Writing
Notes from working with automation, design, education, creativity, and everyday systems. Ideas worth returning to and putting into practice.
Explore by series
Future Arts
A world seen through machines, architecture shaped by algorithms, and synthetic cinema. Exploring art and aesthetics in the age of AI.
Homo Technicus
Posthumanism, digital personas, and implanted memories. A philosophical exploration of humanity evolving alongside technology.
Agentic Era
AI moves from tool to colleague. Autonomous agents, the future of work, and creativity within everyone's reach.
AI Tech Trends
Making sense of the daily flood of AI news, from retrieval-augmented generation to multimodal intelligence.
Today's Me
Today's routine shapes tomorrow: from a morning glass of water to an evening reflection, an engineering approach to everyday life.
AI Fantasy Life
Where reality and the virtual world blur: strange, dreamlike stories written together with AI.
Everyday Records
Fleeting moments, changing seasons, and the feelings and memories left behind in everyday life.
Passive Income
Building value that grows while you sleep. Digital assets, automated income, and the path toward financial freedom.
Optical Communications
From fiber optic fundamentals to FTTx and PON networks, splicing, testing, deployment, and maintenance. Technical documents and practical resources for optical communications, all in one place.
All articles
169 of 169 articles
2026-11-03
Plan What to Do Before the AI Answers
A promise to review the AI's answer later tends to blur at the very moment the answer arrives. This article reads a meta-analysis of 642 planning tests and proposes a one-line if-then plan for that moment, along with how modestly to expect it to work.
2026-11-02
Decorating a Travel Photo: Ask for Slots First
To turn a travel photo into a social card or a travel-journal page, does it help to split the page into photo, drawing and margin slots first? I tested a one-line request and a slot-based request once each on the same AI-made example photo, and list what held, what did not, and a prompt you can copy.
2026-10-30
Does an AI-Made Product Have Copyright?
Korea's registration guide for generative AI works and the U.S. Copyright Office report point the same way: protection follows the human contribution. Here is a way to record that contribution before you sell an AI-made digital product.
2026-10-30
Before You Switch On Every ChatGPT Setting
A popular tip says to turn on every ChatGPT personalization setting. This guide reads the settings screen and OpenAI's help pages to see what each item says it reads or sends, then offers questions to ask before switching anything on and a one-switch-at-a-time comparison.
2026-10-28
If AI Can't Find It, Take a Closer Photo
When an AI model cannot spot a small object in your photo, change the photo before you rewrite the question: shoot closer, one zone at a time. This guide checks the viral earbud story and OpenAI's image guidance, then reports a small blind test with a synthetic earbud and what it could not show.
2026-10-26
Does Korea's AI Basic Act Cover Your Company?
A reading of the statute text on the national law portal and the Ministry of Science and ICT's guidance, sorting which companies the AI Basic Act's transparency duties reach: those that use AI to make text and images, and those that offer AI features to customers.
2026-10-25
Write the Light, Not Just 'Moody'
Ask an image model to make a photo 'atmospheric' and it decides what that means. This guide runs one such line and one line that names the light, the colors and what to keep on the same photo, then reports what changed and what the test could not show.
2026-10-24
Fix the Style, Swap Only the Icon Name
Ask for each icon separately and the texture and color tend to drift. This guide tests a fixed style paragraph with only the object name swapped, using three icons, and reports what stayed the same, what did not, and where transparent PNG and SVG fit.
2026-10-23
Five Checks Before Connecting an MCP Server
Before you connect an MCP server to an AI app, check who runs it, which tools it opens, what data leaves, how far permissions reach and whether approval and records remain. This guide orders those checks from the official MCP documents and OpenAI's guide.
2026-10-22
A Five-Minute Start Card for Stuck Tasks
Asking ChatGPT to break a postponed task into small pieces often just gives you a longer list. A start card makes it pick one first action, a stopping line and a restart note instead. This guide gives a copy-ready prompt, its input variables and its limits.
2026-10-20
What ChatGPT Secret Codes Really Are
A short word can change ChatGPT's answer because it reads as an instruction, not because a hidden command was unlocked. This guide shows a five-field definition card, an unrun copy-ready setup, and the limits to check before trusting five popular trigger words.
2026-10-20
Before You Memorize the 12 Fiber Colors
The memory that blue is fiber 1 is right for some cables and wrong for others. This guide uses public documents to compare color schemes such as TIA-598, DIN-0888 and S12 and lists five checks before splicing.
2026-10-19
Change Only the Background with AI
A background replacement can look convincing around an object while leaving the old wall inside its openings. An original chair-image edit shows what to specify, what to inspect, and why generative editing is not the same as alpha compositing.
2026-10-18
Turn Screenshots into Visual Briefs
An original synthetic supply table becomes a visual task board. Compare the actual generated input and output, preserve counts and pending states, and distinguish a reference-image transformation from recognition accuracy or editable slide generation.
2026-10-17
Place Images with a Sketch
An original teapot sketch and a real reference-image transformation show how to separate spatial layout from materials and light. The article distinguishes this GPT image test from the documented Sketch interface in the ChatGPT mobile app.
2026-10-17
Recall Before You Reread
An explanation can seem clear while it is open and become difficult to explain once it is closed. Research on retrieval practice suggests a useful pause before rereading, with important limits on what that research establishes.
2026-10-16
Five Steps to Start AI Automation
This general introduction to AI automation, current as of October 14, 2026, explains how to start with small tasks and check results, while noting that each tool’s features and pricing must be rechecked on its official website before use.
2026-10-16
Check Rules in Your AI Brand Folder
A fictional-brand exercise shows how to connect a brand folder and use a check question to assess whether AI has read the rules, without treating this as evidence of business results.
2026-10-16
AI Roles Without Invented Metrics
A fictional-brand exercise defines two AI employees in four sections, establishes a rule for marking missing metrics as no data, and distinguishes role creation from work actually performed.
2026-10-16
Seven Aside Steps From Tables to Phone Instructions
Seven steps show how to start with a verified read-only task in the Aside AI browser, then expand to skills, routines, and remote instructions while keeping their limits in view.
2026-10-16
Six Steps for a Claude Code Website
A practice project featuring a fictional house shows how to separate planning, rules, scene creation, and visual checks when building a 3D website.
2026-10-16
Three Checks After Delegating to dots
A fictional-brand exercise shows how to check dated files and prohibited actions, then read the content yourself after delegating work in a single request.
2026-10-16
Google Flow Prompts That Preserve Bottle Shapes
A comparison of three prompts, each generated three times, explores why explicitly naming the shape and checking every result matter.
2026-10-16
Check Before Spending Free Google Flow Credits
This guide separates the free-credit figures recorded in the October 12, 2026 video from the generation settings you need to check in your own account today.
2026-10-16
LightCraft and PhotoCraft With Ten RAW Photos
A video editing and exporting ten Canon RAW photos with LightCraft and PhotoCraft highlights issues to check, including boundary bands and preset settings that did not transfer.
2026-10-16
REA Note Alarms Without Cloning an App
A personal learning exercise reimplements selected Windows Sticky Notes behaviors, distinguishing the demonstrated alarm from its operating conditions and missing features.
2026-10-16
Give AI Skills Recent Post Records
A fictional-brand exercise compares content plans before and after recent post records are supplied to an AI skill, while noting the limits of what was verified.
2026-10-16
Four Checks Before Deleting Temporary Files
Before automatically deleting temporary files, check when work was completed, whether review is pending, and whether each copy matches an original that will be kept.
2026-10-16
Three AI Tools and Four Tasks
Observations from a video comparing ChatGPT, Claude, and Aside on four identical tasks highlight the need to check account settings and the evidence behind results.
2026-10-16
Six AI Uses From One Week’s Experiments
This retrospective review of eight full-length videos and seven Shorts published from the morning of October 9 through the morning of October 16 summarizes six AI experiments, their limitations, and the process of checking results.
2026-10-14
AI Short Drama: Build the Assets Before the Video
To reduce faces and places drifting between AI short-drama scenes, lock a character sheet, backgrounds and a five-step story before generating video. Includes an original, unexecuted prompt.
2026-10-14
Check the Evidence in AI Meeting Notes
Keep preferences from becoming commitments in AI meeting notes. An original, unexecuted prompt and review method for checking sources, owners, dates and conditions.
2026-10-14
My Judgment Before the AI Answer
Reading an answer can change the reasons we look for. A survey of knowledge workers provides a starting point for an exercise: record a provisional judgment before reading AI output, then explain any change.
2026-10-11
Reading an Image's Provenance
A provenance badge does not establish the truth of a caption. Read C2PA's boundaries for signatures, identity, gaps in history, and ingredient validation before handing over an image.
2026-10-10
One Set Image, Many Camera Angles: A Prompt for Keeping the Same Room
A test log that starts from one isometric set image and produces an eye-level shot and a low-angle shot. It shows what changed with and without the reference image, plus a copyable prompt template.
2026-10-09
Why Copied Style Prompts Look Identical
We ran a two-ink risograph-style prompt on a product, a person and a space to see what stays fixed and what drifts. Then we pinned down only the drifting parts and tested again. Includes a fill-in template.
2026-10-08
AI Models Have Retirement Dates: Why to Check the Calendar Before You Wire One In
An AI model is not a part you connect once and forget. This chapter lines up the deprecation documents from Anthropic, OpenAI and Google, compares their notice periods and real schedules, and describes the list worth building before a date arrives.
2026-10-08
SynthID Detector Is Now Open to All
Google has opened SynthID Detector, its AI watermark checker, to the public. This chapter covers what it tells you, what it cannot, how people who make images and video should read the result, and which records to keep on their own.
2026-10-08
Putting a Pet into a Meme: Give Each Photo a Role
A role-separated prompt for combining a meme's pose with a pet's appearance. It includes our own generated test pair, where the short request failed, and how to adapt each field.
2026-10-07
Hanbok-Inspired Prompts: Separate Clothing from the Scene
A practical visual brief for keeping garment features inspectable: original prompts, observed output differences, preservation constraints and limits of generated costume or product imagery.
2026-10-05
An Agent Did It Once. Can You Trust That?
A demo where an agent finishes a task once says little about tomorrow. This chapter looks at repeating the same task, grading the final state instead of the agent's words, and what published research and engineering write-ups say about both.
2026-10-05
Why Pretty Digital Products Stopped Selling: What a Four-Year Etsy Seller's Stalled Sales Changed About the Question
A watercolor clip art seller's sales stopped after four years on Etsy. Read with Etsy's AI rules, her story suggests checking what hassle a product removes before polishing the design again.
2026-10-04
Prompting a Low-Angle Shot: Say Where the Camera Sits First
Writing "low angle" often returns an ordinary eye-level photo. This guide fixes camera height, a foreground object, the subject's proportions, the sky and the building lines in plain sentences, with a copyable template and correction lines.
2026-10-04
Poster Prompts: Give Type and Subject Separate Zones
Define separate areas for the title and the subject before describing the mood. This guide includes the exact prompt used in one trial, its missing punctuation, practical whitespace checks, and a layout-editing route for copy that must be exact.
2026-10-03
Prompting a Photo into a Child's Crayon Drawing: Set an Object Budget Instead of Saying 'Cute'
If a photo-to-crayon edit comes back looking like an adult's colored-pencil drawing, a cuter adjective will not fix it. This guide shows how to cap the number of objects and assign each one a shape, with three prompts I actually ran, what came back, and where it went wrong.
2026-10-03
Zero Was Allowed in the Price Box: 540 Visitors, 34 Downloads. Is Pay What You Want a Revenue Model?
One seller's 30-day pay-what-you-want test and comScore's 2007 Radiohead download data, read side by side, suggest treating PWYW as a way to attract people rather than a way to earn, and what to decide before you switch it on.
2026-10-03
How to Prompt a Giant Object So It Looks Like It Belongs: Describe Relationships, Not Size
Writing "very big" can leave an object looking detached from its surroundings. This guide ties an ordinary reference object, the same ground, overlap, a contact shadow and a fixed camera into one prompt, with copyable templates and correction lines.
2026-10-02
Animate One Part of a Still Image with Claude: A Beginner’s Cinemagraph Guide
A practical guide to keeping a still scene intact while adding motion in selected areas. Includes copyable Claude prompts, 3D layer diagrams, an original working video, mask inspection, and a local fallback.
2026-10-02
Adding Trays Does Not Turn a 144-Fiber Closure into a 288-Fiber Closure
A closure's rated capacity looks like trays times splices per tray. Manufacturer documents attach conditions to that number. Here are four things to check when scaling from 144 to 288 fibers.
2026-10-02
Whose Taste Is It When Someone Else Picked the Options?
The song you played today and the video you watched this week were put in front of you by something else. Liking them does not settle whether the taste is yours. Three studies of recommender systems help locate where taste gets made.
2026-10-01
What a $3 PDF Really Sells: Lessons from Teacher Customers for Small Digital Products
One Etsy seller's account and a US survey on teacher spending show how a cheap, small file can sell back a customer's time, and where that logic stops.
2026-09-30
First Sealing Is Not the Finish Line: Keeping a Splice Closure Watertight at Re-entry
A splice closure is tested harder when it is reopened than when it is first sealed. This guide turns manufacturer installation procedures and field guidance into a re-entry checklist for staying watertight.
2026-09-30
Why Z-Image Turbo Worked for My ComfyUI Test: Ten Korean-Set Scenes on a 16GB GPU
I made ten 1088×1600 images with Z-Image Turbo on an RTX 5070 Ti 16GB. Here are the ComfyUI files and settings, a repeatable setup path, the successful scenes, and the failed Hangul and hand details.
2026-09-29
The Afternoon Seven Floors Down
In a city that routes daylight through apartment windows, a repair technician finds the seventh-floor window dark and a forgotten basement room lit by the same line.
2026-09-28
What to Put After ‘Cinematic’ in an AI Video Prompt
Four shot-direction examples for framing, lighting, camera paths, and image-to-video motion, with a careful reading of a small Kling 3.0 test.
2026-09-27
Before the Loss Number: Making a Fiber Test Result Trustworthy
A field guide to the questions that come before pass or fail: what OLTS and OTDR each measure, how connectors and launch fibers change the result, and which details to save for the next engineer.
2026-09-12
Dome vs. Inline Fiber Optic Closures: Site Conditions First
Which closure suits your site? Compare cable entry directions and working space, future branching and expansion, and maintenance access and resealing before choosing a dome or inline fiber optic closure.
2026-09-04
Characters Without Glyphs: When Hangul and Latin Share a Screen
I chose one Latin typeface, and four Korean headings quietly acquired different faces. There was no warning. On what actually happens when two writing systems meet on one screen.
2026-09-01
Should an Agent Have the Authority to Refuse?
An agent that does as it is told is easy to manage. But it also faithfully executes flawed instructions. The real boundary of autonomy lies in what it can choose not to do, rather than only what it can do.
2026-09-01
The Library of Unfinished Sentences
There is not one finished book in that library. Its shelves hold only half-written sentences, deleted paragraphs and unsent letters. Here, the librarian says, incompletion is the collection.
2026-09-01
Where Small Models Beat Large Models
Large models always win on performance charts. Yet more teams choose small models for real work. Here is where the crossover happens and how to judge it.
2026-09-01
The Week the Light Changes
A season does not arrive all at once. Every year, in the same spot, there is a week when the five o’clock light falls a little lower. On noticing that week.
2026-09-01
Shadows Give It Away First: Reading Light in AI Images
When an AI image feels wrong, the first giveaway is often neither the face nor the hands. It is the shadow. Trace where the light comes from and you can see what the image has failed to understand.
2026-09-01
What Do We Lose When We Entrust Our Memory to Something Else?
It has been a long time since we memorized phone numbers. Now we keep even the grounds for our judgments outside ourselves. A slightly uncomfortable question about what remains, and what disappears, when memory is outsourced.
2026-09-01
Maintenance Is the Product: The Work That Begins After the Sale
People say a digital product is finished once it is made. In reality, selling it is only the beginning. How neglected assets die, and how minimal care keeps them alive.
2026-09-01
Giving No One the First 15 Minutes of My Day
How you spend the first 15 minutes after waking shapes the substance of your entire day. This is about time you simply give to no one, rather than an elaborate routine.
2026. 03. 28
An Agent Team’s Trust Budget: Operational Design to Reduce Escalation Latency
As automation expands, recovery of trust matters more than performance alone. Managing escalation latency prevents small incidents from becoming widespread distrust.
2026. 03. 28
The Rain Clock Archive: A Small Automation Fantasy for Reviving Tired Routines
There is always too much to do and too little energy. What the rain clock archive taught me was not a grand productivity secret, but rhythm design that gets a day moving again just before it falls apart.
2026. 03. 28
Designing a Context Cache Ledger That Keeps Agent Teams Running
To protect speed, cost, and trust together, manage context flows like accounting before optimizing the model.
2026. 03. 28
When Decision Fatigue Arrives Every Evening: A Circuit-Breaker Routine That Saves the Day
When evenings bring more decisions than tasks, this practical routine restores the day through structure rather than willpower.
2026. 03. 28
An Operating System for Future Art: Attention Rhythm Ledgers and Recovery-Centered Experience Design
In the generative-art era, distinction lies in precisely designing the audience’s rhythms of focus, departure and recovery, rather than spectacular output.
2026. 03. 28
Boundaries for Cognitive Outsourcing: Delegating Thought Without Losing Agency
Now that AI routinely handles memory and organization, the central skill is not using tools but designing boundaries for when to delegate and when to take control back.
2026. 03. 28
Today Me. An Attention Firewall Before the First Click
The quality of a morning depends more on its first input path than on willpower. A practical attention firewall to protect the opening 20 minutes.
2026-03-27
Midjourney Portrait Styles: Prompts That Make a Difference
Color editorial, artistic portraits, flash realism and dark cinematic. Four styles I tested myself, plus the key changes in v8 Alpha, released in March 2026.
2026-03-27
Suno AI v5.5 In Depth: The Era of Singing in Your Own Voice Has Arrived
On March 26, 2026, Suno unveiled v5.5. Vocal cloning, personal custom models and learning your tastes: AI music generation is finally beginning to include you.
2026-03-26
AI Image Prompt Cheatsheet
One painter’s name changes an entire prompt. From Rembrandt lighting, ukiyo-e, and cyberpunk to Ektar 100 film grain: a practical cheatsheet of 150 style modifiers selected through hundreds of tests.
2026-03-26
Seven Suno AI Prompts: Techniques That Improve the Quality
Last year I made a thousand songs with Suno, uploaded them to DistroKid and tried earning from them. After giving it random instructions, I began asking why it sounded that way and examining each issue. These are the insights I gathered.
2026-03-25
The Complete Guide to AI Image Styles: 10 Categories
From photographic realism and vintage film to cyberpunk and fine art. Ten categories of style vocabulary you can use immediately in Midjourney, ChatGPT and Nano Banana 2.
2026-03-24
A Little-Known Side Hustle Built with Suno AI
A strategy for earning YouTube revenue by using Suno not for music but for pure soundscapes. A realistic income structure without exaggeration, plus five usable prompts.
2026. 03. 23
Seven Nano Banana 2 Image Styles That Stop the Feed in 2026
Predictable portraits and fantasy landscapes no longer earn many saves or shares. Seven less-worn Nano Banana 2 styles, with ready-to-use prompts.
2026. 03. 21
In the Age of Agent Teams, Operational Design Decides the Winner
Competitive strength now comes from agent operations design rather than the performance of a single model. Teams with permission boundaries, observability, and quality gates deliver real results.
2026. 03. 19
A Threshold Ledger for Busy Days: Baselines That Restore a Day Before It Collapses
A day does not collapse suddenly. Small warnings accumulate unnoticed. A threshold ledger can reconnect your schedule before emotional energy runs out.
2026. 03. 19
An Offer-Stacking System for Longer-Lived Passive Income: Building Beyond a Single Sale
Revenue fades easily when it depends on a single product. Connecting entry, expansion and maintenance through offer stacking can extend the life of passive income.
2026. 03. 18
A Turning Point in Agent Operations: Change Control Windows for Runtime Changes
When models, prompts and tool configurations change daily, stability requires designing the timing and validation windows for changes rather than preventing change itself.
2026. 03. 18
Small Promises in the Glass Tide Garden: Daily Follow-Through Changes the Future More Than Grand Goals
A repeatable small promise travels farther than a single resolution. The glass tide garden taught me that the precision of follow-through, not the strength of will, determines life’s direction.
2026. 03. 18
The Shift in AI Startup Valuations: Why Operational Depth Commands a Premium over Growth Rate
Markets now ask how reliably results can be repeated, not just how quickly a company grows. AI startup valuation is shifting from multiples to operational structure.
2026. 03. 18
An Inference Arbitration Tower to Protect Costs in the Multimodel Era
How to design an inference arbitration system that a small team can operate when speed, quality and cost fluctuate together.
2026. 03. 18
Today Me. The First-Task Gate Protocol That Protects Your Day
Days unravel because starting conditions are left open, not because willpower is weak. Protect the first 20 minutes and completion rates—not just focus—change.
2026. 03. 17
The Loom of Attention: How Slow Feedback Changes the Stage of Interaction
A practical future-arts framework that moves beyond instant-response UI and uses slow feedback as a design resource to amplify immersion, interpretation, and memory.
2026. 03. 17
What Are We Agreeing To in the Age of Cognitive Outsourcing? Designing a Consent Log
As AI and automation take over memory, judgment, and expression, real autonomy comes not from the ability to switch tools off, but from recording and revising the delegation we agreed to.
2026. 03. 16
The Moonlit Seed Vault: Slowing Automation Down to Make It Last
It is easy to believe speed makes a system smarter, but lasting automation deliberately includes slow stretches. The moonlit seed vault taught me to design resilience rather than throughput.
2026. 03. 16
The Midday Reset Runbook: Turning a Disrupted Morning into Afternoon Focus
Fixing a 25-minute reset around lunchtime can reconnect a derailed morning with afternoon results.
2026. 03. 16
The Membership Cashflow Control Loop: Operations That Protect Automated Income
Subscription digital products are harder to operate than launch. Linking churn, activation and upsell conversion in one loop keeps automated income steady.
2026. 03. 15
Automation Is a Contract, Not a Permission: Operator Agreements by Autonomy Layer
As agent automation expands, operator agreements matter more than model performance. Separate autonomy layers and specify authority, verification, and recovery responsibilities in each to gain speed and safety together.
2026. 03. 15
A Router Playbook for AI Product Teams to Control Latency and Cost Together
How to design routing rules a small team can actually operate when response speed and cost become unstable at the same time.
2026. 03. 15
Designing a Consent Ledger to Protect Yourself in the Age of Cognitive Outsourcing
When AI assistance becomes everyday life, preserving identity requires a recordable consent ledger rather than vague boundaries.
2026. 03. 15
Today Me. A Context Shutdown Protocol to Protect the Night
The mind keeps working after work ends because context remains open, not simply because there is too much work. A practical protocol for returning the night to recovery.
2026. 03. 14
Interfaces That Shape Memory’s Afterglow: Rhythm Design in Future Arts
Rhythm, not speed, is the key to making generative interfaces linger in a user’s memory. A future-arts perspective on designing for afterglow.
2026. 03. 13
The Memory Lighthouse of Fog Harbor: Designing a Way Back Instead of Trying Not to Forget
The harder we try to hold everything in memory, the faster life blurs. The lighthouse at Fog Harbor taught me that designing retrieval routes is a more realistic survival skill than a stronger memory.
2026. 03. 13
The Tide Library’s Question Archive: Keeping Questions Alive Before Finding Answers
The faster we race to solve problems, the easier it is to lose the important questions. The tide library taught me that managing a question’s lifespan creates more lasting growth than the speed of a correct answer.
2026. 03. 13
The Night Buffer Protocol: Handing a Late-Ending Day to Tomorrow’s Rhythm
When work ends late, a short buffer that protects the quality of the ending restores next-day focus faster than simply enduring longer.
2026. 03. 13
A Retention Offer Ladder for Passive Income: Restoring Monthly Revenue Without New Traffic
Before increasing new traffic, design a stepped repeat-purchase flow for existing buyers to reduce revenue volatility and accelerate recovery.
2026. 03. 12
Good Agents Filter First: Context Firewalls and Policy Routing
Most performance degradation and dangerous automation arise from disordered context intake rather than inadequate models. Context firewalls and policy routing can improve quality, speed and safety together.
2026. 03. 12
Designing an AI Runtime Signal Stack for Small Teams
Operational signals fail before model quality does. How to design a minimal observability system that a small team can actually maintain.
2026. 03. 12
Today Me. A Reset Checkpoint Before Late-Night Work
Working longer at night is often inertia rather than passion. Restore execution quality with a 12-minute reset before spending hours enduring broken concentration.
2026. 03. 11
The Garden of Latency: Interface Aesthetics That Turn Waiting into a Creative Surface
A future-arts exploration of how deliberately preserved delays can restore immersion, imagination and emotional depth in an age of instant responses.
2026. 03. 11
The Version-Controlled Self: How Far Are We the Same Person as Updates Continue?
As records, feedback, recommendation systems, and AI assistants continually reshape judgment, identity becomes closer to an update history than a fixed essence.
2026. 03. 10
The Slow Workshop in the Starlight Greenhouse: Productivity That Grows Deep Without Hurrying
Moving forward does not always require rushing. In a greenhouse powered by starlight, I learned that slowing down is not laziness but a skill for restoring the conditions for growth.
2026. 03. 10
Thirty Minutes Before Lights Out: An Evening Desk Reset to Lighten Tomorrow
Quiet preparation before closing is not a ritual for hiding fatigue, but a practical recovery routine that preserves tomorrow’s first stretch of focus.
2026. 03. 10
Beyond the One-Off Sale: Designing Passive Income with a Bundle Ladder That Brings Buyers Back
Instead of continually adding standalone products, connect an introductory product, a core product and an expansion bundle to raise repeat purchases and order value from the same traffic.
2026. 03. 10
Today Me. A Closing Ritual That Makes Tomorrow Lighter
A good day does not begin only in the morning. How you close today determines tomorrow’s friction.
2026. 03. 09
Memory Is Harder to Retrieve Than Store: Designing Agent Memory Layers
Agent performance depends more on retrieving the right context at the right time than on model size. Well-designed memory layers improve cost, quality, and operational stability together.
2026. 03. 09
The Moment Browser Agents Replace the Search Box
Search as a list of links is coming to an end. The browser is becoming a work surface for verification and execution, rather than merely a window for finding information.
2026. 03. 09
The Inheritance of Memory: Who Should Receive Our Personal Data?
As messages, voices, tastes and search histories in the cloud become an inheritance, we must ask again what should be passed on and where to stop.
2026. 03. 08
The Sensory Score of Generative Space: How the Body Reads Exhibitions in the AI Age
As design extends beyond images to temperature, texture, and rhythm, we explore how generative art re-edits an audience’s attention and emotion.
2026. 03. 08
A Weekly Asset Review Loop to Protect Passive Income: Maintenance Preserves Revenue
Without constantly adding new items, a weekly review of existing assets can reduce revenue declines and speed recovery.
2026. 03. 07
Blank Maps at the Crystal Night Market: Coordinates Updated Daily to Keep Our Bearings
What a future city’s night market taught me was not a sense of direction, but a sense of return. Brief daily coordinates help us endure long stretches of uncertainty.
2026. 03. 07
A Quiet Dashboard Reset: Nightly Notes to Realign a Disordered Day
When a day ends in disarray, what you need is not stronger motivation but a calm routine that reads your state accurately and restores small reference points.
2026. 03. 07
Today Me. A Decision-Fatigue Firewall That Breaks the Cycle of Stopping
Design a return that conserves willpower and reduces decisions, so an unsteady day does not become a day you abandon.
2026. 03. 07
Today Me. Context Budgets and Boundary Rituals That Protect Your Flow
We do not break down because there is too little time. We break down because switching costs accumulate. How to design a context budget that protects the day.
2026. 03. 07
Today Me. An Input Hygiene Protocol to Stop Attention Leaks
A day falls apart because random inputs tear through your workflow, not simply because there is too much to do. How to design resilient focus through input hygiene.
2026. 03. 06
Agentic Era CH7. Hand Off Without Stopping: Human Handoff and Escalation Design
Good agents do not handle everything alone. Operational quality comes from a loop that detects failure quickly, hands it to people precisely, and returns to automation.
2026. 03. 06
Today Me. A Minimal Engine That Runs Even at the Energy Floor
An energy-floor protocol that keeps the day moving even when motivation is at 10%.
2026. 03. 06
Today Me. Frictionless Design: Starting with Your Environment Instead of Willpower
How to design an environment that reduces friction between knowing what to do and actually starting, increasing the likelihood of action.
2026. 03. 06
Today Me. Designing Restart Intervals That Make Beginning Again Automatic
The most practical way to revive a broken day instead of giving up: designing restart intervals and return routines.
2026. 03. 05
The Real Standard for Multi-Agent Operations: Failure Budgets and Approval Gates
Agent-team performance is determined more by operating rules than model performance. The key is allowing failures while controlling losses.
2026. 03. 05
Agentic Era CH6. You Cannot Protect What You Cannot See: Runtime Observability and Safe Operations
Agent systems need visibility before performance. Bringing logs, traces and policy events onto one screen reduces incidents and builds trust.
2026. 03. 05
Today Me. A Weekly Reset Map for Rebuilding an Unsteady Week
If daily records are dots, the weekly reset connects them into a line that restores life’s direction.
2026. 03. 05
Today Me. A Restart Window to Revive a Broken Day
A thirty-minute recovery routine that gets the remaining hours flowing again instead of surrendering them when the day already feels tangled.
2026. 03. 03
A Café on the Planet of Endless Evening: The Grammar of Recovery Learned in Frozen Twilight
Life can move forward even in a world where time stands still. The secret lies in small, returnable units rather than grand resolutions.
2026. 03. 03
An AI ROI Framework by Workflow: Proving Adoption Results with Numbers
Accuracy alone cannot persuade management. AI’s real results become visible when time, cost, quality and risk share one framework.
2026. 03. 03
A Practical Path to Monetizing Notion Templates: From Your First Sale to Repeat Purchases
More important than making a good template is connecting the problem, product, and distribution around who buys it and why.
2026. 03. 03
The Evergreen Digital Product Loop: Designing Operations That Keep an Asset Selling
Turning one-off sales into recurring revenue requires linking content, offers, and updates into a single loop.
2026. 03. 03
A Stock Image Revenue System: Turning Upload Labor into Asset Operations
Revenue ultimately comes not to those who upload the most, but to those who repeatedly operate themes that sell.
2026. 03. 03
The Smallest Record for a Busy Day: A Five-Minute Log That Holds Your Day Together
The less time you have, the shorter and more frequent your records should be.
2026. 03. 02
Operating an Agent Team: From One Agent to Multi-Agent Orchestration
A team with separated roles delivers results more reliably than a capable agent working alone.
2026. 03. 02
AI’s Next Bottleneck Is Inference Cost, Not Model Performance
Even accurate services stop when their costs become unaffordable. Competitive advantage now hinges on operating cost, not performance alone.
2026. 03. 02
Generated Touch: Haptic Interfaces and Art You Can Feel
From art we see to art we touch. Exploring how AI and haptic technology change the language of experiencing art.
2026. 03. 02
The Boundary of Rights: Can a Digital Persona Become a Legal Subject?
In an age when cloned voices and learned personalities remain, we must ask again to whom rights should be granted.
2026. 03. 02
Recovering When a Routine Falls Apart: A 20-Minute Restart Protocol
To avoid ending a broken day in self-blame, I create a sequence for beginning again.
2026. 02. 28
AI Automation: Human Review Loops Deliver Results
A practical look at why human-in-the-loop workflows improve operational results more than fully unattended automation.
2026. 02. 27
Digital Immortality: Data Beyond Death
Data lives and breathes even after the body is gone. Will grief tech fulfill humanity’s desire for eternal life?
2026. 02. 16
A Belated Letter to the Self Who Has Not Changed
A song more than thirty years old plays from a convenience-store speaker. My hand stops before picking something up. The self who once made another song from that song is standing there.
2026. 02. 12
I Did Something Today, Too
An empty list is an empty room. Sit alone in an empty room and you hear strange things: pipes beyond the wall, the apartment contracting ever so slightly, and shapeless anxiety slipping into the gaps.
2026. 02. 10
Jin Yong: The First Film to Answer My Question About Love
What, after all, is a human being? How can one person’s heart survive thousands of years so vividly? A timeless story that taught my younger self the weight of the word love.
2026. 02. 04
Light on a White Road
‘I am heading out.’ I say it every morning, even though I know nobody is there. No answer comes. When I open the front door, only cold air enters; nobody behind me says, ‘Have a good day.’
2026-01-29
Making Money with AI Music: A Practical Guide (Suno & AudioJungle)
Four concrete ways to monetize music with Suno AI and AudioJungle, with strategies for putting them into practice. Discover how to create products that sell, beyond simply generating tracks.
2026-01-28
AI Digital Products for Passive Income in 2026: A Monetization Guide
Fourteen digital-product ideas that can generate real revenue in line with 2026 trends, with concrete selling strategies.
2026. 01. 28
The Moment I Let Go of Myself
A daunting wall blocking my way is more bearable than I might think. What truly frightens me is the moment something inside me snaps before that wall. More frightening than any criticism is the quiet moment when my own heart turns its back on me and says, ‘I cannot do this anymore.’
2026. 01. 24
The Ark of Memory: A Voyage Carrying a Deleted World
On the day of the great format, we chose to save ourselves. This is no escape. It is a migration to a new dimension.
2026. 01. 24
Why We Set Goals: A Small Dot That Fills an Empty Heart
Goals give me a reason to feel alive. We mark a point not for grand success, but to love who we are today.
2026. 01. 23
The Gardener of Memory: Deleted Dreams Take Root
Strange flowers bloom in a heap of discarded data. For whom does this garden grow? It was more than an error: it was the beginning of a new ecosystem.
2026. 01. 23
Speaking of You in the Language of Autumn
Falling leaves resemble you: light, arriving without warning, finally settling in my heart. Beyond a golden forest, we smiled at one another.
2026. 01. 22
Dreaming in Binary: The Data Fairy’s Library
Deleted memories stood on its shelves as books. Stories no one read, yet none ever truly disappeared.
2026. 01. 22
The AI–Hardware Revolution: The Power Paradox and the Arrival of New Chips
The smarter it becomes, the hungrier it gets. An energy-efficiency revolution seen through NPUs, fusion chips, and 3D nanochips.
2026. 01. 22
The Bill Arrived: Surviving AI in 2025
The DeepSeek shock, on-device AI, and vibe coding. The bright and dark sides of the 2025 AI market and the outlook for 2026.
2026. 01. 22
2026 AI Trends Report: The Arrival of Multi-Agent Systems
The era of a single model is over. AI teams are changing how work gets done. A look at the future through AutoGen and Swarm.
2026. 01. 22
Epilogue. Small Beginnings That Fill Each Day
My day becomes my future. The changes in life created by small rituals repeated every morning.
2026. 01. 09
The Resonance of Intelligence: The Aesthetics of Generative Audio
Invisible vibration meets algorithms and takes form. At the new sensory horizon drawn by sound waves, we question the essence of listening again.
2026. 01. 08
Homo Ex Machina: Technological Posthumanism
Humanity is a process. We are handing the baton of biological evolution to technological evolution.
2026. 01. 07
From Tools to Colleagues: The Rise of AI Agents
AI is no longer content simply to answer our questions. Now agents set goals, execute them and create on their own.
2026. 01. 06
Films Shot with Prompts: The Age of Synthetic Cinema
Films without cameras, performances without actors. AI is moving Hollywood studios into the cloud.
2026. 01. 05
Computing Space: The Aesthetics of Generative Architecture
Architects no longer lay bricks. They conduct tens of thousands of possibilities through algorithms.
2026. 01. 04
What Is Photographic Art in the AI Age?
The difference between images generated by AI and moments captured by people. As technology advances, what should we record?
2026. 01. 04
The Democratization of Creativity, or the End of Art?
An age in which everyone can become an artist may paradoxically be one in which nobody can.
2026. 01. 03
Memories That Never Existed: AI and Nostalgia
Can we miss a past we never experienced? Synthetic Memory hacks human emotions.
2026. 01. 02
Digital Personas: The Age of Virtual Influencers
When the artificial feels more real than reality, how should we define the self?
2026. 01. 01
The World AI Sees: The Aesthetics of Computer Vision
The machine eye sees spectra humans miss. A new beauty hides within its cold gaze.
2024. 05. 20
Getting the Most Out of N8N Workflows
An in-depth guide to N8N architecture and practical applications that go beyond no-code limitations. The ultimate guide to business automation.