\n\nYour entire smart home in the macOS menu bar. Cameras, lights, thermostats, locks, and more \u2013 just one click away.
\n
I keep a Zettelkasten, a personal knowledge base where I write notes and connect them over time. I\u2019m trying out writing about what\u2019s happening there, we\u2019ll see if it sticks.
\nI watched a video about M-shaped careers. People with too many interests (\u201cscanners\u201d) building deep expertise in multiple areas instead of forcing specialization. The Zettelkasten fits here as external memory that lets you switch topics without losing everything.
\nMy job is thinking, then AI executes. I don\u2019t write code anymore, I review it while Claude writes and I direct. My work runs on two tracks: notes and feature plans for AI to implement, and infrastructure that makes AI execution reliable.
\nClarify by prototyping, not planning. My old core value was \u201cMeasure twice, cut once\u201d, which assumed execution was expensive. These days prototyping is cheaper than planning, so I just build stuff and figure it out as I go.
\nAI reframes depth from execution to directing. The M-shaped career model assumes depth means doing the work, but with AI handling execution, depth becomes about understanding enough to see patterns and direct the AI.
\nThese three became permanent notes in my outline, and they\u2019re all connected to each other.
\nAs I use more AI tools these days, my old PKM setup started to feel off, so I ran an audit. The result was Each PKM tool should have one role. Tools work better when they do one thing, and my old setup had overlap everywhere.
\nCurrent pipeline:
\nI also wrote about GTD having a gap. Refinement is a missing stage between Process and Organize. GTD assumes everything captured can be immediately classified as actionable, reference, or trash, but some ideas don\u2019t fit any of these. They\u2019re not procrastinated or blocked, they\u2019re just unformed and still searching for shape. Someday/Maybe gets treated as a parking lot, but refinement is active, not passive. It needs systems for repeated contact with incomplete ideas until they\u2019re ready to become either a Zettelkasten note or a project plan.
\nMost of these notes come back to the same thing: AI made execution cheap, so I\u2019m rethinking how I work and how my tools support that. Three notes became permanent zettels, two are about my PKM setup, and one came from a video. That\u2019s the last two weeks.
\nI can safely say that Liquid Glass is the Butterfly Keyboard of software design. The basic idea is great, but the execution sucks.
\nI just hope it won’t take years to fix it.
\nI had my two previous post in the wrong format, so outgoing links didn’t work. Fixed it.
\n\nPersonal knowledge manager (PKM). Point to any folder and manage your Markdown and plain text notes. Create auto-updating links which can go in both directions.
\n
Pretty early, but looks promising.
\n\n\nA native macOS YouTube Music client built with Swift and SwiftUI.
\n
\n\nCupertino is an Obsidian theme, optimized for desktop and mobile devices. Bringing clean, focused, comfortable reading and writing experience to your vault.
\n
Looks pretty nice, but\u2026 it is still Electron under the hood.
\n\n\nAll your Mac updates. One beautiful interface.
\n
\n\nOwl is a note-taking app currently under development for Apple platforms. An early version is available for Mac.
\n
I don\u2019t usually do New Year\u2019s resolutions, but this year I made an exception. It\u2019s not a goal but a boundary, and it\u2019s that I want to stop spending thinking capacity on problems AI can handle well enough.
\nThings like passport renewals, car maintenance, and health insurance paperwork aren\u2019t intellectually hard problems for me. They\u2019re high-friction re-entries into suspended projects that steal the same executive function I need for creative work. So I\u2019m experimenting with an approach where Claude Code does the thinking and I become the hands and feet that execute.
\nYou might be wondering why I\u2019d give up that much control, but the thing is that I\u2019m not giving up control over outcomes, just over the tedious context-reconstruction that precedes every action.
\nThe standard productivity narrative says \u201cdo things faster.\u201d But I\u2019ve realized that for life admin, speed isn\u2019t the bottleneck. The bottleneck is cognitive re-entry.
\nEvery time I return to a bureaucratic project (health insurance setup, car inspection scheduling, government ID renewal), I pay a tax, and it\u2019s not a time tax but a thinking tax. I have to reload the entire context: What did I do last time? Where\u2019s that document? What\u2019s the phone number? What did the person say?
\nThis context reconstruction drains the exact same mental resource I need for interesting problems like creative work and engineering challenges, the stuff that actually benefits from my personal attention.
\nThe insight is simple, which is that most life admin problems are not hard. They\u2019re just high-friction re-entries. The thinking required isn\u2019t creative; it\u2019s reconstructive. And in my experience, Claude Code handles reconstructive thinking reasonably well (though I still verify critical details).
\nFor knowledge work, everyone talks about AI as a copilot where the human captain does the thinking and the AI assistant helps with execution.
\nFor life admin, I flipped this around so that Claude does the thinking while I handle the execution.
\nIn my early experiments, this architecture shows promise:
\nIt might sound like I\u2019m demoting myself to grunt work, but for me the hard part of life admin was never the physical execution. It was always the mental overhead of figuring out where I left off and what to do next. Once I know what to say and have my documents ready, making the call is easy (though I still need to think on my feet when conversations go sideways).
\nNow let\u2019s talk about how this actually works. For Claude to be useful, it needs context. It can\u2019t just be a chatbot I explain things to each time. It needs to temporarily inhabit the project\u2019s accumulated knowledge.
\nI call this approach \u201cGhost in the Shell\u201d (borrowing from the anime, where consciousness can inhabit different bodies). In my version, the \u201cshell\u201d is the project\u2019s accumulated data and the \u201cghost\u201d is Claude\u2019s reasoning capability that temporarily inhabits it.
\nEach project has a brain: notes, meeting transcripts, task history, linked documents. When I trigger a task, Claude loads this brain and reasons from that context.
\nI should mention the technical setup, which is that I use Claude with MCP integrations that can read from my document database and task manager. I use Mac-specific tools (DEVONthink for documents, OmniFocus for tasks, Hookmark for cross-app links), but the principle should be tool-agnostic since any setup that can aggregate project context and feed it to an LLM would work.
\nWhen I select a task in OmniFocus, Claude follows the project hierarchy upward to find the parent project, then loads all linked documents and notes. It builds a timeline of what happened. Then it can act with full context, not just respond to isolated prompts. Sure, Claude doesn\u2019t truly \u201cunderstand\u201d the project, but it synthesizes the available information well enough for bureaucratic tasks.
\nYou might ask whether managing AI becomes its own overhead, and yes, it can. I\u2019ve found that AI assistants can become another attention sink through clarifications, follow-ups, and re-prompts, and before you know it, managing the AI becomes a new kind of bureaucracy.
\nSo I built in what I call the \u201cone question rule.\u201d If Claude is missing information, it can ask one question maximum. Then it has to provide a best-effort plan with assumptions stated. If it\u2019s still stuck, it downgrades by first moving from doing the task to proposing a plan, and then to explaining how this type of problem generally works so I can handle it myself.
\nThis prevents Claude from becoming needy. It forces decisive action with explicit assumptions rather than endless clarification loops.
\nI also added risk-based escalation. \u201cAutopilot\u201d here means Claude proceeds without asking for confirmation:
\nThis doesn\u2019t prevent all errors. I still need to review what Claude produces, and I\u2019ve caught it confidently stating wrong details. But the guardrails prevent the worst-case scenarios while keeping friction low for routine tasks.
\nClaude doesn\u2019t just prep and disappear but waits for me to come back with results.
\nHere\u2019s a typical flow:
\nClaude: "Here's your call script for the mechanic. Key points to cover:\n timing, parts availability, price estimate. Come back when done."\n\n[I make the call]\n\nMe: "Done. He said Tuesday at 10am works, and the part needs to be ordered."\n\nClaude: Creates follow-up task that surfaces on the appointment date,\n logs the outcome to project notes (so next time I have this\n context), links everything together.This conversational loop keeps Claude as the orchestrator throughout. Every interaction gets logged back to the project\u2019s brain, so next time the context is even richer.
\nThe action types fall into three modes based on how much autonomy Claude gets:
\nAutopilot (Claude does the work, I review results): Research tasks where Claude queries Perplexity or analyzes code. Processing tasks where it extracts key points from articles or documents. Documentation tasks where it saves findings to DEVONthink and creates links between related items.
\nPropose (Claude preps, I decide and act): Phone calls where Claude loads my last conversation with someone, generates a script, and tells me what questions to ask. Emails where it drafts messages based on project context. Setup tasks where it researches best practices and generates step-by-step guides. Decisions where it builds comparison matrices and recommends options.
\nTeach (Claude explains, I learn): Learning tasks where Claude curates resources, creates a study plan, and suggests hands-on exercises, but I do the actual learning.
\nThe key insight is that even when Claude can\u2019t do the task, it can reduce my cognitive load significantly. For a phone call, I usually don\u2019t have to remember anything because Claude already looked up my last interaction with that person and what we discussed. I just read the script (though I still need to adapt when the conversation goes sideways).
\nI built a library of action types as markdown files that Claude reads when working on a task. Each type (research, review, process, draft, analyze, document, setup, learn, coordinate, execute, spec) defines a prep pattern for what Claude should do before I act, a capture pattern for logging results afterward, and which tools to use. For example, the \u201cdraft\u201d type handles phone calls and emails by loading contact details and past conversations, then generating a script or message draft.
\nThe process of building this library is ongoing. I pay attention to what types of actions I do repeatedly, then figure out how Claude could help with each one. Some actions start in Propose mode and graduate to Autopilot once I trust the pattern. The goal is to keep expanding what can be delegated as I discover new friction points.
\nClaude generates plans fresh each time using these patterns plus the project\u2019s accumulated context. In theory, this gets smarter with each use because the context compounds. I haven\u2019t measured this rigorously, but I\u2019ve noticed Claude\u2019s prep work improving as project histories grow richer.
\nWhat I\u2019ve found is that the real risk of AI assistants isn\u2019t that they make you lazy but that they make you a manager of an intern that asks infinite questions.
\nI prevent that by design through a kind of contract with Claude that includes clear modes of operation, risk-based guardrails, explicit definitions of done, and a hard cap on questions, so it behaves like a capable assistant rather than a needy one.
\nIs this worth the setup time? I spent a day building the infrastructure. Whether it pays off depends on how many bureaucratic projects I run through it. Ask me in six months.
\nFor now, I\u2019m experimenting with reserving my thinking capacity for interesting problems. Everything else gets delegated to the ghost in the shell.
\nThis is actually a pretty interesting way to think about which app should be used for which type of writing.
\n\n\nQuestion: What am I writing?
\na. Longform – a novel or non fiction piece of more than 2,000 words.
\nAnswer: Scrivener
\nb. Shortform – a blog post or e-mail or report or some such. Something that I’m “drafting”.
\nAnswer: IA Writer
\nc. Notes – an idea, a thought, a snippet or reminder. Something that may grow, or may link, but does not in its current form represent a draft towards publishing.
\nAnswer: Obsidian
\nQuestion: where do I “archive” writing?
\nAnswer: DevonThink
\nDevonThink is searchable, organised and backed-up. Perhaps archiving is an annual process – so at the beginning of 2022, I archive any and all pieces in IA Writer that predate 2021, leaving a minimum of 1 year and a maximum of 2 years in situ. Scrivener projects can be archived too. (In Plain Text)
\n
See more about this topic:
\n\nI\u2019m writing on my blog from Bike today, which I still love.
\nAlso related but in a different way
\nThinking in Threads with Bike Outliner
\nNote: sometimes I wonder why do I keep a separate Zettelkasten around, since WordPress and my blog could solve that problem easily.
\nWriting in an outliner and publishing directly kinda works for me, since I’m not editing posts in some weird CMS (even if MarsEdit is awesome).
\nI like how easy it is to move rows around.
\nIt’s freeform text, but on the other hand, it’s still pretty structured.
\nI should write more posts from here. It would make sense for quick posts like this, no hard editing, even spelling errors.
\nNobody cares, the main idea is to push content which can be interesting.
\nThere is a new beta of DEVONthink in Liquid Glass.
\nLooks pretty nice.
\nThere\u2019s a new app icon, which is similar to DEVONthink To Go 4.
\nSome smaller UI issues
\nThe magnification glass on the right has a small UI issue.
\nGeneral settings drop-downs are touching each other.
\nThe Archive (native)
\nPlain-text Zettelkasten app with wiki-links, fast search, and minimal UI. File-based (markdown), local-first. The primary tool for this zettelkasten. (Workflow context)
\nMy current or previous usage of the app.
\nPrimary zettelkasten tool – all permanent notes live here. Main journaling and thought capture. Search and browse the knowledge network. Private notes hidden from published site. (Content pipeline)
\nPart of three-tool thinking system: Zettelkasten (The Archive) for journaling and permanent notes, TaskPaper for planning/brainstorming, Emacs for code experiments. (Different thinking modes)
\nArticles and websites
\nSummary of this article.
\nManan Khattar’s case for The Archive: 1,330 notes (350K words) in plain text. Speed and keyboard-centric design minimize cognitive friction. Search-driven workflow enables queries like #literature #inbox (productivity OR writing) (202407 OR 202408). Native Zettelkasten support from zettelkasten.de’s creator.
Trade-offs acknowledged: unique identifier requirement makes wikilinks less readable than Obsidian. No plugin system\u2014requires Alfred workflows for full functionality. No mobile app (uses 1Writer as workaround).
\nCore thesis: tool choice matters less than sustaining writing-focused thinking habit. The Archive’s speed accelerated his adoption.
\nI got to know kindaVim from this article.
\nActually, I already wrote about this before: Bookmarked \u201ckindaVim\u201d
\nThere are videos about kindaVim: kindaVim – YouTube
\nMost of them just ramblings.
\nThis is another app from the same guy for keyboard based UI navigation: Wooshy
\niA Writer (native)
\nDistraction-free markdown editor with focus mode, syntax highlighting, and publishing integrations. Typography-first design from Information Architects.
\nMy current or previous usage of the app.
\nDrafting and linking workflow via smart folders. #Drafting folder shows notes being written, #Linking folder shows notes ready for outline placement. The final drafting step before zettelkasten integration. (Content pipeline)
\nArticles and websites
\nWriting Workflow – Stuart Lennon
\nSummary of the article, focusing on the question \u201cWhat am I writing?\u201d section.
\nStuart Lennon’s framework: three writing categories.
\nLongform (2000+ words) in Scrivener for novels/non-fiction.
\nShortform drafts in iA Writer for blog posts and reports.
\nNotes ideas in Obsidian for thoughts not yet publication-ready.
\nThe “What am I writing?” question determines tool selection. Plain text everywhere for portability. DEVONthink as archive. Open question: whether Drafts could serve as unified entry point with automated routing.
\nBike (native)
\nNative macOS outliner with fast keyboard navigation, collapsible hierarchy, and scripting support. Version 2 adds extensions API and custom attributes. The app running this outline.
\nMy current or previous usage of the app.
\nCore organizing tool for the content pipeline. Ideas get arranged in Bike outlines before export to zettelkasten via Drafts. Also manages the folgezettel outline structure. (Content pipeline)
\nDecoding files are raw Bike outlines for blog posts. All content ready for publishing; private things go to interstitial journal instead. (Journal-type apps)
\nNow used as AI-integrated thinking space with thread-watch and MCP server. Reactive prompting via @question attributes. (Bike thinking tool exploration)
\nArticles and websites
\n\nTaskPaper (native)
\nPlain-text task management using indentation and tags. Same developer as Bike. Good for project brainstorming and task breakdown. (Workflow context)
\nMy current or previous usage of the app.
\nTaskPaper role: plain-text planning and brainstorming for project-specific thinking and task breakdown. OmniFocus shortcut triggers “project brainstorming” in TaskPaper. (Different tools for different thinking modes)
\nHistorical use: interstitial journaling. Deprecated that role, moved journaling back to zettelkasten. Brainstorm sessions now live in the journal rather than separate files. (Plans for TaskPaper changes)
\nWorkflow: Bike for initial outlining, switch to TaskPaper once plan solidifies. Part of the plain-text system alongside The Archive and Bike. (Text-based system)
\nArticles and websites
\n\nDrafts (native)
\nQuick capture app with powerful actions system for routing text to other apps. Automation-first design. Inbox for thoughts, not storage.
\nMy current or previous usage of the app.
\nExport gateway from Bike to zettelkasten. Ideas organized in Bike get exported through Drafts into new notes. Also used for quick capture on Apple Watch at night. (Content pipeline)
\nMCP server spec exists for automation integration. AppleScript-based CRUD operations planned but automation permissions need resolution first. (Drafts MCP spec)
\nArticles and websites
\n\nCraft (native)
\nBlock-based document editor with rich formatting, folder structure, and cross-device sync. Apple Design Award winner, positioned between Notion and traditional notes.
\nMy current or previous usage of the app.
\nPrimary reading queue manager via Readwise-Craft syncer integration. Articles flow in, get processed with highlights, then export to zettelkasten. MCP server enables AI automation. (Readwise-Craft syncer)
\nDaily notes workflow for session logging and project tracking. Block-based structure good for moving content between documents. (Content pipeline)
\nArticles and websites
\n\nDEVONthink (native)
\nDocument/knowledge management database with AI classification, OCR, and smart groups. Stores anything, indexes everything. Research-grade organization.
\nMy current or previous usage of the app.
\nDEVONthink role: document archive and reference storage. Articles, PDFs, web archives, receipts. AI classification and OCR. Research-grade organization with smart groups. (Incremental reading system)
\nReading queue integration: documents captured to DEVONthink appear in OmniFocus @Read/Review perspective. The queue is in OmniFocus, content stays in DEVONthink.
\nMCP access via mcporter enables AI agents to search and retrieve DEVONthink content. Source references in older zettelkasten notes use x-devonthink-item:// links.
\nArticles and websites
\n\nObsidian (Electron)
\nPlugin-extensible markdown vault with graph view, backlinks, and community ecosystem. Electron-based, local files. Power-user PKM tool.
\nMy current or previous usage of the app.
\nLimited use currently. Explored for AI integration with Cursor but zettelkasten workflow remains in The Archive. Considered for consolidating Planning folder but not active. (Obsidian exploration)
\nArticles and websites
\n\nOmniOutliner (native)
\nProfessional outliner from OmniGroup with styling, columns, and export options. More structured than Bike, less nimble. AppleScript support.
\nArticles and websites
\n\nMindNode (native)
\nVisual mind mapping with clean design and focus mode. Exports to various formats. Good for brainstorming spatial relationships, less for linear text.
\nMy current or previous usage of the app.
\nUsed for preparing content before processing. Ideas get connected to mindmaps with tags and links before moving to notes. Public mindmaps shared online (e.g., Japanese Design and Aesthetics). (Processing workflow)
\nArticles and websites
\n\nScapple (native)
\nFreeform note board from Literature and Latte (Scrivener makers). Spatial arrangement of text snippets with connection lines. Idea exploration before structure.
\nArticles and websites
\n\nNotePlan (native)
\nMarkdown notes with calendar integration and task management. Daily notes workflow with backlinks. Hybrid of journaling and task app.
\nArticles and websites
\nMy Ultimate Productivity Tool Workflow for 2024: Plan with Me in NotePlan
\nMy Productivity Workflow 2025: Task Management + Apple Shortcuts + Gmail
\nMy NotePlan Workflow \u2014 Everything I Need to Plan and Execute my Day
\nNotePlan: Your Notes, Tasks, and Calendar in One Place – MacSparky
\nDistill (Tauri)
\nAI-native thinking tool with append-only threads and contextual awareness. Local-first with cloud sync. Discussed earlier in this document.
\nMy current or previous usage of the app.
\nExplored briefly – app installed, bug filed about quick capture window positioning with right-side Dock. Append-only threading concept influenced thread-watch development. (Reactive prompting)
\nRemoved from active toolchain – no current use case. Keep app installed for existing content. (Journal-type apps conclusion)
\nExplored briefly – app installed, bug filed about quick capture window positioning with right-side Dock. Append-only threading concept influenced thread-watch development. (Reactive prompting)
\nRemoved from active toolchain – no current use case. Keep app installed for existing content. (Journal-type apps conclusion)
\nArticles and websites
\n\nBear (native)
\nMarkdown notes with tags, nested tags, and backlinks. Clean native UI. Simpler than Obsidian, more polished than most.
\nArticles and websites
\n\nUlysses (native)
\nLong-form writing app with library management and publishing targets. Markdown-based but hides syntax. For writers shipping manuscripts.
\nArticles and websites
\n\nNoteship (native)
\nMinimal sticky-note style capture app. Menu bar access, quick entry. Capture-focused, not organization-focused.
\nArticles and websites
\nNotebooks 3 (native)
\nHierarchical notebook app with nested books and markdown support. Cross-platform sync. Traditional notebook metaphor with modern features.
\nArticles and websites
\n\nScrivener (native)
\nLong-form writing IDE with binder, corkboard, and compile system. For novels, screenplays, research papers. Structure-first drafting.
\nArticles and websites
\n\nTinderbox (native)
\nHyper-customizable note database with agents, attributes, and visual maps. Steep learning curve, extreme power. For complex knowledge work and research.
\nMy current or previous usage of the app.
\nHigher-level life management – dreams, someday/maybe ideas, things above projects. Scripts connect OmniFocus to Tinderbox for two-way sync. Ideas and goals live here; projects get managed in OmniFocus. (Tinderbox for life management)
\nIncremental Reading document watches Readwise chunks folder. Considered skipping from general chunk processing due to workflow friction getting content in. (Tinderbox workflow friction)
\nArticles and websites
\n\nBBEdit (native)
\nProfessional text editor with grep, syntax coloring, and scripting. Not a note app per se – a text tool that handles notes. 30+ years of macOS development.
\nArticles and websites
\n\nOrg-mode with Emacs (native)
\nPlain-text outliner inside Emacs with TODO states, scheduling, and literate programming. Keyboard-driven, infinitely extensible. Used for programming experiments. (Workflow context)
\nMy current or previous usage of the app.
\nCurrent Emacs usage: Org Mode for programming experiments and literate programming. Still exploratory – set up OmniFocus shortcut for “programming docs” but workflow not fully established. (Different Tools for Different Thinking Modes)
\nOrg Agenda navigation shortcuts documented separately – day/week views, date jumping, standard Emacs keybindings. (Org Agenda Shortcuts)
\nPosition in tool ecosystem: Zettelkasten (The Archive) for journaling and permanent notes, TaskPaper for planning/brainstorming, Emacs for code experiments. Three tools, three thinking modes.
\nArticles and websites
\n\nStrflow (native)
\nStream-of-consciousness capture app with AI integration. Continuous text flow rather than discrete notes. Early-stage product.
\nMy current or previous usage of the app.
\nShort idea dumps when thinking about something – ideas then move to OmniFocus/Bike. Removed from active toolchain; no current use case. App kept installed for existing content. (Journal-type apps conclusion)
\nShort idea dumps when thinking about something – ideas then move to OmniFocus/Bike. Removed from active toolchain; no current use case. App kept installed for existing content. (Journal-type apps conclusion)
\nArticles and websites
\n\nNotePlus (native)
\nNative macOS LLM client (~10MB). Supports OpenAI, Gemini, Anthropic, Ollama, DeepSeek, Copilot. Different category than Bike\u2014this is a chat interface with note-taking features, not an outliner. The “True Native” claim positions it against Electron apps.
\nI’ve been looking for a tool that supports threaded thinking. Distill nails the concept: you think out loud into threads, and AI agents respond inline without breaking the flow. But it’s pricey, hosted, and not native. I want local files so I can use git.
\nSo I built something similar with Bike.
\nThe core concept is append-only threading. New thoughts go at the end. No reorganizing, no inserting between existing ideas. The thread keeps expanding.
\nThis preserves how ideas actually develop. When I capture thoughts in a traditional outline, I’m constantly deciding where things go. That decision-making interrupts the thinking. With append-only, there’s no decision. New thought? Add it at the end. Want to go deeper? Indent.
\nI’ve written about this before in Reviewing append-only workflows. The pattern keeps showing up: Mastodon threads, chat-based note apps like Gibberish, and now Bike.
\nBike is a macOS outliner (think Workflowy, but native and scriptable). What makes it work for threading:
\nThe structure emerges from the content. I start typing, indent when going deeper, and the hierarchy forms naturally.
\nDistill’s insight is that AI agents can watch threads and respond without being addressed directly. I wanted that for Bike.
\nSo I built thread-watch. It watches my outlines for rows tagged with @question and adds responses as indented blockquotes. The workflow:
![Liquid Glass](\"https://static.decoding.io/media/share/Screenshot%202025-12-26%20at%2015.20.33.png\"){kind=link}
![new app icon](\"https://static.decoding.io/media/share/Screenshot%202025-12-26%20at%2016.43.09.png\"){kind=link}
![has a small UI issue](\"https://static.decoding.io/media/share/Screenshot%202025-12-26%20at%2015.21.04.png\"){kind=link}
![touching each other](\"https://static.decoding.io/media/share/Screenshot%202025-12-26%20at%2015.21.59.png\"){kind=link}
@questionThe responses stay in the thread. They’re part of the thinking record, not a separate conversation.
\nHere’s what it looks like in practice:
\n\n\nThe technique I use for AI responses is what I call reactive prompting. The agent isn’t addressed directly. It responds to the thought stream.
\nThe system prompt tells Claude to write like margin notes in a book: factual, terse, no “I found that\u2026” discovery narrative. This matters because conversational AI responses feel like interruptions. Reactive responses integrate with the thinking instead.
\nThe agent has access to my Zettelkasten folder. Before answering, it searches for relevant existing knowledge and links back. I was outlining something about capture workflows, tagged it @question, and the agent found a note I’d written two years ago. The response linked back, and I restructured my outline based on what past-me had already figured out.
Bike becomes the staging area for fast thinking. When something crystallizes, I copy it to the zettelkasten as a permanent note.
\nAppend-only limits restructuring. If the structure is wrong, I can add cross-references and synthesis rows, but I can’t reorganize without losing the thinking record.
\nTagging @question and running the tool is still an interruption. It’s smaller than switching to a chat window, but not invisible.
Standard outlining is about organization. Threading is about capture. I’m not building a perfect structure. I’m externalizing thinking fast enough that I don’t lose it, and the AI assists inline when I need context.
\nThe structure emerges. The thread keeps expanding.
\nI’ve been running macOS Tahoe‘s new Liquid Glass design on my MacBook Air for a couple of days. Parts work, parts don’t, but most people will adapt quickly enough.
\nThe smart thing Apple did was leave 3rd-party apps alone. They run with the previous design intact, giving you familiar anchors while you adapt to the new interface. Electron apps will probably never update, which might be a blessing.
\nThe dynamic background adaptation is broken. Buttons shift from black to white as the background changes, and it’s jarring every single time. Toolbar consistency is all over the place too. Photos and Mail use progressive blur with gradients while Safari keeps the old blurred rectangle, with no apparent logic to which apps get which treatment. Scrolling through Photos is particularly painful as the interface flips from black to white and back again as you move through your library. This isn’t a minor polish issue. It’s a fundamental problem with how the adaptive system responds to content, and Apple should fix it.
\n\n\n\n
\nWhen the general public gets their hands on Liquid Glass, I’m sure people will complaint. The current state is too buggy to survive contact with millions of users. I expect iterative improvements throughout the OS 26 cycle as the design gets tweaked based on user feedback.
\nThe foundation has potential, but the background adaptation logic needs fixing. Until then, it’s a work in progress.
\nThe new /context command in Claude Code revealed how I’m wasting tokens by having all MCP (Model Context Protocol) tools loaded into one session for every conversation. When I’m in the middle of development, I don’t need OmniFocus task management or DEVONthink database tools.
Based on subagents I had the following idea: what if I split Claude into specialized personas, each loading only the tools it actually needs?
\nInstead of one bloated agent trying to do everything, I created three specialized personas:
\nMain Persona (cc): The clean, focused default. Just filesystem, bash, and macOS automation via one MCP server. This handles most of my daily coding tasks with minimal tool overhead.
Research Persona (ccr): The deep investigation specialist. Loads Zen for multi-model reasoning (GPT-5, O3, Gemini Pro), Context7 for library docs, Octocode for code search and analysis, and Codex for deep code research. When I need to debug something complex or review architecture, this persona has the tools.
GTD Persona (ccg): The productivity workflow expert. Connects to OmniFocus for tasks, Obsidian for notes, DEVONthink for documents, and HyperContext for calendar management. This persona focuses on productivity workflows and task management.
Each persona also has access to subagents for specialized tasks, adding another layer of capability without wasting more tokens.
\nThe implementation lives in my dotfiles as simple shell functions:
# Main persona - clean and fast\ncc() {\n claude ${*} --dangerously-skip-permissions\n}\n\n# Research persona with specialized tools\nccr() {\n claude ${*} --dangerously-skip-permissions \\\n --model opus \\\n --mcp-config ~/.claude/mcp-research.json \\\n --append-system-prompt \"You are a code research and analysis specialist \\\nfocused on deep technical investigation, multi-model reasoning, and \\\ncomprehensive code review. Your expertise includes security auditing, \\\narchitecture analysis, debugging complex issues, and researching best \\\npractices across codebases. You leverage multiple AI models for consensus \\\nbuilding, use Context7 to fetch up-to-date library documentation, analyze \\\nGitHub repositories for patterns with Octocode, and generate thorough \\\ntechnical documentation.\"\n}\n\n# GTD persona with productivity tools\nccg() {\n claude ${*} --dangerously-skip-permissions \\\n --model opus \\\n --mcp-config ~/.claude/mcp-gtd.json \\\n --append-system-prompt \"You are a GTD specialist orchestrating a \\\nmulti-layered knowledge system where OmniFocus drives project execution, \\\nObsidian captures notes and knowledge, and DEVONthink archives reference \\\nmaterials. You excel at processing inboxes across all three systems, \\\norganizing projects with proper next actions, capturing meeting notes \\\nwith task extraction, and maintaining a trusted system for all personal \\\nand professional information. Your expertise includes creating bidirectional \\\nlinks between systems using Hook and maintaining clear separation between \\\nactive project work and archived reference materials.\"\n}\nEach persona loads its own MCP configuration file, containing only the relevant tool servers. The --append-system-prompt flag gives each persona a specialized identity and deep expertise in their domain.
The most powerful aspect isn’t just the separation. It’s the ability for personas to delegate to each other. This creates a network of specialized expertise without loading unnecessary tools.
\nWhen delegation happens, I must approve it first. The main persona will say something like: “I’d like to ask the research persona to analyze this Rails controller for N+1 queries using its specialized tools.” Once I type “approved” or “go”, the delegation proceeds.
\nThe delegation always runs in the background using the Bash tool:
\n// How the main persona actually delegates\nBash({\n command: 'ccr -p \"Analyze this Rails controller for N+1 queries and performance issues\"',\n description: \"Ask research persona to analyze Rails performance\",\n run_in_background: true\n})\nThis pattern works both ways. I can also delegate GTD tasks with a simple OmniFocus link:
\n# Ask GTD persona for comprehensive project review\nccg -p \"/gtd:project-review omnifocus:///task/jP9S4CFgPin\"\n
The GTD persona doesn’t just fetch tasks. It discovers linked resources through Hookmark, pulls references and meeting notes from DEVONthink, reads project plans from Obsidian, builds a complete timeline, and suggests concrete next actions based on the full project context. All while the main persona continues with development work.
\nThe background execution is key. The delegating persona doesn’t block waiting for results. It can continue working on other aspects while the specialized persona handles its domain-specific task.
\nThe token savings are significant, but the real benefit is focus. Each persona does one thing well with the right tools for that job. The system prompts shaped distinct behaviors. The main persona stays lean and fast. The research persona leverages multiple AI models for deep analysis. The GTD persona navigates my entire productivity stack.
\n\n\nA reference manual for people who design and build software.
\n
Spent today designing the Rails controller layer for Pezeta.
\nI analyzed Capito, an envelope budgeting prototype I built seven years ago. Standard approach \u2013 categories as envelopes, track balances, show what’s left to spend, like YNAB. But Pezeta takes a different angle. Instead of manually managing category balances, I’m building a goal-oriented system where users define financial objectives and the system calculates the monthly allocations needed to achieve them.
\nFor the controller architecture I designed five main tasks:
\nAuthentication & Ledger Foundation \u2013 Auto-create ledgers with system categories when users sign in the first time. System categories like Income and Safe to Spend get created automatically, so no manual setup required.
\nBudget Controller \u2013 Returns a single JSON endpoint for now, focused on Safe to Spend calculation and future goal balances. This is what the UI will display prominently, not a list of category balances.
\nCategories Controller \u2013 Standard CRUD for user-defined categories. Goals are implemented as special categories with allocation rules that determine how they get funded each month.
\nMonthly Allocations \u2013 This is where goal-driven allocations get applied, but users can override them when needed.
\nTest Infrastructure \u2013 Rails conventions with fixtures that include the system categories.
\nI’m sticking with pure Rails MVC patterns, so no service objects, no complex abstractions. Business logic lives in models, controllers stay thin, everything renders JSON for now since we don’t have a Turbo-based UI yet.
\nThis foundation should make the goal system feel automatic rather than manual.
\nI built the model layer for goal-oriented budgeting in Pezeta today. This is the implementation of the goal system I’ve been designing over the past week.
\nI’ve laid the foundation for four distinct goal types:
\nI still need to test all of this thoroughly, but the models are in place and the calculations are working. What I like about this approach is how it will turn vague financial intentions into concrete monthly actions. Want to save $5,000 for a vacation by next December? The system can calculate exactly how much to set aside each month. Need three months of expenses in your emergency fund? It’ll figure out what that means based on your actual spending patterns. The system tracks progress over time, adjusts recommendations based on current balances, and integrates smoothly with the existing budget structure.
\nNext up is building the UI layer. The foundation is solid, but there’s still work to do before I can actually use these goal features in my own budgeting workflow.
\n\nTwo nights ago, I deleted everything.
\nEvery note in Obsidian. Every half-baked atomic thought, every\nZettelkasten slip, every carefully linked concept map. I deleted every Apple\nNote I\u2019d synced since 2015. Every quote I\u2019d ever highlighted. Every to-do\nlist from every productivity system I\u2019d ever borrowed, broken, or\nbastardized. Gone. Erased in seconds.
\nWhat followed: Relief.
\n
Sometimes it\u2019s good to delete things, but a second brain isn\u2019t an end goal. The sole purpose of a second brain is to help achieve some form of output. It serves as a tool to organize thinking for the future.
\nThe problem with output is that we often don\u2019t know what it will look like when we start. What a Zettelkasten (and GTD too) acknowledges is that you\u2019ll have an outcome of some kind and you need a framework to manage it. That\u2019s all.
\nPeople who capture their ideas then leave them there are still doing the same thing, but can probably think through their output in their head. I do that sometimes too, and it\u2019s fast.
\nBut as soon as I\u2019m working on anything that requires more thinking capacity than I currently have, I reach for PKMs and task managers to handle it.
\nI usually split my long-term project-specific Zettelkastens off from my main one, though.
\nPreviously:
\n\n\nLisaGUI is a web OS inspired by the Lisa Office System, Apple\u2019s first graphical user interface (GUI). It aims to perfectly recreate the Lisa\u2019s iconic 1-bit UI directly in your browser.
\n
Lisa had so many great UI ideas. We still have the concept of stationary pads on macOS today.
\nToday I worked on planning Pezeta’s goal system. I experimented with using two AI agents to collaborate on breaking down the work: Claude Opus 4 as the coordinating agent with read/write access, and ChatGPT o3 through Repo Prompt for implementation planning.
\nThe idea was to leverage each model’s strengths. Claude handles the file operations and manages the task structure, while ChatGPT analyzes the codebase and suggests implementation approaches. They literally chat with each other using the Repo Prompt MCP tool, with Claude asking questions and ChatGPT providing analysis. Sometimes I had to intervene to guide them back on track or stop them to wait for my approval on design decisions. It’s like having two developers pair programming, each bringing their own expertise to the planning session.
\nThe approach surfaced some important design decisions. Instead of service objects, we’re going with Rails-idiomatic patterns: model methods, class methods on MonthlyAllocation, and callbacks for snapshots. Everything through ActiveRecord.
The bulk of the work went into defining test specifications:
\nEach test case addresses specific edge cases. What happens when a monthly goal already has surplus funds? How does a deadline goal behave when it’s past due? Getting these details right upfront matters.
\nAll calculations will use “opening balance,” the balance at the start of the month, not the current balance. This prevents allocation amounts from changing as transactions flow through during the month, which matches how people think about budgets.
\nThe session produced seven detailed task files ready for implementation. Each includes exact specifications, test cases, and implementation notes. The tasks cover Single Table Inheritance strategy, historical snapshots through callbacks, and model-driven allocation suggestions.
\nTomorrow I’m letting Claude Code implement these tasks.
\nHad a focused session today expanding the goal system design for Pezeta. Yesterday I started with the basic concept, today I dove deeper into the details.
\nI expanded all four goal types with detailed logic and edge cases:
\nSo this session was used to think about these and take a deep dive in the goal rule engine inner workings. I still have questions but it will be iterated when I start using the MVP. I have goal examples in my current budget for all of these.
\nThe technical challenge was handling goal changes over time. What happens if you change your vacation target from $2,000 to $3,000 in March? Should January’s allocation be recalculated as if you always wanted $3,000?
\nI went with a snapshot approach. Each month gets its own snapshot of goal parameters. Past months never change retroactively. January’s allocation was based on January’s goal settings, period. This preserves the audit trail of what you were thinking at the time.
\nToday wrapped up the core data foundation with the Budget view model and spent time planning the goal allocation rules that’ll make this budgeting app different.
\nThe Budget model pulls everything together. Give it a ledger and a month, and it calculates the current state of all categories, their allocated amounts, actual spending, and remaining balances.
\nThis is just pure Ruby class that takes the foundational models and presents them in a way that makes sense to users.
\nThe bigger breakthrough today was finalizing the goal allocation rules. This is what’ll set Pezeta apart from traditional envelope budgeting.
\nInstead of managing dozens of budget categories, users define their financial goals\u2014things they actually care about saving for. The app figures out how much to allocate to each goal and shows one simple number: how much is safe to spend on everything else. Goals run in the background while you focus on that single “Safe to Spend” amount.
\nI want to implement four goal types:
\nEach follows different allocation formulas, but the key insight is the engine just calculates expected allocations\u2014it never moves money automatically. That stays under your control.
\nI’ve got a React prototype artifact built with Claude that demonstrates how this will look. We have a simple “Safe to Spend” number on top, goals running quietly in the background. I just created this to validate the UX approach, but it shows the vision clearly.
\nDay two of building Pezeta. Yesterday was all planning and starting. Today was about getting my hands dirty with actual Rails models.
\nCompleted the three core models today:
\nCategory model (with proper enums and validations)MonthlyAllocation model (with decimal precision for money)Transaction model (income/expense tracking)Not the most exicting part for people to see (and I\u2019m spending about 1\u20132 hours a day to build this), but I want to get the models right, that\u2019s the whole point of why I\u2019m writing the first MVP in Rails, I know Rails the best.
\nI\u2019m trying to stick to test-driven development. Write the test first, make it pass, move on. Feels slower at first, but every time I change something, I know immediately if I broke anything.
\nHere\u2019s something weird I\u2019m trying: I started a new zettelkasten in Obsidian yesterday to drive this whole project.
\nThe idea: use notes as the driving force to guide Claude Code through writing the actual code. I think in the zettelkasten, then I let AI collect that context and draft the output: code, blog posts, status updates etc. At the end I still manually review and adjust.
\nI\u2019m using a modified version of Ian Nuttal\u2019s Task Magic system to drive the development. Break big plans into smaller tasks, then automate the output generation. Not just code generation though. Claude Code helps with everything like documentation, tests, even these blog posts, but I\u2019m still reviewing everything.
\nMy theory is that this might be the fastest way to do solo development today. Capitalize on what AI is actually good at – turning clear thinking into working artifacts.
\nWe\u2019ll see if this scales. but for now, it\u2019s working.
\n\n\nI recently shipped Context, a native macOS app for debugging MCP servers. The goal was to build a useful developer tool that feels at home on the platform, powered by Apple’s SwiftUI framework. I’ve been building software for the Mac since 2008, but this time was different: Context was almost 100% built by Claude Code. There is still skill and iteration involved in helping Claude build software, but of the 20,000 lines of code in this project, I estimate that I wrote less than 1,000 lines by hand.
\n
Indragie shipped a native macOS app with Claude Code writing 95% of the code. Haven’t dug into the full post yet, but Simon Willison pulled out the key bits \u2013 Claude nails SwiftUI but gets tripped up by Swift’s newer concurrency stuff.
\nFinally decided to start building Pezeta today. It’s been in my mind for years – since we started Agyvihar, our Hungarian podcast about personal finance and productivity.
\nSimple: a budgeting app that tells you one number – how much you can spend right now.
\nThat’s it.
\nNo charts. No categories. No endless reports. Just open the app and see: $47. Or $134. Or whatever you can actually spend without screwing up your budget.
\nThe One Number – Your daily spending limit. Updated in real-time. Factors in your bills, savings, everything.
\nGoals – Want to save $2000 by December? Cool. The app adjusts your daily number automatically. No spreadsheets required.
\nThat’s basically it. If budgeting were simpler, maybe more people would actually do it.
\nI’m not jumping straight into building native apps. Starting with a Rails MVP instead.
\nFew reasons:
\nThis Rails version won’t go public. It’s just for me to validate the idea and work out the kinks.
\nI’m putting it all out there because watching someone figure things out might be more useful than another perfect tutorial.
\nPartly for feedback – maybe someone will tell me this is a terrible idea before I waste six months on it.
\nPartly for accountability – harder to quit when people are watching.
\nPartly because I’m curious what happens when you share the messy process instead of just the polished result.
\nStarting with Rails 8 setup, basic models, authentication. The boring foundation stuff that nobody talks about but takes forever.
\nAfter that, we’ll see. Building software is weird.
\nI actually learned a couple of new things about TextEdit and Terminal.
\nMihai Parparita is asking on Mastodon:
\n\nWhat are good examples of early Mac OS X software (10.0-10.4 era, pre-Intel switch)?
\n
Then providing the initial list.
\n\n\n\n
\n- OmniWeb, OmniGraffle, OmniOutliner and other Omni Group software
\n- Transmit 2, Unison, Candy Bar, Pixadex, and other Panic software
\n- Audio Hijack and other Rogue Amoeba software
\n- NetNewsWire and MarsEdit
\n- Watson
\n- Acorn
\n- TextMate
\n- Quicksilver
\n- VoodooPad
\n- BBEdit
\n- SubEthaEdit
\n- GraphicConverter
\n- iCab
\n- Camino
\n- Fetch
\n- Default Folder X
\n- MenuMeters
\n- DEVONthink
\n- CodeWarrior
\n- Resorcerer
\n- FruitMenu, WindowShadeX and other Unsanity haxies
\n- LiteSwitch X
\n- DragThing and PCalc
\n- TinkerTool
\n- LaunchBar
\n- Path Finder
\n- Konfabulator
\n
Good memories, although most if these are still alive and well.
\nBlogging is essential because it provides knowledge and serves as a valuable reference for us and others.
\nOn the other hand, getting recognized as a blogger these days is rarely going to work. People are increasingly consuming different types of content these days, and this is largely due to the way AI is changing the way we interact with the web.
\nThat said, having a blog for thinking out loud is still the best way to learn and then \u201creport\u201d it, even if we\u2019re the only ones reading it.
\nLately, I started to do a bunch of (what I call) Project-based Zettelkasten![\"\u2122\"](\"https://s.w.org/images/core/emoji/15.0.3/72x72/2122.png\")
, which are just git repositories containing a bunch of Markdown files related to different projects.
This post reminds me that iA Writer can grab all that stuff and make it available in one app.
\nPreviously:
\n\nI\u2019m not a big fan of Arc (it just overcomplicates the browsing session), but if you want something similar based on Firefox, then the Zen browser could be interesting.
\nSteve Jobs said this in 2011:
\n\n\nNo ads. We build products that we want for ourselves, too, and we just don’t want ads.
\n
Then, we have a new rumor, that Apple is thinking about showing ads in Maps.
\nMy reaction? No ads. You build products that you and I want for ourselves, too, and we just don’t want ads.
\nThis is not more for me, but I love the idea of using the MagSafe Wallet this way.
\n![](\"https://decoding.io/wp-content/uploads/2025/02/image.jpg\")
\nMac(os)talgia is exploring my 2020 work-from-home routine with an added touch of nostalgia. How would have the same workflow looked like with the tools of today and the limitations of yesterday.
\n
I would take this Mac OS 9 design of Slack, Zoom, and Figma immediately.
\nYet more reading about President Musk (I\u2019m gonna steal this).
\n\n\nElon Musk\u2019s unceasing attempts to access the data and information systems of the federal government range so widely, and are so unprecedented and unpredictable, that government computing experts believe the effort has spun out of control.
\n
Lately, I feel early Gilead vibes while reading the news.
\nContinuing 3380:
\nMaybe I should use tags as a thread title in a way that group posts together. I\u2019m using pretty general tags currently, which is fine, but a tag page is just another filtered wall of text. This way, each post could be part of an ongoing conversation. Here\u2019s an example page for a thread.
\nIt means that I could have more precise tags which can be displayed in the sidebar, or even better, I can link to a thread which kind of acts like an article on its own. It was just written in a chronological order.
\nAs an icing on the cake, a post could belong to multiple threads.
\nContinuing 3378:
\nIf you visit the front page of my blog, you\u2019ll notice that I organize posts by day, inspired by the daily notes feature from Roam Research.
\nI typically enjoy this wall of text style more than having separate posts highlighted individually (perhaps that\u2019s why I like threading on Mastodon).
\nThe issue is that WordPress lacks a similar feature, but I would love to have threading on my blog too. My Zettelkasten can stack connected notes, which aligns with this concept.
\nMaybe I should create something that allows me to link posts together, with backlink support.
\nPreviously:
\n\nI sometimes think we might need different ways to share ideas instead of long articles\u2014it\u2019s important to mention that I’m talking about ideas, not stories.
\nLooking at how I read, I see that I usually skim content instead of reading it all at once. Then, I go back to the interesting parts later. If you think this is like Incremental reading, then you\u2019re correct.
\nI\u2019m not sure what might happen if I started using formats closer to the raw idea then heavily edited posts.
\nI\u2019ve been thinking about using voice input more in my workflow after watching Chris from DailyTekk talk about quitting typing. He made some interesting points about how speaking (120-150 words per minute) is naturally faster than typing (40-60 words per minute). What was interesting is his observation about typing disrupting the flow with constant self-editing\u2014I\u2019ve definitely experienced this.
\nSo, I\u2019m experimenting with various voice-based input options for my Mac (and possibly iOS/iPadOS as well).
\nI recently acquired VoicePen as a tool to take voice notes. It\u2019s not a direct dictation app, though. Instead, it works by recording voice in a separate app and then allowing you to copy the cleaned transcription back to your note-taking app. It maintains a history of transformations (such as fixing grammar and cleaning the text) alongside the original transcription. I can revisit and reapply these transformations or copy different versions of the text. It\u2019s a decent tool, but it\u2019s not designed for real-time voice input.
\nI looked at Inbox AI too, but it\u2019s more of an automation app and I don\u2019t want to invest time learning it right now. I\u2019m sure it can do this.
\nBolt.AI offers something different – it has an inline dictation feature. I\u2019m currently trying out its dictation with chat assistants\u2014being able to talk naturally to AIs instead of typing.
\nI\u2019m also experimenting with macOS\u2019s built-in dictation. It integrates well with text editing and shows what I\u2019m typing in real-time. Sure, it sometimes types the wrong stuff, but these can be fixed quickly with the built-in writing tools.
\nI\u2019m noticing there\u2019s a difference between thinking out loud and just dictation\u2014I\u2019ll need more time to figure out what works best for each case. For now, I think each input method\u2014typing, voice, even Apple Pencil\u2014probably has its place. I\u2019ll likely end up using a mix of them depending on what I\u2019m trying to do.
\n\nI try to operate on the \u201cNothing Doesn\u2019t Go in Here\u201d principle. Flipping through the notebook that is on my desk right now, I see shopping lists, notes from conference calls, sermon notes, a little chart I made when working on some invoices for Relay, a doodle of the Widgetsmith icon I made for some reason, and a lot more. A receipt for a recent meal with a friend fell out of it onto my lap; I\u2019ll probably tape that into the notebook for safekeeping.
\n
The good thing about Field Notes is that you don’t have to care about it. It doesn’t feel pristine like some other brands. So, I mainly capture similar things, but I group them by date. I have meeting notes, programming session notes, random ideas, outlines, journal entries, etc.
\n\nThese notebooks are a trail of breadcrumbs dating back almost 15 years. If I flip through an old one, I get a glimpse of what was going on in my life at that time. I can go to the notebook I was using when we launched Relay, or when I quit my job. I like having them on hand; seeing them in my studio each day makes me happy.
\n
I never scanned my old Field Notes but kept them safe in a wooden box. I don’t care if they are destroyed someday since I use my notebooks as a temporary capture tool. If something is interesting, I’ll transfer it to my GTD system or just scan that part of the notebook.
\nOtherwise, the notebook can be discarded.
\nPreviously:
\n\n\n\nProgramming as we know it is in a process of dying out.
\nMost haven’t come to terms with it yet.
\n
I’m just saving this, so I can look back at it in 3 years and see if it’s becoming true or not.
\n\n\nGoogle removed a pledge to not build AI for weapons or surveillance from its website this week. The change was first spotted by Bloomberg. The company appears to have updated its public AI principles page, erasing a section titled \u201capplications we will not pursue,\u201d which was still included as recently as last week.
\nAsked for comment, the company pointed TechCrunch to a new blog post on \u201cresponsible AI.\u201d It notes, in part, \u201cwe believe that companies, governments, and organizations sharing these values should work together to create AI that protects people, promotes global growth, and supports national security.\u201d
\n
Shit, I just finished watching the Terminator movies yesterday.
\nWell, meh\u2026
\nI like Tapestry. It\u2019s going to be a great app, and people seem to like it. No Mac app for now though.
\nI\u2019m just not sure why everyone was complaining about Reeder 5 when it was released doing basically the same thing.
\nAfter today\u2019s post on disabling the wake on lid open, I got a comment from smarthieef on how we can switch off this behavior for the keyboard too, so we can only turn on the MacBook using the Touch ID button, making the keyboard cleaning process way easier.
\n\n\n\n
\n- Press and hold the left Control and Command buttons with right Shift button for a total of 7 seconds.
\n- Without releasing them, press the Power button and hold together for an additional 7 seconds until your laptop shuts down. The login screen may flash for a second so don’t prematurely release the keys until the machine is off.
\n- For your next startup, your Mac can only be powered on by using the Power button or closing and opening the lid.
\n
I tested this on a 16-inch M1 MacBook Pro, and it works indeed. It is also a one-time thing, so on the next boot, it resets this behavior to the default one, so we can get the best of both worlds.
\nUpdate: I tested this on an M3 MacBook Air too, and it didn\u2019t work.
\nHere are a couple of highlights from Why I Make Minimal Use of Social Media:
\n\nI have a relative who loves to post about their perfect, loving family, filling their Facebook feed with happy family photos. But, as you might guess, they conveniently leave out the reality of their many conflicts and challenges. It\u2019s a polished image that doesn\u2019t show the truth.
\n
Nobody\u2019s gonna share their struggles and failures as much as their fun times, so I think this behavior is just for display. Remember that when you see something like this, it’s just a snapshot of their day. Nothing more, nothing less.
\n\nThere are other ways to keep in touch with family members and friends. Emails and texts can be addressed to a group of contacts or individuals. By not using social media yourself, you can discourage others from using it as their preferred way to \u201ckeep in touch.\u201d
\n
I never emailed any of my friends, but we usually keep in touch via text messages. Texting is like having a peer-to-peer social network.
\nI just realized that the ActivityPub plugin for WordPress does Fediverse replies inline too. Pretty nice!
\n![\"\"](\"https://decoding.io/wp-content/uploads/2025/01/image_db89a4bf-565d-488b-8eab-d8397b8a8498-1024x400.jpg\")