AI

• The Death of Clever Code

  One positive product of working with Agentic tools is they rarely suggest clever code. No arcane one-liners, no “look how smart I am” abstractions. And, well, I’m here for it.

  Before we continue it helps to understand a bit about how LLMs work. These models are optimized for pattern recognition. They’ve been trained on massive amounts of code and learned what patterns appear most frequently.

  Clever code, by definition, is bespoke. It’s the unusual pattern, the one-off trick. There just isn’t enough training data for cleverness. The AI gravitates toward the common, readable solution instead.

  Let me give you an example.

  Show Me the Code

  Here’s a nested ternary:

  const result = a > b ? (c > d ? 'high' : 'mid') : (e > f ? 'low' : 'none');
  

  I’d be impressed if you could explain that correctly on your first try. What happens when there’s a bug in one of those conditions? Good luck debugging that.

  Now here’s the same logic:

  let result;
  if (a > b) {
    if (c > d) {
      result = 'high';
    } else {
      result = 'mid';
    }
  } else {
    if (e > f) {
      result = 'low';
    } else {
      result = 'none';
    }
  }
  

  A lot easier, right? If it’s easy to read, it’s easy to maintain. The AI tooling doesn’t struggle to read either version, but you might, and when there is a bug, explaining exactly what needs to change becomes the hard part.

  Actually wait. It turns out, not all complexity is created equal.

  Two Kinds of Complexity

  Essential complexity is the complexity of the problem itself. If you’re building a mortgage calculator or doing tax calculations, there’s inherent complexity in understanding the domain. You can’t simplify that away, and you shouldn’t try.

  Accidental complexity is the stuff you introduce. The nested ternary instead of the if/else. Five layers of abstraction for the sake of abstraction that only runs in a specific edge case. Generic utility functions where you’ve tried to cover every possible scenario, but realistically you only need two or three cases.

  Ok but what about abstraction, since abstraction is where accidental complexity loves to hide?

  Good Abstraction vs. Bad Abstraction

  Abstraction shows up everywhere in programming, but let’s think about it in two flavors.

  Good abstraction hides details the caller doesn’t need to care about. The interface clearly communicates what it does. Think array.sort(), you look at it and immediately know what’s happening. Those dang arrays getting some sort of sorted. You know exactly what it does without caring about the implementation.

  Bad abstraction hides details you do need to understand in order to use it correctly. Think of a processData() method that’s doing six different things with an internal state that’s nearly impossible to test. And splitting it into processData1() through processData6() doesn’t help either. That’s just moving the vegetables around on your plate which doesn’t mean you’ve actually finished dinner.

  AI Signals

  So why does any of this matter for working with AI coding tools?

  Because if the agents keep getting your code wrong, if they consistently misunderstand what a function does or there are incorrect modifications, that’s a signal.

  It’s telling you that your code has some flavor of cleverness that makes it hard to reason about. Not just for the AI, but for your team, and for you six months from now.

  The goal is to code where the complexity comes from the problem, not from the solution. The AI struggling with your code is like a canary in the coal mine for maintainability.

  Mar 25, 2026 / AI / Programming / Code-quality

  Enjoyed this?

• Your AI Agent Needs a Task Manager

  If you’ve spent time working with AI coding tools, you’ve probably hit the compaction wall. Suddenly, your agent knows what it’s currently working on but has completely forgotten the five other things connected to it.

  This is the memory problem, and it’s a big one.

  The Context Window Isn’t Enough

  Your AI agent needs some sort of memory system that lives outside the context window. When you’re working on simple, one-off tasks, the chat-as-workspace approach works fine. You ask a question, you get an answer, you move on. But the moment you’re tackling a complex set of related tasks? It breaks down fast.

  I’ve been thinking about this through the lens of a framework I’m calling the Agentic Maturity Model. The short version is that there are distinct levels to how teams and developers use AI agents, and moving between levels isn’t about using “better” tools, but rather it’s a shift in how you approach the work.

  Four months ago, there were no real options. The good news? It seems like all the model providers recognize this is the next frontier. Memory and persistence are where I’m looking for the actual progress to happen next.

  Claude Code has certainly gotten better in these areas over the last couple of months. They’ve added an auto memory feature in beta. They added a lightweight Tasks system based on a Todo system called Beads built by Steve Yegge. His key idea was that the task state should live outside the context window.

  These are meaningful building blocks towards an actual working memory system that persists across sessions and survives compaction.

  We’re Almost There

  The tooling and harneses we built on top of the LLMs are already changing how software gets built, but where we are headed? Here is what I think:

  • auto-improving memory: where the agent learns your patterns, your codebase, your preferences
  • persistent task tracking that survives compaction: Tasks, todos, issues, whatever you want to call them. The point is they exist outside the conversation.

  When those two pieces come together properly, the workflow for everyone will change again.

  Your agent doesn’t just respond to the current prompt. It knows where it is in a larger plan, what’s been done, what’s blocked, and what’s next. That’s the difference between a helpful chatbot and an actual collaborator.

  We are so close I can taste the blood in the water, oh wait, that’s mine. ☠️

  Mar 24, 2026 / AI / Developer-tools / Software-development

  Enjoyed this?

• Using Claude to Think Through a Space Elevator

  When I say I wanted to understand the engineering problems behind building a space elevator, I mean I really wanted to dig in. Not just read about it. I wanted to work through the challenges, piece by piece, with actual math backing things up.

  So I decided to see what Claude and I could do with this kind of problem.

  Setting it Up

  I have an Obsidian vault that Claude Code/CoWork has access to, and I started by asking it to help me understand the core challenges of building a space elevator. First things first: clearly state all the problems. What are the engineering hurdles? What makes this so hard?

  From there, I started asking questions. Could we use an asteroid as the anchor point and manufacture the cable in space? How would we spool enough cable to reach all the way down to Earth? Would it make more sense to build up from the ground, down from orbit, or meet somewhere in the middle?

  I’ll admit I made some mistakes along the way. I confused low Earth orbit with geostationary orbit at one point but Claude corrected me and explained the difference. That’s part of what makes this approach work. You’re not just passively reading; you’re actively thinking through problems and getting corrected when your mental model is off.

  Backing It Up With Math

  Here’s where it got really interesting. I told Claude: don’t just describe the problems. Prove them. Back up every challenge with actual math and physics calculations.

  I also told it not to try cramming everything into one massive document. Write an overview document first, then create supporting documents for each problem so we could work through them individually.

  So Claude started writing Python code to validate all the calculations. I hadn’t planned on that initially, but once it started writing code, I jumped in with my typical guidance. Use a package manager, write tests for all the code.

  What we ended up with is a Python module covering about 12 of the hardest engineering challenges for a space elevator. There’s a script that calls into the module, runs all the math, and spits out the results. It’s not a complete formal proof of anything, but it’s a structured way to think through problems where the code can actually catch mistakes in the reasoning.

  And it did catch mistakes. That’s the whole point of this approach, you’re using the calculations as a check on the thinking, not just trusting the narrative.

  Working Through Problems Together

  As we worked through each challenge, I kept asking clarifying questions. What about this edge case? How would we handle that constraint?

  It was genuinely collaborative, me bringing curiosity and some engineering intuition, Claude bringing the ability to quickly formalize ideas into code and calculations.

  The code isn’t public or anything. But the approach is what I think is worth sharing.

  The Hard Part Is Still Hard

  My main limiting factor is time. The math looks generally fine to me, but if I really wanted to verify everything thoroughly, I’d need to spend a lot more time with it. A mathematician or physicist who’s deeply familiar with these calculations would be much faster at spotting issues. Providing guidance like, “no, you shouldn’t use this formula here, that approach is wrong.”

  I can do that work. It’s just going to take me significantly longer than someone with that specialized background.

  This is what I mean when I talk about working with agentic tools on hard problems. It’s not about asking an AI for the answer. It’s about using it as a thinking partner; one that can write code, run calculations, and help you check your reasoning as you go.

  For me, that’s the real power of tools like Claude. Not replacing expertise, but amplifying curiosity.

  Mar 23, 2026 / AI / Claude / Space / Engineering

  Enjoyed this?

• Voice-to-Text in 2026: The Tools and Models Worth Knowing About

  As natural language becomes a bigger part of how we build software, it’s worth looking at the state of transcription models. What’s the best way to get voice to text right now?

  For a lot of people, talking to your computer is faster than typing. You can stream-of-thought your way through an idea, prompt your tools, and get things moving without your fingers being the bottleneck. If you haven’t tried it yet, it will change how you work with your machine. I’m not exaggerating.

  The Tools

  Here’s what people are actually using for desktop voice-to-text:

  • Willow Voice — Popular choice, lots of people swear by it
  • SuperWhisper — My current pick
  • Wispr Flow — Another well-regarded option
  • Voice Ink — Worth a look?
  • Aiko — From an Open Source dev, Sindre Sorhus
  • MacWhisper — Solid Mac-native option

  I’ve tried several of these, and the biggest pain point for people is going to be that many require monthly subscriptions. I’ve been happy with SuperWhisper and it is worth mentioning they still have a pay for it once (Lifetime) option, so you don’t get locked into monthly payments forever. That said, Willow Voice and Wispr Flow both have strong followings.

  The Models Behind the Magic

  Most of these tools started with OpenAI’s Whisper, the voice model released and open-sourced back in 2022. With Whisper, you could run solid transcription locally on your own hardware.

  But we’re a few years past that now, and there are some more models to choose from. Here is a summary table of the current state of the transcription models.

  Model	Company	Released	Local Run?	Used in Desktop Tools?	Best For
  Whisper Large-v3	OpenAI	Nov 2023	Yes	Yes (The Standard)	Multilingual accuracy (99+ langs)
  Whisper v3 Turbo	OpenAI	Oct 2024	Yes	Yes (Fast Settings)	Best speed-to-accuracy ratio for local use
  Nova-3	Deepgram	Apr 2025	Self-Host	Limited (API-based)	Real-time agents; handling messy background noise
  Parakeet TDT 1.1B	NVIDIA	May 2025	Yes	Developer-focused / CLI	Ultra-low latency; significantly faster than Whisper
  SenseVoice-Small	Alibaba	July 2024	Yes	Emerging (Fringe)	High-precision Mandarin/English and emotion detection
  Canary-1B	NVIDIA	Oct 2025	Yes	Developer-focused	Beating Whisper on technical jargon & punctuation
  Voxtral Mini V2	Mistral	Feb 2026	Yes	Yes (Privacy apps)	High-speed local transcription on low-VRAM devices
  Granite Speech 3.3	IBM	Jan 2026	Yes	No (Enterprise focus)	Reliable technical ASR with an Apache 2.0 license
  Scribe v2	ElevenLabs	Jan 2026	No	Via API	Extremely lifelike punctuation and speaker labels

  
  ---
  

  We’re at an interesting inflection point. You can articulate your thoughts faster by speaking than typing, its becoming a real productivity gain. It’s not just an accessabiltiy aid anymore. People who can type well enough are using these tools on a daily basis.

  That’s all for now!

  Mar 17, 2026 / Productivity / AI / Tools / Voice

  Enjoyed this?

• Your Context Window Is a Budget — Here's How to Stop Blowing It

  If you’re using agentic coding tools like Claude Code, there’s one thing you should know by now: your context window is a budget, and everything you do spends it.

  I’ve been thinking about how to manage the budget. As we are learning how to use sub-agents, MCP servers, and all these powerful capabilities we haven’t been thinking enough about the cost of using them. Certainly the dollars and cents matters too if you are using API access, but the raw token budget you burn through in a single session impacts us all regardless. Once it’s gone, compaction kicks in, and it’s kind of a crapshoot on whether it knows how to pick up where we left off on the new session.

  Before we talk about what you can do about it, let’s talk about where your tokens go, or primarily are used.

  Why Sub-Agents Are Worth It (But Not Free)

  Sub-agents are one of the best things to have in agentic coding. The whole idea is that work happens in a separate context window, leaving your primary session clean for orchestration and planning. You stay focused on what needs to change while the sub-agent figures out how.

  Sub-agents still burn through your session limits faster than you might expect. There are actually two limits at play here:

  • the context window of your main discussion
  • the session-level caps on how many exchanges you can have in a given time period.

  Sub-agents hit both. They’re still absolutely worth using and working without them isn’t an option, but you need to be aware of the cost.

  The MCP Server Problem

  MCP servers are another area where things get interesting. They’re genuinely useful for giving agentic tools quick access to external services and data. But if you’ve loaded up a dozen or two of them? You’re paying a tax at the start of every session just to load their metadata and tool definitions. That’s tokens spent before you’ve even asked your first question.

  My suspicion, and I haven’t formally benchmarked this, is that we’re headed toward a world where you swap between groups of MCP servers depending on the task at hand. You load the file system tools when you’re coding, the database tools when you’re migrating, and the deployment tools when you’re shipping. Not all of them, all the time.

  There’s likley more subtle problems too. When you have overlapping MCP servers that can accomplish similar things, the agent could get confused about which tool to call. It might head down the wrong path, try something that doesn’t work, backtrack, and try something else. Every one of those steps is spending your token budget on nothing productive.

  The Usual Suspects

  Beyond sub-agents and MCP servers, there are the classic context window killers:

  • Web searches that pull back pages of irrelevant results
  • Log dumps that flood your context with thousands of lines
  • Raw command output that’s 95% noise
  • Large file reads when you only needed a few lines

  The pattern is the same every time: you need a small slice of data, but the whole thing gets loaded into your context window. You’re paying full price for information you’ll never use.

  And here’s the frustrating part — you don’t know what the relevant data is until after you’ve loaded it. It’s a classic catch-22.

  Enter Context Mode

  Somebody (Mert Köseoğlu - mksglu) built a really clever solution to this problem. It’s available as a Claude Code plugin called context-mode. The core idea is simple: keep raw data out of your context window.

  Instead of dumping command output, file contents, or web responses directly into your conversation, context-mode runs everything in a sandbox. Only a printed summary enters your actual context. The raw data gets indexed into a SQLite database with full-text search (FTS5), so you can query it later without reloading it.

  It gives Claude a handful of new tools that replace the usual chaining of bash and read calls:

  • ctx_execute — Run code in a sandbox. Only your summary enters context.
  • ctx_execute_file — Read and process a file without loading the whole thing.
  • ctx_fetch_and_index — Fetch a URL and index it for searching, instead of pulling everything into context with WebFetch.
  • ctx_search — Search previously indexed content without rerunning commands.
  • ctx_batch_execute — Run multiple commands and search them all in one call.

  There are also slash commands to check how much context you’ve saved in a session, run diagnostics, and update the plugin.

  The approach is smart. All the data lives in a SQLite FTS5 database that you can index and search, surfacing only the relevant pieces when you need them. If you’ve worked with full-text search in libSQL or Turso, you’ll appreciate how well this maps to the problem. It’s the right tool for the job.

  The benchmarks are impressive. The author reports overall context savings of around 96%. When you think about how much raw output typically gets dumped into a session, it makes sense. Most of that data was never being used anyway.

  What This Means for Your Workflow

  I think the broader lesson here is that context management is becoming a first-class concern for anyone doing serious work with agentic tools. It’s not just about having the most powerful model, it’s about using your token budget wisely so you can sustain longer, more complex sessions without hitting the wall.

  A few practical takeaways:

  • Be intentional about MCP servers. Load what you need, not everything you have.
  • Use sub-agents for heavy lifting, but recognize they cost session tokens.
  • Avoid dumping raw output into your main context whenever possible.
  • Tools like context-mode can dramatically extend how much real work you get done per session.

  We’re still early in figuring out the best practices for working with these tools. But managing your context window? That’s one of the things that separates productive sessions from frustrating ones.

  Hopefully something here saves you some tokens.

  Mar 15, 2026 / AI / Programming / Developer-tools / Claude

  Enjoyed this?

• AI-Powered Process Orchestration Across the Enterprise | Appian

  Simplify digital operations with Appian’s agentic automation platform - purpose-built for enterprise growth.

  Mar 13, 2026 / AI / links / agent / automation / platform

• How to Write a Good CLAUDE.md File

  Every time you start a new chat session with Claude Code, it’s starting from zero knowledge about your project. It doesn’t know your tech stack, your conventions, or where anything lives. A well-written CLAUDE.md file fixes that by giving Claude the context it needs before it writes a single line of code.

  This is context engineering, and your CLAUDE.md file is one of the most important pieces of it.

  Why It Matters

  Without a context file, Claude has to discover basic information about your project — what language you’re using, how the CLI works, where tests live, what your preferred patterns are. That discovery process burns tokens and time. A good CLAUDE.md front-loads that knowledge so Claude can get to work immediately.

  If you haven’t created one yet, you can generate a starter file with the /init command. Claude will analyze your project and produce a reasonable first draft. It’s a solid starting point, but you’ll want to refine it over time.

  The File Naming Problem

  If you’re working on a team where people use different tools: Cursor has its own context file, OpenAI has theirs, and Google has theirs. You can easily end up with three separate context files that all contain slightly different information about the same project. That’s a maintenance headache.

  It would be nice if Anthropic made the filename a configuration setting in settings.json, but as of now they don’t. Some tools like Cursor do let you configure the default context file, so it’s worth checking.

  My recommendation? Look at what tools people on your team are actually using and try to standardize on one file, maybe two. I’ve had good success with the symlink approach , where you pick your primary file and symlink the others to it. So if CLAUDE.md is your default, you can symlink AGENTS.md or GEMINI.md to point at the same file.

  It’s not perfect, but it beats maintaining three separate files with diverging information.

  Keep It Short

  Brevity is crucial. Your context file gets loaded into the context window every single session, so every line costs tokens. Eliminate unnecessary adjectives and adverbs. Cut the fluff.

  A general rule of thumb that Anthropic recommends is to keep your CLAUDE.md under 200 lines. If you’re over that, it’s time to trim.

  I recently went through this exercise myself. I had a bunch of Python CLI commands documented in my context file, but most of them I rarely needed Claude to know about.

  We don’t need to list every single possible command in the context file. That information is better off in a docs/ folder or your project’s documentation. Just add a line in your CLAUDE.md pointing to where that reference lives, so Claude knows where to look when it needs it.

  Maintain It Regularly

  A context file isn’t something you write once and forget about. Review it periodically. As your project evolves, sections become outdated or irrelevant. Remove them. If a section is only useful for a specific type of task, consider moving it out of the main file entirely.

  The goal is to keep only the information that’s frequently relevant. Everything else should live somewhere Claude can find it on demand, not somewhere it has to read every single time.

  Where to Put It

  Something that’s easy to miss: you can put your project-level CLAUDE.md in two places.

  • ./CLAUDE.md (project root)
  • ./.claude/CLAUDE.md (inside the .claude directory)

  A common pattern is to .gitignore the .claude/ folder. So if you don’t want to check in the context file — maybe it contains personal preferences or local paths — putting it in .claude/ is a good option.

  Rules Files for Large Projects

  If your context file is getting too large and you genuinely can’t cut more, you have another option: rules files. These go in the .claude/rules/ directory and act as supplemental context that gets loaded on demand rather than every session.

  You might have one rule file for style guidelines, another for testing conventions, and another for security requirements. This way, Claude gets the detailed context when it’s relevant without bloating the main file.

  Auto Memory: The Alternative Approach

  Something you might not be aware of is that Claude Code now has auto memory, where it automatically writes and maintains its own memory files. If you’re using Claude Code frequently and don’t want to manually maintain a context file, auto memory can be a good option.

  The key thing to know is that you should generally use one approach or the other. If you’re relying on auto memory, delete the CLAUDE.md file, and vice versa.

  Auto memory is something I’ll cover in more detail in another post, but it’s worth knowing the feature exists. Just make sure you enable it in your settings.json if you want to try it.

  Quick Checklist

  If you’re writing or revising your CLAUDE.md right now, here’s what I’d focus on:

  • Keep it under 200 lines — move detailed references to docs
  • Include your core conventions — package manager, runtime, testing approach
  • Document key architecture — how the project is structured, where things live
  • Add your preferences — things Claude should always or never do
  • Review monthly — cut what’s no longer relevant
  • Consider symlinks — if your team uses multiple AI tools
  • Use rules files — for detailed, task-specific context

  That’s All For Now. 👋

  Mar 6, 2026 / AI / Programming / Claude-code / Developer-tools

  Enjoyed this?

• Claude Code Skills vs Plugins: What's the Difference?

  If you’ve been building with Claude Code, you’ve probably seen the terms “skill,” “plugin,” and “agent” thrown around. They’re related but distinct concepts, and understanding the difference will help you build better tooling. Let’s focus on skills versus plugins since those two are the most closely related.

  Skills: Reusable Slash Commands

  Skills are user-invocable slash commands, essentially reusable prompts that run directly in your main conversation. You trigger them with /skill-name and they execute inline. They can be workflows or common tasks that are done frequently.

  Skills can live inside your .claude/skills/ folder, or they can live inside a plugin (where they’re called “commands” instead). Same concept, different home.

  The important frontmatter you should pay attention to is the allowed-tools property. This defines which tool calls the skill can access, and there are three formats you can use:

  1. Comma-separated names — Bash, Read, Grep
  2. Comma-separated with filters — Bash(gh pr view:*), Bash(gh pr diff:*)
  3. JSON array — ["Bash", "Glob", "Grep"]

  I don’t think there’s a meaningful speed difference between them? The filtered format might take slightly longer to parse if you have a huge list, but in practice it’s negligible. Pick whichever is most readable for your use case.

  The real power here is that skills can define tool calls and launch subagents. That turns a simple slash command into something that can orchestrate complex workflows.

  Plugins: The Full Package

  A plugin is a bigger container. It can bundle commands (skills), agents, hooks, and MCP servers together as a single distributable unit. Every plugin needs a .claude-plugin/plugin.json file; which is just a name, description, and author.

  Plugins are a good way to bundle agents with skills. If your workflow needs a specialized agent that gets triggered by a slash command, a plugin is a good option for that.

  Pushing the Boundaries of Standalone Skills

  However, I wanted to experiment with what’s actually possible using standalone skills, so I built upkeep. It turns out that you can bundle actual compiled binaries inside a skill directory and call them from the skill. That opens up a lot of possibilities.

  Here’s how I did it:

  • The skill has a prerequisite section that checks for a bin/ folder containing the binary
  • A workflow calls the binary, passing in the commands to run
  • Each step defines what we expect back from the binary

  You can see the full implementation in the SKILL.md file. It’s a pattern that lets you distribute real functionality, not just prompts, through the skill.

  Quick Summary

  • Skills are slash commands. Reusable prompts with tool access that run in your conversation.
  • Plugins bundle skills, agents, hooks, and MCP servers together with a plugin.json.
  • Skills are more flexible than you might expect, you can call subagents, distribute binaries, and build real workflows.

  If you’re just getting started, skills are the easier entry point. When you need to package multiple pieces together or distribute agents alongside commands, that’s when you reach for a plugin.

  Have fun building!

  Mar 5, 2026 / AI / Development / Claude-code

  Enjoyed this?

• Claude Code Now Has Two Different Security Review Tools

  If you’re using Claude Code, you might have noticed that Anthropic has been quietly building out security tooling. There are now two distinct features worth knowing about. They sound similar but do very different things, so let’s break it down.

  The /security-review Command

  Back in August 2025, Anthropic added a /security-review slash command to Claude Code. This one is focused on reviewing your current changes. Think of it as a security-aware code reviewer for your pull requests. It looks at what you’ve modified and flags potential security issues before you merge.

  It’s useful, but it’s scoped to your diff. It’s not going to crawl through your entire codebase looking for problems that have been sitting there for months.

  The New Repository-Wide Security Scanner

  Near the end of February 2026, Anthropic announced something more ambitious: a web-based tool that scans your entire repository and operates more like a security researcher than a linter. This is the thing that will help you identify and fix security issues across your entire codebase.

  First we need to look at what already exists to understand why it matters.

  SAST tools — Static Application Security Testing. SAST tools analyze your source code without executing it, looking for known vulnerability patterns. They’re great at catching things like SQL injection, hardcoded credentials, or buffer overflows based on pattern matching rules.

  If a vulnerability doesn’t match a known pattern, it slips through. SAST tools also tend to generate a lot of false positives, which means teams start ignoring the results.

  What Anthropic built is different. Instead of pattern matching, it uses Claude to actually reason about your code the way a security researcher would. It can understand context, follow data flows across files, and identify logical vulnerabilities that a rule-based scanner would never catch. Think things like:

  • Authentication bypass through unexpected code paths
  • Authorization logic that works in most cases but fails at edge cases
  • Business logic flaws that technically “work” but create security holes
  • Race conditions that only appear under specific timing

  These are the kinds of issues that usually require a human security expert to find or … real attacker.

  SAST tools aren’t going away, and you should still use them. They’re fast, they catch the common stuff, and they integrate easily into CI/CD pipelines.

  Also the new repository-wide security scanner isn’t out yet, so stick with what you got until it’s ready.

  Mar 4, 2026 / DevOps / AI / Claude-code / security

  Enjoyed this?

• Ever wanted your CLAUDE.md to automatically update from your current session before the next compact? There’s a skill for that and it’s been helpful. In case you missed it, here’s a link to the skill:

  autoskill

  Mar 3, 2026 / AI / Claude-code

• Managing Your Context Window in Claude Code

  If you’re using Claude Code, there’s a feature you should know about that gives you visibility into how your context window is being used. The /context skill breaks everything down so you can see exactly where your tokens are going.

  Here’s what it shows you:

  • System prompt – the base instructions Claude Code operates with
  • System tools – the built-in tool definitions
  • Custom agents – any specialized agents you’ve configured
  • Memory files – your CLAUDE.md files and auto-memory
  • Skills – any skills loaded into the session
  • Messages – your entire conversation history

  Messages is where you have the most control, and it’s also what grows the fastest. Every prompt you send, every response you get back, every file read, every tool output; it all shows up in your message history.

  Then there’s the free space, which is what’s left for actual work before a compaction occurs. This is the breathing room Claude Code has to think, generate responses, and use tools.

  You’ll also see a buffer amount that’s reserved for auto-compaction. You can’t use this space directly, it’s set aside so Claude Code has enough room to summarize the conversation and hand things off cleanly.

  Why This Matters

  Understanding your context usage helps you work more efficiently. A few ways to keep your context lean:

  • Start fresh sessions for new tasks instead of reusing a long-running one
  • Be intentional about file reads — only read what you need, not entire directories
  • Use sub-agents — when you delegate work to a sub-agent, it runs in its own context window instead of yours. All those file reads, tool calls, and intermediate reasoning happen over there, and you just get the result back. It’s one of the best ways to preserve your primary context for the work that actually needs it.
  • Trim your CLAUDE.md — everything in your memory files loads every session, so keep it tight

  I’ll dig into sub-agents more in a future post. For now, don’t forget about /context

  Mar 2, 2026 / AI / Claude-code / Developer-tools

  Enjoyed this?

• I published an Agentic Maturity Model on GitHub, a mental framework for thinking about and categorizing AI tools. It’s open to contributions and I’m looking for coauthors.

  Agentic Maturity Model

  Mar 1, 2026 / AI / Open-source / Agentic

• The Anthropic Hive Mind

  As you’ve probably noticed, something is happening over at Anthropic. They are a spaceship that is beginning to take off.

  Feb 20, 2026 / AI / Claude / links / anthropic

• Are We Becoming Architects or Butlers to LLMs?

  In a recent viral post , Matt Shumer declares dramatically that we’ve crossed an irreversible threshold. He asserts that the latest AI model…

  Feb 18, 2026 / AI / links / automation / reflections

• Don’t sleep on OpenClaw. There are a ton of people building with it right now who aren’t talking about it yet. The potential is real, and when those projects start surfacing, it’s going to turn heads. Sometimes the most exciting stuff happens quietly before it hits the mainstream.

  Feb 14, 2026 / AI / Open-source / Openclaw

• REPL-Driven Development Is Back (Thanks to AI)

  So you’ve heard of TDD. Maybe BDD. But have you heard of RDD?

  REPL-driven development. I think most programmers these days don’t work this way. The closest equivalent most people are familiar with is something like Python notebooks—Jupyter or Colab.

  But RDD is actually pretty old. Back in the 70s and 80s, Lisp and Smalltalk were basically built around the REPL. You’d write code, run it immediately, see the result, and iterate. The feedback loop was instant.

  Then the modern era of software happened. We moved to a file-based workflow, probably stemming from Unix, C, and Java. You write source code in files. There’s often a compilation step. You run the whole thing.

  The feedback loop got slower, more disconnected. Some languages we use today like Python, Ruby, JavaScript, PHP include a REPL, but that’s not usually how we develop. We write files, run tests, refresh browsers.

  Here’s what’s interesting: AI coding assistants are making these interactive loops relevant again.

  The new RDD is natural language as a REPL.

  Think about it. The traditional REPL loop was:

  1. Type code
  2. System evaluates it
  3. See the result
  4. Iterate

  The AI-assisted loop is almost identical:

  1. Type (or speak) your intent in natural language
  2. AI interprets and generates code
  3. AI runs it and shows you the result
  4. Iterate

  You describe what you want. The AI writes the code. It executes. You see what happened. If it’s not right, you clarify, and the loop continues.

  This feels fundamentally different from the file-based workflow most of us grew up with. You’re not thinking about which file to open, You’re thinking about what you want to happen, and you’re having a conversation until it does.

  Of course, this isn’t a perfect analogy. With a traditional REPL, you have more control. You understood exactly what was being evaluated because you wrote it.

   >>> while True:
    ...     history.repeat()
  

  Feb 13, 2026 / AI / Programming / Development

  Enjoyed this?

• I usually brainstorm spec docs using Gemini or Claude, so if you are like me, this prompt is interesting insight into your software decisions.

  Based off our previous chats and the previous documents you've helped me with, provide a detailed summary of all my software decisions and preferences when it comes to building different types of applications.

  Feb 12, 2026 / AI / Development

• Here’s a tip: if you ask Claude (via an API not Code) to Vibe a typing hacker game make sure to tell it not to return valid exploits. I asked Claude to use actual Python code snippets in the game today and… GitHub’s security scanner was not happy with me. Oopsie doopsie. Lesson learned!

  Feb 12, 2026 / AI / Claude / Coding

• Knowledge Without a Knower

  How do we define knowledge in the age of AI? Can new knowledge even be created if we’re outsourcing our thinking to the models or the systems we built around the models?

  Let’s start with what knowledge actually is. Traditionally, to know something, you have to believe it’s true and have some justification for that belief. It’s implicit knowledge earned through experience, study, or reasoning.

  AI doesn’t work that way. To the tools, it’s a probabilistic map of patterns extracted from massive amounts of text. There’s no belief, no understanding in the human sense. It’s knowledge without a knower.

  That distinction matters more than we might think.

  From Retention to Curation

  The way we work with knowledge is shifting. For centuries, the paradigm was retention: memorize facts, write things down, build personal libraries of information.

  Now we have tools that can do that for us, often better and faster than we ever could.

  So what’s our new role?

  Curation.

  The skills that matter now are about what we can retrieve, what we can verify, and what we can synthesize.

  We don’t need to remember everything, we need to know how to find it, evaluate it, and combine it in useful ways.

  The Skills We Actually Need

  If we’re not going to be the primary repositories of knowledge anymore, what should we focus on?

  Spotting bullshit. This might be the most important skill of the next decade. When the tool outputs something that doesn’t match what we know to be true, can we catch it? AI systems are confident even when they’re wrong. They don’t hedge. They don’t say “I’m not sure about this.” So we need that internal alarm that goes off when something doesn’t add up.

  Asking good questions. This has always been important, but it’s now essential. Understanding the problem means knowing where the gaps in your knowledge actually lie. A well-formed question is half the answer. An AI can give you a thousand responses, but only a good question will get you a useful one.

  Reasoning about reasoning. How did the system arrive at that answer? What steps did it take? Why does it think that’s the case? We need to be able to trace the logic, not just accept the output. This is meta-cognition applied to our tools.

  The Human in the Loop

  New knowledge will continue to need humans. Not for the grunt work of data processing or pattern matching, AI can handle that better than we ever could.

  Instead our role is to identify the anomalies. We need to become detectives, finding the errors in the data. Skepticism will be extremely valuable in the times ahead.

  Being a critical thinker. We need to be able to evaluate the evidence, weigh the pros and cons, and make informed decisions.

  In computing, we see error correcting used in the semiconductor industry, and we see a different technique also used in the quantum computing industry. And while reducing the amount of errors in a given system will continue to be important, what we really after here?

  Well, the truth, right?

  I propose we come up with a new name for truth. I think it should be called “HAT” or a “human accepted truth.”

  The aggregate of HATs is what we shall call “knowledge.” Knowledge is the sum of all human accepted truths.

  Feb 10, 2026 / AI / Knowledge / Thinking

  Enjoyed this?

• The Broken Promise of Reach

  AI is changing things we take for granted: the relationship between effort and reach.

  For years, the implicit promise of the internet was straightforward.

  Put in the work, create something valuable, and you’d find your audience. Maybe not millions, but someone.

  The effort you invested had a reasonable correlation to the impact you could achieve. A thoughtful blog post might get shared.

  A well-crafted tutorial could help thousands of developers. The work mattered because it reached people who needed it.

  That equation is no longer guaranteed.

  Now we’re in a world where AI can generate endless content at near-zero cost.

  The supply of words, images, and ideas has become functionally infinite.

  So, what happens to the value of any individual piece?

  • Your carefully researched article drowns in a database of a thousand AI-generated summaries.
  • Your authentic ideas are lost in a sea of algorithmic content designed for engagement.

  So, why put in the work if the reward isn’t there?

  The dream used to be building something sustainable or something big enough to matter. Enough to support yourself while doing work you care about. And it still can be, that, a dream.

  We need to return to the act of finding value in the act itself

  Don’t let your self-worth depend on metrics decided by a platform.

  Your entire creative output shouldn’t be measured in likes, shares, and subscriber counts. That’s when they win.

  We can’t hand over the definition of “success” to the algorithms

  The old promise of platforms will provide is broken.

  We’re not going back.

  Now we need to build something new.

  What we build is up to us.

  Feb 8, 2026 / AI / Indieweb / Publishing

  Enjoyed this?

• First Impressions of OpenAI's Codex App

  I’ve been experimenting with OpenAI’s new Codex app for engineering work that launched on February 2nd, and I’m not impressed.

  No subagents from what I can tell. It gets stuck on stuff that shouldn’t be blockers. The gpt-5.2-codex model feels slow. I don’t care about the existing skills enough to try to set one up for it.

  I did sign up for a free month trial of ChatGPT Plus though, so I’m going to give it a few more attempts before my time runs out. But so far, it doesn’t feel like a force multiplier the way Claude Code or Amp Code does. Even Open Code feels more productive.

  Maybe I’ll have better luck with Codex on the CLI? We’ll see.

  There’s something about flat-fee billing that feels so much better than watching tokens drain away. Less constrained and more open to trying new things I guess.

  I appreciate that they went through the effort of building what looks like a native Swift app instead of yet another VS Code fork.

  I think Codex desktop app is aimed at competing with Cursor’s market share, or maybe it’s some attempt at solving whatever direction Claude Desktop is heading.

  It doesn’t feel like it’s solving my problems as a developer who already has workflows that I know work.

  Back to the CLI pour moi.

  Feb 6, 2026 / AI / Developer-tools / Openai

  Enjoyed this?

• Why IndiePub Matters More Than Ever

  You might have heard the term IndiePub. It’s short for independent publishing—the practice of creating and distributing your work (books, stories, games, articles) without the financial backing or editorial control of a larger corporation.

  It’s closely related to the IndieWeb movement in software development, which has been advocating for personal ownership of digital presence for years.

  In the last two decades, we’ve watched massive platforms rise to dominance. They promised reach and convenience, and they delivered—for a while. But now we’re seeing the consequences: algorithmic timelines that bury your work, arbitrary policy changes that can wipe out years of audience building, and the degradation of platform quality as engagement-farming content floods every feed.

  This platform decay is exactly why anyone publishing content on the internet should own their own distribution.

  I’m not saying you have to stop publishing on platforms. But you shouldn’t publish there first.

  Publish your work on something you control, your own site, your own domain, and then push it out to the platforms. Your home base stays yours. The platforms get a copy.

  The AI Content Flood Changes Everything

  Now we’re facing something new, with the rise of AI-generated content, these platforms are becoming saturated with noise. Human authenticity is at a premium.

  The irony? Human authentic content is what trained these large language models in the first place.

  So there’s a reclaiming happening here. We’re taking our data back.

  Our authentic human content belongs on the platforms we control. The AI stuff can continue flowing through the corporate channels.

  It’s Never Been Easier

  I’m not going to dive deep into the complexities of book publishing or all the decisions writers face when navigating traditional vs. self-publishing. But I will say this: it’s never been easier to find a way to publish your thoughts and ideas online.

  There are definitely some better decisions you can make about where and how to publish. If you’re wondering whether you’ve made the best choice for the type of content you want to put out there, I created a guided decision-making tool to help answer those questions: blog-picker.logan.center

  ---

  IndiePub matters now more than ever because it’s our primary defense against the commoditization of creativity.

  When everything becomes content for platforms, when algorithms decide what gets seen, when AI can generate infinite variations of “good enough” content.

  Own your words. Own your distribution. The platforms are guests at your table, not the other way around.

  Feb 5, 2026 / AI / Indieweb / Publishing / Creativity

  Enjoyed this?

• Why Data Modeling Matters When Building with AI

  If you’ve started building software recently, especially if you’re leaning heavily on AI tools to help you code—here’s something that might not be obvious: data modeling matters more now than ever.

  AI is remarkably good at getting the local stuff right. Functions work. Logic flows. Tests pass. But when it comes to understanding the global architecture of your application? That’s where things get shaky.

  Without a clear data model guiding the process, you’re essentially letting the AI do whatever it thinks is best. And what the AI thinks is best isn’t always what’s best for your codebase six months from now.

  The Flag Problem

  When you don’t nail down your data structure upfront, AI tools tend to reach for flags to represent state. You end up with columns like is_draft, is_published, is_deleted, all stored as separate boolean fields.

  This seems fine at first. But add a few more flags, and suddenly you’ve got rows where is_draft = true AND is_published = true AND is_deleted = true.

  That’s an impossible state. Your code can’t handle it because it shouldn’t exist.

  Instead of multiple flags, use an enum: status: draft | published | deleted. One field. Clear states. No contradictions.

  This is just one example of why data modeling early can save you from drowning in technical debt later.

  Representation, Storage, and Retrieval

  If data modeling is about the shape of your data, data structures determine how efficiently you represent, store, and retrieve it.

  This matters because once you’ve got a lot of data, migrating from one structure to another, or switching database engines—becomes genuinely painful.

  When you’re designing a system, think about its lifetime.

  • How much data will you store monthly? Yearly?
  • How often do you need to retrieve it?
  • Does recent data need to be prioritized over historical data?
  • Will you use caches or queues for intermediate storage?

  Where AI Takes Shortcuts

  AI agents inherit our bad habits. Lists and arrays are everywhere in their training data, so they default to using them even when a set, hash map, or dictionary would perform dramatically better.

  In TypeScript, I see another pattern constantly: when the AI hits type errors, it makes everything optional.

  Problem solved, right? Except now your code is riddled with null checks and edge cases that shouldn’t exist.

  Then there’s the object-oriented problems. When building software that should use proper OOP patterns, AI often takes shortcuts in how it represents data. Those shortcuts feel fine in the moment but create maintenance nightmares down the road.

  The Prop Drilling Epidemic

  LLM providers have optimized their agents to be nimble, managing context windows so they can stay productive. That’s a good thing. But that nimbleness means the agents don’t always understand the full structure of your code.

  In TypeScript projects, this leads to prop drilling: passing the entire global application object down through nested components.

  Everything becomes tightly coupled. When you need to change the structure of an object, it’s like dropping a pebble in a pond. The ripples spread everywhere.

  You change one thing, and suddenly you’re fixing a hundred other places that all expected the old structure.

  The Takeaway

  If you’re building with AI, invest time in data modeling before you start coding. Define your data structures. Think about how your data will grow and how you’ll access it.

  The AI can help you build fast. But you still need to provide the architectural vision. That’s not something you can blindly trust the AI to handle, not yet, anyway.

  Feb 4, 2026 / AI / Programming / Software-development

  Enjoyed this?

• 2026: The Year We Stop Blaming the Tools

  Here’s a hard truth we’re going to have to face in 2026: sometimes the bottleneck isn’t the technology, it’s us.

  I’ve been thinking about how we use tools, how we find the merit in their use. We have access to increasingly powerful tools, but their value depends entirely on our understanding of them.

  A hammer is useless if you don’t know which end to hold. The same goes for AI assistants, automation frameworks, and the growing ecosystem of agentic systems.

  The rapid adoption of tools like OpenClaw’s agentic assistant tells me something important: people and companies are starting to see the real potential in building autonomous systems. Not just as toys or experiments, but as genuine productivity multipliers. That’s a shift from where we were even a year ago.

  I think 2026 will be the year we see more widespread adoption of genuinely useful tools. The Gartner hype cycle is really interesting and how it applies or doesn’t to AI adoption, but I won’t cover it here. I’d like to write more about that in future articles.

  The companies that build genuinely useful tools will be the ones that survive. They’ll be the ones that understand the value of tools and how to use them effectively. They’ll be the ones that embrace the future of work, where humans and machines work together to achieve more.

  It’s not about replacing humans. It’s about humans getting better at wielding the tools we’ve built. That’s always been how technology works. This time is no different.

  Feb 1, 2026 / AI / Tools / automation / 2026

  Enjoyed this?

• The Rise of Spec-Driven Development: A Guide to Building with AI

  Spec-driven development isn’t new. It has its own Wikipedia page and has been around longer than you might realize.

  With the explosion of AI coding assistants, this approach has found new life and we now have a growing ecosystem of tools to support it.

  The core idea is simple: instead of telling an AI “hey, build me a thing that does the boops and the beeps” then hoping it reads your mind, you front-load the thinking.

  It’s kinda obvious, with it being in the name, but in case you are wondering, here is how it works.

  The Spec-Driven Workflow

  Here’s how it typically works:

  1. Specify: Start with requirements. What do you want? How should it behave? What are the constraints?

  2. Plan: Map out the technical approach. What’s the architecture? What “stack” will you use?

  3. Task: Break the plan into atomic, actionable pieces. Create a dependency tree—this must happen before that. Define the order of operations. This is often done by the tool.

  4. Implement: You work with whatever tool to build the software from your task list. The human is (or should be) responsible for deciding when a task is completed.

  You are still a part of the process. It’s up to you to make the decisions at the beginning. It’s up to you to define the approach. And it’s up to you to decide you’re done.

  So how do you get started?

  The Tool Landscape

  The problem we have now is there is not a unified standard. The tool makers are busy building the moats to take time to agree.

  Standalone Frameworks:

  • Spec-Kit - GitHub’s own toolkit that makes “specifications executable.” It supports multiple AI agents through slash commands and emphasizes intent-driven development.

  • BMAD Method - Positions AI agents as “expert collaborators” rather than autonomous workers. Includes 21+ specialized agents for different roles like product management and architecture.

  • GSD (Get Shit Done) - A lightweight system that solves “context rot” by giving each task a fresh context window. Designed for Claude Code and similar tools.

  • OpenSpec - Adds a spec layer where humans and AI agree on requirements before coding. Each feature gets its own folder with proposals, specs, designs, and task lists.

  • Autospec - A CLI tool that outputs YAML instead of markdown, enabling programmatic validation between stages. Claims up to 80% reduction in API costs through session isolation.

  Built Into Your IDE:

  The major AI coding tools have adopted this pattern too:

  • Kiro - Amazon’s new IDE with native spec support
  • Cursor - Has a dedicated plan mode
  • Claude Code - Plan mode for safe code analysis
  • VSCode Copilot - Chat planning features
  • OpenCode - Multiple modes including planning
  • JetBrains Junie - JetBrains' AI assistant
  • Google Antigravity - Implementation planning docs
  • Gemini Conductor - Orchestration for Gemini CLI

  Memory Tools

  • Beads - Use it to manage your tasks. Works very well with your Agents in Claude Code.

  Why This Matters

  When first getting started building with AI, you might dive right in and be like “go build thing”. You keep then just iterating on a task until it falls apart once you try to do anything substantial.

  You end up playing a game of whack-a-mole, where you fix one thing and you break another. This probably sounds familiar to a lot of you from the olden times of 2 years ago when us puny humans did all the work. The point being, even the robots make mistakes.

  Another thing that you come to realize is it’s not a mind reader. It’s a prediction engine. So be predictable.

  What did we learn? With spec-driven development, you’re in charge. You are the architect. You decide. The AI just handles the details, the execution, but the AI needs structure, and so these are the method(s) to how we provide it.

  Jan 31, 2026 / AI / Programming / Tools / Development

  Enjoyed this?