a href="https://lobste.rs/c/mocyzx"&gt;nicoco&lt;/a&gt;). One commenter notes the increasing norm of open data/code in social sciences (&lt;a href="https://lobste.rs/c/wem8vo"&gt;cjr&lt;/a&gt;). Praise is given for Goldbaum's talent on the FT team at Quansight (&lt;a href="https://lobste.rs/c/smbgu9"&gt;kenjin&lt;/a&gt;). The discussion also touches upon Mercurial's continued relevance for large monorepos, as detailed in a FOSDEM talk (&lt;a href="https://lobste.rs/c/jxjnrl"&gt;mk12&lt;/a&gt;). Minor formatting and typo suggestions for the interview text are also made (&lt;a href="https://lobste.rs/c/xwzqwa"&gt;roryokane&lt;/a&gt;, &lt;a href="https://lobste.rs/c/wekqsx"&gt;veqq&lt;/a&gt;), and one user reacts to a comment about GitHub IP blocking (&lt;a href="https://lobste.rs/c/jewbjk"&gt;fedemp&lt;/a&gt;).&lt;/p&gt;</description>
<category>person</category>
<category>interview</category>
</item>
<item><title>agent-shell 0.47 updates</title><link>https://xenodium.com/agent-shell-0-47-1-updates</link><guid>https://lobste.rs/s/lle7ol</guid><pubDate>Thu, 12 Mar 2026 12:33:35 -0500</pubDate><comments>https://lobste.rs/s/lle7ol/agent_shell_0_47_updates</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; agent-shell 0.47 updates xenodium.com ██ ██ ███████ ███ ██ ██████ ██████ ██ ██ ██ ██ ███ ███ ██ ██ ██ ████ ██ ██ ██ ██ ██ ██ ██ ██ ████ ████ ███ █████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ████ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ██ ███████ ██ ████ ██████ ██████ ██ ██████ ██ ██ March 12, 2026 agent-shell 0.47 updates We got quite a few agent-shell additions since the last post , so let's go through the highlights as of v0.47.1 . What's agent-shell ? Agent shell is a native Emacs mode to interact with LLM agents powered by ACP ( Agent Client Protocol ). Your employer can make a difference agent-shell has been attracting quite a few users. Many of you are working in tech where employers are happily paying for IDE subscriptions and LLM tokens to improve productivity. If you are using agent-shell for work, consider getting your employer to give back by sponsoring the project. I also know many of you work at AI companies offering paid agents like Claude Code, Copilot, Gemini, Codex, etc. all supported by agent-shell . Please nudge your employers to help fund projects like agent-shell , which are making their services available to more users. ✨ Sponsor agent-shell ✨ So what's new? claude-code-acp renamed to claude-agent-acp [Action Required] Let's get this one out of the way, as it needs actioning. Both the npm package and the CLI agent have been renamed from claude-code-acp to claude-agent-acp (to align with Anthropic's branding guidelines ). If you're using Claude Code, you'll need to update: npm remove -g @zed-industries/claude-code-acp npm install -g @zed-industries/claude-agent-acp If you had customized agent-shell-anthropic-claude-acp-command , update it to point to claude-agent-acp . New agents supported Auggie ( #179 by @Pacane ). Cline. Factory Droid ( #178 by @ag91 ). GitHub Copilot. Kiro CLI ( #351 by @zmjones ). Mistral Vibe. Pi ( #232 by @conornash ) Bootstrapped sessions (experimental) This was a biggie. How sessions are loaded is now configurable via agent-shell-session-strategy . When set to 'new , starting a new shell delivers a fully bootstrapped session before presenting you with the shell prompt. This means the ACP handshake, authentication, and session creation all happen upfront. You can enable this flow with: (setq agent-shell-session-strategy 'new) What's the benefit? Bootstrapped sessions enable changing models and session modes (Planning, Don't ask, Skip permissions, etc…) before submitting your first prompt. For the time being, the existing (deferred) behaviour is still offered via 'new-deferred . Just set as follows: (setq agent-shell-session-strategy 'new-deferred) Session resume (experimental) Probably the most requested feature and also facilitated by the bootstrapping changes. agent-shell-session-strategy also unlocks session resume. Set it to 'prompt and every time either M-x agent-shell-new-shell or C-u M-x agent-shell are invoked, you'll be offered to resume previous sessions or start a new one. (setq agent-shell-session-strategy 'prompt) Alternatively, you can set to 'latest to always resume the most recent session in current project. Under the hood, there are two ways to pick up from previous session: session/resume (lightweight, no message replay) and session/load (full history replay). By default, agent-shell prefers resuming (controlled by agent-shell-prefer-session-resume ). Please favor resuming for the time being as loading has more edge cases to sort out still. Note: Both resuming and loading sessions are agent-dependent. Some agents may not yet support either, especially as the features aren't yet considered stable in Agent Client Protocol (see session/list spec). This feature was a collaboration between @farra , @travisjeffery , and myself. Clipboard images You can now use agent-shell-send-clipboard-image ( #285 by @dangom ) to send images straight from your clipboard into agent-shell . Clipboard images are saved to .agent-shell/screenshots in your project root and inserted into the shell buffer as context. Note: You'll need either pngpaste or xclip installed on your system for the feature to automatically kick in. In addition, we now have agent-shell-yank-dwim : if the clipboard has an image, it pastes it as context. Otherwise, it yanks text as usual. In other words, copy an image anywhere to your system's clipboard and paste/yank into the buffer as usual (typically via C-y ). Status display + tool calls Status labels and tool call titles rendering got some improvements. Status reporting is generally more compact, redundant text is dropped from tool call titles, and tool status/kind shortening has been consolidated. Image rendering agent-shell now renders images inline. When agents output images (charts, diagrams, screenshots, etc.), they display directly in the shell buffer. You may need to nudge the agent to output image paths in the expected format so agent-shell can pick up. Markdown images: ![alt text](/path/to/image.png) Any of the following in a line of their own are supported also: /path/to/image.png file:///path/to/image.png ./output/chart.png ~/screenshots/demo.png Recognized image formats depend on what your Emacs was built with (typically png, jpeg, gif, svg, webp, tiff, etc. via image-file-name-extensions ). Emacs skills While on the topic of image rendering, this works particularly well when coupled with charting agent skills . I shared some of these over at emacs-skills , demoed in episode 13 of the Bending Emacs series. Table rendering Tables are now rendered using overlays ( #17 by @ewilderj ). Usage tracking Tracking usage now possible ( #270 by @Lenbok ): A color-coded context usage indicator in the header (green -&gt; yellow -&gt; red as context fills up), enabled by default via agent-shell-show-context-usage-indicator . M-x agent-shell-show-usage to check token counts, context window usage, and cost in the minibuffer.r- An optional end-of-turn usage summary can be enabled via (setq agent-shell-show-usage-at-turn-end t) . G&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description></item><item><title>Games in PostScript - Play Chess Against Your Printer</title><link>https://www.youtube.com/watch?v=2YCATafErA8</link><guid>https://lobste.rs/s/efhaws</guid><pubDate>Thu, 12 Mar 2026 18:30:25 -0500</pubDate><comments>https://lobste.rs/s/efhaws/games_postscript_play_chess_against_your</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; A YouTube video from GambiConf 2025 demonstrates the concept of playing games, specifically chess, against a printer using PostScript implementations.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; mdaniel points to the speaker's GitHub repository for PSChess, a chess engine written in PostScript (&lt;a href="https://lobste.rs/c/2hfiwr"&gt;comments&lt;/a&gt;), and also highlights the speaker's article on the complexities of parsing JSON (&lt;a href="https://lobste.rs/c/hg4uvk"&gt;comments&lt;/a&gt;).&lt;/p&gt;</description></item><item><title>The Best Darn Grid Shader (Yet)</title><link>https://bgolus.medium.com/the-best-darn-grid-shader-yet-727f9278b9d8</link><guid>https://lobste.rs/s/jn7yoi</guid><pubDate>Fri, 13 Mar 2026 01:37:56 -0500</pubDate><comments>https://lobste.rs/s/jn7yoi/best_darn_grid_shader_yet</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The article delves into the development and features of "The Best Darn Grid Shader (Yet)".&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description></item><item>
<title>Lowdown Manpage Support</title>
<link>https://kristaps.bsd.lv/lowdown/mdoc.html</link>
<guid>https://lobste.rs/s/rz6gak</guid>
<author>kristaps.bsd.lv via dzwdz</author>
<pubDate>Thu, 12 Mar 2026 04:31:01 -0500</pubDate>
<comments>https://lobste.rs/s/rz6gak/lowdown_manpage_support</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Lowdown Manpage Support simplifies the creation of Unix manpages by allowing authors to write them in Markdown and convert them to `mdoc` or `man` format, thereby circumventing the complexities of traditional manpage languages.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters discuss various alternatives and approaches to manpage generation. &lt;a href="https://lobste.rs/c/bcxwmh"&gt;acatton&lt;/a&gt; mentions `scdoc`, but &lt;a href="https://lobste.rs/c/xk3jih"&gt;dzwdz&lt;/a&gt; points out that `scdoc` lacks the semantic features of `mdoc` that Lowdown offers. &lt;a href="https://lobste.rs/c/nnbjmg"&gt;dzwdz&lt;/a&gt; also notes that Lowdown 3.0.0, released last week, is from an author on the `mandoc` team, suggesting its quality. &lt;a href="https://lobste.rs/c/exlfpt"&gt;telemachus&lt;/a&gt; appreciates tools that lead to more manpages, recalling `ronn` as a similar Markdown-to-manpage tool. &lt;a href="https://lobste.rs/c/f1b6k7"&gt;fanf&lt;/a&gt; highlights Lowdown's understanding of manpage conventions compared to other generators, while &lt;a href="https://lobste.rs/c/9nmh0e"&gt;pyrmont&lt;/a&gt;, creator of Predoc, expresses hope that Lowdown's support encourages more `mdoc`-based manpages. Other tools mentioned include `go-md2man` by &lt;a href="https://lobste.rs/c/odzpr4"&gt;samuelkarp&lt;/a&gt;, AsciiDoc with its `manpage` doctype by &lt;a href="https://lobste.rs/c/uydoy7"&gt;Diti&lt;/a&gt;, and `pandoc` by &lt;a href="https://lobste.rs/c/2wu9co"&gt;jmc&lt;/a&gt;.&lt;/p&gt;</description>
<category>unix</category>
</item>
<item><title>oss-security - Re: Multiple vulnerabilities in AppArmor</title><link>https://www.openwall.com/lists/oss-security/2026/03/12/7</link><guid>https://lobste.rs/s/2oob2h</guid><pubDate>Thu, 12 Mar 2026 17:29:41 -0500</pubDate><comments>https://lobste.rs/s/2oob2h/oss_security_re_multiple</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; A Qualys Security Advisory, titled "CrackArmor," reveals multiple vulnerabilities in AppArmor, a Linux Security Module. These include a "confused-deputy" problem enabling unprivileged attackers to manipulate AppArmor profiles, potentially weakening system defenses, causing denial-of-service, or bypassing user-namespace restrictions. The advisory also details various Local Privilege Escalations (LPEs) through user-space and kernel-space vulnerabilities like uncontrolled recursion, out-of-bounds reads, use-after-free, and double-free issues.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description></item><item>
<title>Building a new Flash</title>
<link>https://bill.newgrounds.com/news/post/1607118</link>
<guid>https://lobste.rs/s/wxxxuc</guid>
<author>bill.newgrounds.com via edwardloveall</author>
<pubDate>Wed, 11 Mar 2026 21:42:35 -0500</pubDate>
<comments>https://lobste.rs/s/wxxxuc/building_new_flash</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; NG Guard&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; &lt;ul&gt;
&lt;li&gt;Internet_Janitor: A new Flash authoring tool would be _fantastic_, but I'd be wary of this. The author is promising an _incredibly_ ambitious and broad collection of functionality implemented _all at once_, offering a few tantalizing static screenshots of the tool, assuring folks it will be open-source when it's done, and asking for donations. The readme is packed with em-dashes and even a distressingly familiar sentence structure: Not just play them back — edit them. This is based on slop, and beyond that probably vaporware. &lt;a href="https://lobste.rs/c/94m2yl"&gt;https://lobste.rs/c/94m2yl&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;Fingel: Yea, "This isn’t a proof of concept or a weekend project. It’s a real authoring environment. Here’s where things stand" is where the record came to a screeching halt for me. I hate that this is happening now. Especially for a project like this that tugs at people's heartstrings. People harbor a lot of feelings for Flash and that era of computing in general, many of them positive. &lt;a href="https://lobste.rs/c/9kmjvd"&gt;https://lobste.rs/c/9kmjvd&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;viraptor: I'd love it if the author actually did it. I'm now worried by the ai gen part. Accidentally I've been doing 3 big chunks of this project separately (reverse engineering + rebuilding a compiler, making a vector editor, making an interactive animation thing). This is now completely doable with some care and the challenge will be in actually giving all of it a coherent UX people want to use (and not running out of money for tokens I guess...) Of course it could be a scam and I wouldn't pay before some early, functioning release is shown. &lt;a href="https://lobste.rs/c/d56jwp"&gt;https://lobste.rs/c/d56jwp&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;hwayne: Who is this person, and why should we trust them to deliver? &lt;a href="https://lobste.rs/c/fhhqcu"&gt;https://lobste.rs/c/fhhqcu&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;trenchant: Looks awesome, doesn't seem to be a way to try it out. I was hoping the next Flash would run in the browser :-) &lt;a href="https://lobste.rs/c/lvvvww"&gt;https://lobste.rs/c/lvvvww&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;calvin: I'm kind of surprised no one capitalized on the decline of Flash, especially now that it's discontinued without any successor from Adobe or otherwise. Both in the niches of 2D animation tools (it seems like Toonboom/Harmony are the last ones left?), and in rich site authoring tools (for web-based RAD and flashy animated sites; in the HTML5 era it seems you have to get devs and designers to do it by hand). &lt;a href="https://lobste.rs/c/vemglp"&gt;https://lobste.rs/c/vemglp&lt;/a&gt;&lt;/li&gt;
&lt;li&gt;viraptor: It's more enterprisey, but for authoring/animation there's &lt;a href="https://rive.app/"&gt;https://rive.app/&lt;/a&gt; Nowhere near the ease of access/use of flash though. It's not something a generation of kids can pirate and create a culture around. &lt;a href="https://lobste.rs/c/vulmh8"&gt;https://lobste.rs/c/vulmh8&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;&lt;/p&gt;</description>
<category>retrocomputing</category>
</item>
<item>
<title>ArcaOS 5.1.2 now available (OS/2 Warp)</title>
<link>https://www.arcanoae.com/arcaos-5-1-2-now-available/</link>
<guid>https://lobste.rs/s/fzgcjh</guid>
<author>arcanoae.com via classichasclass</author>
<pubDate>Thu, 12 Mar 2026 11:28:03 -0500</pubDate>
<comments>https://lobste.rs/s/fzgcjh/arcaos_5_1_2_now_available_os_2_warp</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Arca Noae has released ArcaOS 5.1.2, improving support for installation on modern UEFI-based systems and GPT-based disk layouts. This update is available free of charge for users with active ArcaOS 5.1 Support &amp; Maintenance subscriptions and can be used for new installations or to upgrade prior English versions of ArcaOS 5.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description>
<category>osdev</category>
</item>
<item><title>Blog: Mitigating URL-based Exfiltration in Gemini</title><link>https://bughunters.google.com/blog/mitigating-url-based-exfiltration-in-gemini</link><guid>https://lobste.rs/s/cbryfr</guid><pubDate>Fri, 13 Mar 2026 01:01:40 -0500</pubDate><comments>https://lobste.rs/s/cbryfr/blog_mitigating_url_based_exfiltration</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; This blog post focuses on strategies and implementations for mitigating URL-based exfiltration risks specifically in Gemini.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description></item><item>
<title>A fully snapshotable Wasm interpreter</title>
<link>https://github.com/friendlymatthew/gabagool?tab=readme-ov-file#gabagool</link>
<guid>https://lobste.rs/s/eu5uiz</guid>
<author>github.com by matthewkim</author>
<pubDate>Wed, 11 Mar 2026 08:59:37 -0500</pubDate>
<comments>https://lobste.rs/s/eu5uiz/fully_snapshotable_wasm_interpreter</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Gabagool is a WebAssembly interpreter written from scratch, aiming to be fully spec-compliant and performant with the ability to serialize, suspend, and restore its entire execution state. It also features a time-travel debugger and is tested against the WebAssembly spec test suite, passing 96% of tests related to arithmetic, control flow, memory, tables, globals, function references, imports/exports, and exceptions.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters discussed &lt;a href="https://lobste.rs/c/l0d4gt"&gt;time travel debugging&lt;/a&gt; and its implementation, with &lt;a href="https://lobste.rs/c/v2kae3"&gt;matthewkim confirming&lt;/a&gt; that the snapshot/restore system naturally extends to TTD. &lt;a href="https://lobste.rs/c/gcos1p"&gt;jaredkrinke questioned&lt;/a&gt; whether it keeps a log of operations or only travels back to checkpoints. &lt;a href="https://lobste.rs/c/tp5oat"&gt;matthewkim clarified&lt;/a&gt; that the current idea is to snapshot at every instruction, potentially leading to high memory usage, and sought research on TTD strategies. &lt;a href="https://lobste.rs/c/32eeva"&gt;cmcaine suggested&lt;/a&gt; recording IO and taking snapshots every N instructions. &lt;a href="https://lobste.rs/c/qgozhd"&gt;shanemhansen brainstormed&lt;/a&gt; log-based or persistent data structures for time travel. &lt;a href="https://lobste.rs/c/9qcosb"&gt;jaredkrinke shared&lt;/a&gt; a paper related to Microsoft's x86 time travel debugger. &lt;a href="https://lobste.rs/c/cf233t"&gt;Corbin recommended&lt;/a&gt; checking out "decaying histories" for long-lived programs. &lt;a href="https://lobste.rs/c/nko0fu"&gt;matthewkim noted&lt;/a&gt; he started implementing a debugger using the exponential decay buffer. &lt;a href="https://lobste.rs/c/mnlw3f"&gt;apropos found&lt;/a&gt; the project cool and mentioned tikzjax/web2js as a "terrifying level of jank" for snapshotting Wasm execution for LaTeX compilation. &lt;a href="https://lobste.rs/c/h1wqgb"&gt;gignico asked&lt;/a&gt; if it's similar to TeX's "formats," which &lt;a href="https://lobste.rs/c/nnaaly"&gt;apropos confirmed&lt;/a&gt; the author was aware of, and &lt;a href="https://lobste.rs/c/sxdxy5"&gt;gignico explained&lt;/a&gt; formats as memory dumps of TeX after loading macros. &lt;a href="https://lobste.rs/c/mdo9tj"&gt;amw-zero made&lt;/a&gt; a humorous reference to "gabagool."&lt;/p&gt;</description>
<category>rust</category>
<category>wasm</category>
</item>
<item>
<title>Generative AI vegetarianism</title>
<link>https://sboots.ca/2026/03/11/generative-ai-vegetarianism/</link>
<guid>https://lobste.rs/s/g3qpeu</guid>
<author>sboots.ca via carlana</author>
<pubDate>Wed, 11 Mar 2026 19:35:43 -0500</pubDate>
<comments>https://lobste.rs/s/g3qpeu/generative_ai_vegetarianism</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The author identifies as a "generative AI vegetarian," choosing to avoid generative AI tools and content due to concerns about bias, the degradation of critical thinking, and the tools' tendency to produce predictable rather than insightful results. They explain their reasons, drawing a parallel to dietary vegetarianism, and advocate for supporting organizations that deliberately avoid generative AI.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters debate the effectiveness and appropriateness of the "generative AI vegetarianism" analogy. &lt;a href="https://lobste.rs/c/dul35v"&gt;sarah-quinones&lt;/a&gt; points out the analogy's shortcomings regarding "purity" and the difference between personal choice and political conflict. &lt;a href="https://lobste.rs/c/qajw0q"&gt;hongminhee&lt;/a&gt; criticizes framing structural harms as personal lifestyle choices, advocating for political intervention over individual abstention, though later conceding that some organization is necessary (&lt;a href="https://lobste.rs/c/a4bqjc"&gt;https://lobste.rs/c/a4bqjc&lt;/a&gt;). &lt;a href="https://lobste.rs/c/jplygy"&gt;boramalper&lt;/a&gt; emphasizes the need for organized political force even with mass abstention. &lt;a href="https://lobste.rs/c/f3lwhy"&gt;lproven&lt;/a&gt; defends individual abstention as a matter of clear conscience, while &lt;a href="https://lobste.rs/c/odsjgw"&gt;hongminhee&lt;/a&gt; counters that a clear conscience shouldn't substitute for structural change. &lt;a href="https://lobste.rs/c/ijcdsq"&gt;jmk&lt;/a&gt; compares "AI-free" labels to organic food, which &lt;a href="https://lobste.rs/c/kvomuq"&gt;hongminhee&lt;/a&gt; uses to illustrate the limitation of niche markets in achieving systemic change. &lt;a href="https://lobste.rs/c/3fhxdk"&gt;lproven&lt;/a&gt; reiterates their commitment to personal abstention as a line in the sand against "mechanically-recovered words."&lt;/p&gt;</description>
<category>vibecoding</category>
</item>
<item>
<title>On The Need For Understanding</title>
<link>https://blog.information-superhighway.net/on-the-need-for-understanding</link>
<guid>https://lobste.rs/s/wxk0ka</guid>
<author>blog.information-superhighway.net via achyudh</author>
<pubDate>Wed, 11 Mar 2026 16:25:23 -0500</pubDate>
<comments>https://lobste.rs/s/wxk0ka/on_need_for_understanding</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The author reflects on Gerald Sussman's observation that modern programming often requires "basic science" to understand foreign libraries, contrasting it with a past where systems were built from entirely understood parts. Through personal anecdotes from their programming journey, they argue that while tools have evolved, the need for deep understanding remains crucial, and often involves delving into the source code of dependencies.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; Commenters largely agree with the article's premise, emphasizing the importance of curiosity and deep understanding in software development. &lt;a href="https://lobste.rs/c/0pgi0e"&gt;daniel_alp&lt;/a&gt; shares an experience where painstaking debugging led to a profound understanding of their code, and expresses concern that LLMs might discourage such deep dives, leading to a loss of critical thinking and creativity. &lt;a href="https://lobste.rs/c/vjlapr"&gt;ashwinsundar&lt;/a&gt; highlights the joy of the "hard work and process" of programming, including encountering and systematically solving bugs. &lt;a href="https://lobste.rs/c/tam6jq"&gt;hcs&lt;/a&gt; notes that enjoying such work is a valuable trait for programmers. &lt;a href="https://lobste.rs/c/ba1lkd"&gt;technomancy&lt;/a&gt; suggests that while LLMs might automate "producing lines of code," the "thinking about problems and how they could be solved by code" remains a critical, human-centric career. &lt;a href="https://lobste.rs/c/g1yeuw"&gt;krig&lt;/a&gt; and &lt;a href="https://lobste.rs/c/rlumwn"&gt;tsion&lt;/a&gt; both affirm that delving into the source code of dependencies is an "unreasonably effective skill" and a high-value behavior. &lt;a href="https://lobste.rs/c/lqopsp"&gt;benj&lt;/a&gt; distinguishes between "enterprise" software, where deep understanding might be discouraged, and software as a solution to a problem, where it is essential.&lt;/p&gt;</description>
<category>practices</category>
<category>programming</category>
<category>education</category>
</item>
<item>
<title>Announcing Mercurial sprint in London, May 27-29th</title>
<link>https://mercurial-scm.org/news/2026/0001-london-sprint</link>
<guid>https://lobste.rs/s/z5qqob</guid>
<author>mercurial-scm.org via mk12</author>
<pubDate>Thu, 12 Mar 2026 06:05:29 -0500</pubDate>
<comments>https://lobste.rs/s/z5qqob/announcing_mercurial_sprint_london_may</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; We are happy to announce that we will be holding a Mercurial sprint in London, UK from May 27th to 29th 2026.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; &lt;a href="https://lobste.rs/c/7btyv2"&gt;sebastien: I used Mercurial for many years, and loved both the CLI experience and the customizability. I switched to Git when ButBucket stopped Mercurial support and I discovered that hg-git sometimes would just stop working on a repository. It felt like a step down (was missing hg histedit, hg absorb), but now with jj I fell that I have the next generation of Mercurial with no downside.&lt;/a&gt; &lt;a href="https://lobste.rs/c/ztl4as"&gt;dkl: I'm curious what distinguishes Mercurial from Git in 2026 that might explain continued efforts like this. I quit fighting the tide at least a decade ago, maybe a little more, but I'm curious if anyone can put into words what the "killer apps" or whatever are that are keeping the momentum going here.&lt;/a&gt; &lt;a href="https://lobste.rs/c/fhmxtx"&gt;Sietsebb: For me, Mercurial + hg-git + the Evolve extension is still the nicest way to work with Git repos, and I include several months of Jujutsu experience in that comparison. That Mercurial has a better UI than Git is news that is two decades old by now. It edges out jj for me because its help is more polished. For example, `hg help` groups commands by topic (making commits, looking around, interacting with upstream, etc); whereas `jj help` sorts `jj bookmark` next to `jj bisect`. Or, compare [hg help log](https://repo.mercurial-scm.org/hg/help/log) with [jj help log](https://docs.jj-vcs.dev/latest/cli-reference/#jj-log). Or try finding out why some entries in `jj log` contain a `??` marker. (I know the answer now. It was not easx to find.) I want to give the jj help a copy-editing pass, but it'll be a few months befare I get around to it. ---- Want to try Mercurial as a Git client yourself? These installation instructions will get you set up and running! Tried and tested 2y ago; can't retest now, am an mabile and Termux lacks pipx. If an error has crept in let me know and I'll help you. ``` # build-requirements for debian sudo apt install pipx python3-dev gcc # Latest Mercurial, pre-release version of hg-git pipx install mercurial==6.5.3 pipx inject mercurial hg-evolve dulwich hg clone https://foss.heptapod.net/mercurial/hg-git ~/.local/hg-git # Setup hgrc cat &lt;&lt;EOF &gt;&gt; ~/.hgrc [ui] username = $(whoami) # Your name goes here tweakdefaults = True [extensions] evolve= hggit=~/.local/hg-git/hggit EOF ``` Now try it out: ``` hg clone git+https://github.com/nolar/looptime # random cool repo with short history cd looptime hg log --graph ```&lt;/a&gt; &lt;a href="https://lobste.rs/c/gynpqd"&gt;carlomonte: mercurial still has a clean, direct interface. without a quality alternative as benchmark, git would probably be worse. last but not least, "hg serve". i am relly happy that things keep going. i also hope that this hackathon is about working on mercurial on mercurial terms, and not chasing features from git.&lt;/a&gt;&lt;/p&gt;</description>
<category>event</category>
<category>vcs</category>
</item>
<item><title>MALUS - Clean Room as a Service</title><link>https://malus.sh/</link><guid>https://lobste.rs/s/wxfznm</guid><pubDate>Thu, 12 Mar 2026 11:03:54 -0500</pubDate><comments>https://lobste.rs/s/wxfznm/malus_clean_room_as_service</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; MALUS offers a "Clean Room as a Service" using AI robots to recreate open-source projects from scratch, aiming to liberate companies from license obligations, attribution, and copyleft. This service provides legally distinct code with corporate-friendly licensing, ensuring zero exposure to original source code and full indemnification.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; &lt;a href="https://lobste.rs/c/fgyvy2"&gt;altano&lt;/a&gt;: "The further in I got the less sure I was that this was satire. Weird times."&lt;/p&gt;</description></item><item>
<title>Secure Communication, Buried In A News App</title>
<link>https://hackaday.com/2026/03/09/secure-communication-buried-in-a-news-app/</link>
<guid>https://lobste.rs/s/d2skzd</guid>
<author>hackaday.com via ajdecon</author>
<pubDate>Wed, 11 Mar 2026 18:54:29 -0500</pubDate>
<comments>https://lobste.rs/s/d2skzd/secure_communication_buried_news_app</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; The Guardian, in collaboration with the University of Cambridge, has developed CoverDrop, a secure communication system embedded within its news app. This system aims to provide deniable secure messaging for journalists and sources by sending constant, seemingly meaningless encrypted data alongside actual messages, making it difficult for external observers to detect covert communication. The goal is to offer a higher level of protection and deniability than traditional encrypted messaging apps, which can draw unwanted attention to users. The codebase is available on GitHub.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; &lt;a href="https://lobste.rs/c/ptkrvv"&gt;fanf&lt;/a&gt; sees CoverDrop as an "Interesting successor to SecureDrop," while &lt;a href="https://lobste.rs/c/9egknd"&gt;legoktm&lt;/a&gt; (who works on SecureDrop) clarifies it's "More like an alternative/complement to SecureDrop rather than a successor," covering different threat models. They also note that SecureDrop is still active and used by The Guardian.&lt;/p&gt;</description>
<category>cryptography</category>
</item>
<item><title>Learn Haskell in Two Weeks</title><link>https://vitez.me/learn-haskell-in-two-weeks</link><guid>https://lobste.rs/s/gpwpji</guid><author>vitez.me via abhin4v</author><pubDate>Thu, 12 Mar 2026 23:12:55 -0500</pubDate><comments>https://lobste.rs/s/gpwpji/learn_haskell_two_weeks</comments><description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; Mitchell Vitez describes Mercury's "Learn Haskell by Exercises" (LHbE) program, an intensive two-week, exercise-driven training for new hires to quickly become proficient in Haskell for web development. The program emphasizes active practice, daily one-on-one mentorship, and a structured curriculum covering topics from basic types to monad transformers. It avoids passive learning methods like books and lectures in favor of hands-on exercises, guided struggle, and layered feedback to foster effective mental model development.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description><category>practices</category><category>haskell</category></item><item>
<title>Claude Code isn’t going to replace data engineers (yet)</title>
<link>https://rmoff.net/2026/03/11/claude-code-isnt-going-to-replace-data-engineers-yet/</link>
<guid>https://lobste.rs/s/mcwe1g</guid>
<author>rmoff.net by rmoff</author>
<pubDate>Thu, 12 Mar 2026 04:25:16 -0500</pubDate>
<comments>https://lobste.rs/s/mcwe1g/claude_code_isn_t_going_replace_data</comments>
<description>&lt;p&gt;&lt;strong&gt;The Gist:&lt;/strong&gt; This article investigates whether AI, specifically Claude Code, can replace data engineers. The author tested Claude's ability to build a dbt (data build tool) project using DuckDB for UK Environment Agency flood monitoring data, providing a detailed prompt with requirements like proper staging, handling data quality issues, SCD type 2 snapshots, historical backfill, documentation, tests, and source freshness checks. Claude successfully generated a working dbt project, demonstrating its capability in generating complex data engineering code, though the article implies further discussion on the nuances of its performance.&lt;/p&gt;&lt;p&gt;&lt;strong&gt;The Lobsters Take:&lt;/strong&gt; No comments yet.&lt;/p&gt;</description>
<category>vibecoding</category>
</item>
<item>
<title>Faster asin() Was Hiding In Plain Sight</title>
<link>https://16bpp.net/blog/post/faster-asin-was-hiding-in-plain-sight/</link>
<guid>https://lobste.rs/s/bunmdv</guid>
<author>16bpp.net via knl</author>
<pubDate>Wed, 11 Mar 2026 11:00:20 -0500</pubDate>
<comments>https://lobste.rs/s/bunmdv/faster_asin_was_hiding_plain_sight</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/bunmdv/faster_asin_was_hiding_plain_sight"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>math</category>
<category>performance</category>
</item>
<item>
<title>//go:fix inline and the source-level inliner</title>
<link>https://go.dev/blog/inliner</link>
<guid>https://lobste.rs/s/vjsm2q</guid>
<author>go.dev via cgrinds</author>
<pubDate>Wed, 11 Mar 2026 13:04:41 -0500</pubDate>
<comments>https://lobste.rs/s/vjsm2q/go_fix_inline_source_level_inliner</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/vjsm2q/go_fix_inline_source_level_inliner"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>go</category>
</item>
<item>
<title>Why should we have user age tracking in Operating Systems and websites?</title>
<link>https://wiki.alcidesfonseca.com/blog/age-verification-in-operating-systems-and-the-internet/</link>
<guid>https://lobste.rs/s/oyb3u8</guid>
<author>wiki.alcidesfonseca.com by alcides</author>
<pubDate>Wed, 11 Mar 2026 05:58:59 -0500</pubDate>
<comments>https://lobste.rs/s/oyb3u8/why_should_we_have_user_age_tracking</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/oyb3u8/why_should_we_have_user_age_tracking"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>law</category>
<category>osdev</category>
</item>
<item>
<title>Okmain: you have an image but you want a colour</title>
<link>https://dgroshev.com/blog/okmain/</link>
<guid>https://lobste.rs/s/t43mh5</guid>
<author>dgroshev.com by dangroshev</author>
<pubDate>Wed, 11 Mar 2026 11:20:45 -0500</pubDate>
<comments>https://lobste.rs/s/t43mh5/okmain_you_have_image_you_want_colour</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/t43mh5/okmain_you_have_image_you_want_colour"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>python</category>
<category>design</category>
<category>rust</category>
<category>vibecoding</category>
</item>
<item>
<title>Moonforge, A Yocto-Based Linux OS</title>
<link>https://www.igalia.com/2026/03/09/Introducing-Moonforge-A-Yocto-Based-Linux-OS.html</link>
<guid>https://lobste.rs/s/tyeo20</guid>
<author>igalia.com via lemon</author>
<pubDate>Wed, 11 Mar 2026 09:06:36 -0500</pubDate>
<comments>https://lobste.rs/s/tyeo20/moonforge_yocto_based_linux_os</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/tyeo20/moonforge_yocto_based_linux_os"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>linux</category>
<category>release</category>
</item>
<item>
<title>Temporal: The 9-Year Journey to Fix Time in JavaScript</title>
<link>https://bloomberg.github.io/js-blog/post/temporal/</link>
<guid>https://lobste.rs/s/ei0ans</guid>
<author>bloomberg.github.io via joomy</author>
<pubDate>Wed, 11 Mar 2026 11:07:15 -0500</pubDate>
<comments>https://lobste.rs/s/ei0ans/temporal_9_year_journey_fix_time</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/ei0ans/temporal_9_year_journey_fix_time"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>javascript</category>
<category>web</category>
</item>
<item>
<title>SQLite WAL-reset database corruption bug</title>
<link>https://sqlite.org/wal.html#walresetbug</link>
<guid>https://lobste.rs/s/mqpba7</guid>
<author>sqlite.org via cve</author>
<pubDate>Wed, 11 Mar 2026 04:18:52 -0500</pubDate>
<comments>https://lobste.rs/s/mqpba7/sqlite_wal_reset_database_corruption_bug</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/mqpba7/sqlite_wal_reset_database_corruption_bug"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>databases</category>
</item>
<item>
<title>AI should help us produce better code</title>
<link>https://simonwillison.net/guides/agentic-engineering-patterns/better-code/</link>
<guid>https://lobste.rs/s/tiktds</guid>
<author>simonwillison.net by simonw</author>
<pubDate>Tue, 10 Mar 2026 17:38:37 -0500</pubDate>
<comments>https://lobste.rs/s/tiktds/ai_should_help_us_produce_better_code</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/tiktds/ai_should_help_us_produce_better_code"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>vibecoding</category>
</item>
<item>
<title>I Ditched Elasticsearch for Meilisearch</title>
<link>https://anisafifi.com/blog/i-ditched-elasticsearch-for-meilisearch-heres-what-nobody-tells-you/</link>
<guid>https://lobste.rs/s/zzhh6z</guid>
<author>anisafifi.com via fanf</author>
<pubDate>Wed, 11 Mar 2026 22:37:14 -0500</pubDate>
<comments>https://lobste.rs/s/zzhh6z/i_ditched_elasticsearch_for_meilisearch</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/zzhh6z/i_ditched_elasticsearch_for_meilisearch"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>databases</category>
</item>
<item>
<title>Using Unicode Half-Stars Symbols in Ratings</title>
<link>https://hyperborea.org/tech-tips/half-stars/</link>
<guid>https://lobste.rs/s/ndtuji</guid>
<author>hyperborea.org via abnercoimbre</author>
<pubDate>Wed, 11 Mar 2026 12:20:13 -0500</pubDate>
<comments>https://lobste.rs/s/ndtuji/using_unicode_half_stars_symbols_ratings</comments>
<description>&lt;p&gt;&lt;a href="https://lobste.rs/s/ndtuji/using_unicode_half_stars_symbols_ratings"&gt;Comments&lt;/a&gt;&lt;/p&gt;</description>
<category>browsers</category>
</item>
</channel>
</rss>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment