How to Build Agentic Graphs
A practical guide to designing agent graphs: parallelism, conflicts, model selection, and cache management, based on my painful but real experience with graphs over 4 months and 250k LoC shipped to prod.
A practical guide to designing agent graphs: parallelism, conflicts, model selection, and cache management, based on my painful but real experience with graphs over 4 months and 250k LoC shipped to prod.
I break down how to choose between GPT-5.6 models (Luna, Terra, Sol) and reasoning levels through a simple analogy with human attention: some tasks need more "memory for rules," others need more "time to think."
Explaining Kent Workflows - a new paradigm for working with agents that lets you turn manual agent management into a fully automated process where agents work in loops.
Released Kent 2.0 - formerly Builder, now Kent with a new site at kent.sh. This release includes: system prompt customization, new subagents with config examples, performance and terminal fixes, system sleep prevention, and more.
Firefox CLI is a CLI interface for controlling a real Firefox session with AI agents - a full equivalent of Agent Browser for Chrome, but better. Install once via an extension; supports Claude Code and Codex.
MCP has fundamental architectural problems - it pollutes context, kills flexibility, and creates security risks. For agents with terminal access, CLI is a far better alternative, and it's time to stop using MCP.
Complex systems of agents, slash commands, and orchestrators aren't progress - they're micromanagement that gets in the way of frontier models doing their best work. Here's why the simple approach works better, and how I actually work.
Explanation of Kent's advanced compaction algorithm and how it lets the agent run for dozens of sessions in a row without losing context or making unnecessary stops.
Introducing Kent - a free, open-source coding agent built for professional engineers. Outperforms Codex and is already available on GitHub.
Switching to a tiling window manager became a necessity for me as the number of parallel agentic workflows kept growing. Here's a full breakdown of three options for macOS (Aerospace, Amethyst, Yabai) - pros, cons, and conclusions.
Google released Gemini 3.1 Pro with fixed tool calling, Anthropic and Google started aggressively banning subscription use outside their own tools, while OpenAI keeps actively supporting OpenClaw and is gearing up for GPT-5.3.
Spent 5 days and $150+ testing OpenClaw with different AI models. GPT is a robot without personality, Gemini has great character but can't work properly, Opus is perfect but costs $100/day. OpenClaw is still full of bugs (4000 issues, 3000 PRs).
Real case study of Android app optimization: how to reduce cold start time from 17 to 1.7 seconds, improving retention from 82% to 98.5%. Concrete solutions: migration to Splash Screen API, background task system, reactive data loading, and backend optimization.
OpenClaw (formerly ClawdBot) by Peter Steinberger is a personal AI assistant with file-system access, background agents, and remote control. This post covers its origins, hype, and my hands-on take.
I migrated 150k lines of code to AGP 9.0 and wrote a comprehensive, step-by step guide to migrating large Kotlin Multiplatform projects from AGP 8 to AGP 9, covering all breaking changes including removed build types, BuildConfig issues, NDK restrictions, and the new DSL syntax.
I found my calling as a consultant - the story of how I saved a business millions and created my best product - myself.
Agent skills are a way to progressively disclose context to AI through markdown files, which saves tokens and boosts performance. I explain why you need them, when to use them, how to create them, and debunk the myths.
In December, AI models crossed a threshold after which I became useless as a programmer. Now I push code to prod that Ive never read.
Introducing ksrc - a CLI utility that simplifies searching and viewing Kotlin library source code from 15+ commands down to two for your AI agents.
Starting in 2026, Im going independent for at least six months - will be building my own product full-time and building in public.
Im sharing how the Pulse feature in ChatGPT Pro proactively puts together a daily personalized digest for me - from Crashlytics crash fixes to Kotlin, Compose, and Jetpack updates. I explain why this is one of the most useful AI use cases for me.
Detailed comparison of Codex and Claude Code after three months of using both tools. Breaking down the pros and cons of each, model features, harnesses, pricing, and revealing who I paid $200/month to
Comparison of MVIKotlin, FlowMVI, Orbit MVI, and Ballast based on research across 70 Kotlin architecture libraries and 100+ criteria, with examples, pros, cons, and use cases for each. Includes a public spreadsheet comparing 70 Kotlin MVI and state management libraries.
SharedPreferences in Android are leading to ANRs through a hidden mechanism that Google tried to lie to you about. I completely expose them with proofs.
After a year of work, I reached 0% freezes (App Not Responding or ANRs) in my app. Sharing a complete guide with 12 tips: from removing SharedPreferences and enabling StrictMode to fighting memory leaks.
I rebuilt a legacy game engine from scratch using MVI architecture and Kotlin, 7000+ -> 400 lines, 20s load time -> to 1.75s, crashes from 8% to 0.01%, and 1700% better event processing throughput.
A comprehensive guide to making Android notifications and alarms completely reliable, based on years of real-world experience building an alarm clock app with zero notification failure reports.
A comprehensive guide on dealing with toxic family dynamics through structured conversations, exploring your emotions, and enforcing boundaries - including specific frameworks for creating accountability and knowing when its time to walk away.
An 8-question template for an emotions journal, based on 4 years of experience in Gestalt therapy. Helps fully work through and resolve uncomfortable emotions through sequential entries.
Built a Wear OS companion app in one evening, ran into terrible Java-based Google APIs and half-documented Horologist. Sharing my experience and step-by-step guide on how to avoid the main pitfalls when developing for watches.
I wont make you a reproducer. Stop asking.
In August, I switched to the Piantor Pro split keyboard with 40 keys and the Colemak DH layout. It turns out that traditional QWERTY keyboards are relics of the past, while split keyboards are much more ergonomic, portable, and efficient.
Representing events as state creates more problems than using event streams like Channel, and I explain why popular solutions often backfire in real-world applications.
Learn the correct way to load and observe data in Kotlin apps using, and avoid 4 common anti-patterns that cause hidden bugs, resource leaks, and production crashes.
Full guide on how to manage, update and represent state and avoid races in Kotlin MVI/MVVM apps using coroutines.
Anthropic published research showing that large language models are capable of detecting manipulations of their thinking process and recognizing their own intentions, which may indicate elementary forms of self-awareness.