iiinova — where your digital self keeps evolving

Laurynas Keturakis @laulau.land · 26/05/2026

extremely feb 2026 energy but I just published a jj workspace manager pando uses OS copy-on-write and stores workspace info in one place so I can run multiple throwaway explorations on my machine without too much ceremony dl it, rip it, rewrite it in zig: github.com/laulauland/p...

github.com

GitHub - laulauland/pando: Per-feature copy-on-write workspaces with native jj integration

Per-feature copy-on-write workspaces with native jj integration - laulauland/pando

000

Laurynas Keturakis @laulau.land · 16/04/2026

slightly nitpicking but I don't think subsidized is the correct term though? The API costs are not served at net 0 value for Anthropic to begin with

110

Reposted by Laurynas Keturakis

dan @danabra.mov · 06/04/2026

this is so painful to read, i love it github.com/stephenturne...

github.com

918222

Laurynas Keturakis @laulau.land · 05/04/2026

If anyone would be down to co-organize a reading club for this in Amsterdam area, hit me up

040

Laurynas Keturakis @laulau.land · 01/04/2026

PSA no model drops this week but the new Designing Data Intensive Applications (aka the Hog Book) is out on paperback

000

Laurynas Keturakis @laulau.land · 22/03/2026

marimo notebook making my m1 pro go brrr

a table showing training loss decreasing row by row

010

Reposted by Laurynas Keturakis

Matt Webb 🌸🌼🌸 @genmon.fyi · 20/03/2026

“Intelligent drum & bass” is such a weirdly judgemental name for a music genre

262

Laurynas Keturakis @laulau.land · 13/03/2026

vacation reading on computational literacy: finding diSessa’s framing “building tools to build intuition” especially relatable lately (book: “Changing Minds: Computers, Learning, and Literacy”)

000

Laurynas Keturakis @laulau.land · 07/03/2026

same lol bsky.app/profile/laul...

030

Laurynas Keturakis @laulau.land · 07/03/2026

I saw a phrase somewhere that really resonated: “you need to know what’s there to prompt what’s next”. I find myself learning a lot and being inspired from just reading the LLM code. It’s also fun to ask LLMs to write ast-grep rules to never write try-catch again

130

Laurynas Keturakis @laulau.land · 07/03/2026

I also liked it as a genuinely interesting speculative design for what multiagent system and a “selfdriving codebase” would look like. It’s crazy but it’s also genuine

110

Laurynas Keturakis @laulau.land · 05/03/2026

any time codex/claude makes an elaborate awk/jq/sed sequence to debug a large stream of text I say a little thank you prayer in my mind for the unix geeks who have practiced this craft long before we had the magic token machines

040

Laurynas Keturakis @laulau.land · 02/03/2026

I've had a brief Nix era in 2023/24 that I since reverted from due to the usual Nix stuff (obscure errors, one particular way of doing things, pain of making changes mid-flight) I am now very tempted to return to it - most of the stuff that was a pain to deal as a human dev is excellent for agents

000

Laurynas Keturakis @laulau.land · 01/03/2026

with all the buzz around harness engineering and self-driving codebases, i keep returning to this text: petevilter.me/post/datalog...

petevilter.me

Codebase as Database: Turning the IDE Inside Out with Datalog | Pete Vilter

Introspectable and extensible IDEs with logic programming

0201

Laurynas Keturakis @laulau.land · 21/02/2026

everyone’s exploring different agent coordination messaging platforms when all we should do is get them to send diff patches over email git send-email exists for a reason

010

Laurynas Keturakis @laulau.land · 19/02/2026

this is very cool! Did you use TransformeLens/SAELens for any part of your work?

100

Reposted by Laurynas Keturakis

Szymon Kaliski @szymonkaliski.com · 12/01/2026

prototyping co-drawing with Gemini Flash 3 at Google in these demos "thinking" is disabled, which makes the model return tokens very quickly (all videos are realtime), and I find these rapid responses pretty good for the use-cases I'm experimenting with, like: executing simple diagrams ...

612619

Laurynas Keturakis @laulau.land · 15/02/2026

sunday evening stream, hacking on pi extensions, trying this out for the first time hiiii www.twitch.tv/laulau_land

twitch.tv

laulau_land - Twitch

Working on pi extensions

010

Laurynas Keturakis @laulau.land · 10/02/2026

ast-grep scan will continue until morale (and codeslop) improves

000

Laurynas Keturakis @laulau.land · 06/02/2026

just released a small update to jj-hunk - programmatic hunk selector for your jj-using agent New: -r flag. Split any revision, not just your working copy. support for YAML specs stable hunk IDs you can install it with: cargo install jj-hunk npx skills add laulauland/jj-hunk

000

Laurynas Keturakis @laulau.land · 28/01/2026

fine, here's another one: shell completions for running ! comands inside prompts. Hooks into your shell completions directly (unless you're on zsh, which idk why you'd be there) npmjs.com/pi-shell-com...

npmjs.com

000

Laurynas Keturakis @laulau.land · 28/01/2026

2. pi-threads is simpler - just a direct rip from one of Amp's features that enables pi to explore previous conversation threads www.npmjs.com/package/pi-...

100

Laurynas Keturakis @laulau.land · 28/01/2026

for longer conversations I noticed this method retains a lot more detail in my compaction summaries so I don't fear that dreaded context threshold anymore I also can tune it a bit more easily by prescribing more ways to explore the conversation with the bash tools

100

Laurynas Keturakis @laulau.land · 28/01/2026

pi (a very good and very shitty coding agent from @mariozechner.at) has a very powerful extension system that I already raved about. Now that it supports npm/git-based installs I published a couple of mine that I've been using daily

140

Laurynas Keturakis @laulau.land · 27/01/2026

Java still holds software development in its iron grip

120

Laurynas Keturakis @laulau.land · 27/01/2026

<niche-CI-content-alert> wild that the most common standardized format for reporting test results is a *loose* standard JUnit XML really hoping this changes in 2026 now that we're mostly focused on reviewing/reporting on code </niche-CI-content-alert>

130

Laurynas Keturakis @laulau.land · 25/01/2026

and of course it has the ace jump functionality too

000

Laurynas Keturakis @laulau.land · 25/01/2026

added revision search to Tatami that supports search by change id, keywords in messages, and also niche but fun stuff like revset queries (you can query "pol-" to go to the parent of "pol" revision)

100

Laurynas Keturakis @laulau.land · 18/01/2026

getting to the point where I'm really liking the experience of Tatami - a GUI I'm building for jj - uses jjui/vim-style keybindings for navigation - has overview and detail screens - fast! Uses @tanstack.com DB to interface via @tauri.app IPC - uses diffs.com of course

110

Laurynas Keturakis @laulau.land · 11/01/2026

still figuring out edge cases and tweaking the skill file but it's been great for asking Claude to "please build me a narrative of all the changes you've made that makes sense for me"

100

Laurynas Keturakis @laulau.land · 11/01/2026

$ jj-hunk list {"src/lib.ts": [{"index": 0, ...}, {"index": 1, ...}, {"index": 2, ...}]} $ jj-hunk split '{"files": {"src/lib.ts": {"hunks": [0, 2]}}, "default": "reset"}' "refactor" Hunks 0 and 2 go into the commit, hunk 1 stays in working copy.

100

Laurynas Keturakis @laulau.land · 11/01/2026

with jj-hunk you can list all of the file changes per hunk in json and then express the changes you want committed or left in working copy github.com/laulauland/j...

github.com

GitHub - laulauland/jj-hunk: Programmatic hunk distribution in jj

Programmatic hunk distribution in jj. Contribute to laulauland/jj-hunk development by creating an account on GitHub.

100

Laurynas Keturakis @laulau.land · 11/01/2026

needed a tool to make it easier for agents to reorganize commits and hunks into logical narratives that make it easier to review code jj split is great but requires interactive TUI that's still a tricky territory for Claude so here's jj-hunk - programmatic hunk selection for jj

110

Laurynas Keturakis @laulau.land · 04/01/2026

I wrote up how my workflows with agents and coding evolved over 2025 gearing up for the first week of work already but if you're still in the mood to read blog posts - here you go: laulau.land/working-with...

laulau.land

how I worked with agents in 2025

2025 has transformed how I work as an engineer. Many previous assumptions I had about my profession have gone up in the air yet some have remained. What follows...

010

Laurynas Keturakis @laulau.land · 04/12/2025

it's atuin wrapped time! story of my 2025 shell is that I adopted jj halfway through the year (you can see lg - lazygit - decline) and mostly switched projects from pnpm (shorthand: pn) to bun

030

Laurynas Keturakis @laulau.land · 07/10/2025

brb adding a claude code hook that exits with error 2 the moment it outputs "comprehensive"

100

Laurynas Keturakis @laulau.land · 07/10/2025

overall feels like a Good Thing, I'm glad that they chose MCP to power this and I feel more certain than ever that TypeScript is the best tool to build MCPs - the DX of linking tool outputs and getting the typed data in UI is unbeatable lots to do for mcp-lite!

020

Laurynas Keturakis @laulau.land · 07/10/2025

you need to link resources with your UI components and you need to wire up your resources with tools - it's odd that they chose to link the latter using a special _meta field rather than MCP embedded/linked resources but maybe I'm missing something

The image shows a code snippet written in JavaScript (or TypeScript). It defines a tool registration using server.registerTool with the name "pizza-albums". The tool has a title "Show Pizza Album" and metadata (_meta) that includes an openai/outputTemplate field pointing to "ui://widget/pizza-albums.html", along with two other fields for invocation messages: "Hand–tossing an album" and "Served a fresh album". The tool expects an input schema with one field pizzaTopping of type string. The async function returns an object with content (an array containing a text message: "Rendered a pizza album!") and an empty structuredContent object.

A highlighted line draws attention to the "openai/outputTemplate": "ui://widget/pizza-albums.html", entry.

130

Laurynas Keturakis @laulau.land · 07/10/2025

ok so raw thoughts while exploring OpenAI's new Apps SDK the difficult thing right now is with the existing tooling you have to run two servers: one to serve your assets, one to drive the MCP server logic

120

Laurynas Keturakis @laulau.land · 06/09/2025

ok so I accidentally read this as “personality discord design hackathon”

110

Laurynas Keturakis @laulau.land · 05/09/2025

launched mcp-lite today: github.com/fiberplane/mcp it's a lighter, minimal library to create your MCP servers in TypeScript that can work on any runtimes and can work with any Standard Schema vendors

github.com

GitHub - fiberplane/mcp: Lightweight, composable MCP framework for TypeScript

Lightweight, composable MCP framework for TypeScript - fiberplane/mcp

282

Laurynas Keturakis @laulau.land · 02/09/2025

if you're publishing an npm library today - what is the standard to inform coding agents about its API (is there)? Something like AGENTS.md but that you can embed in the package.

100

Laurynas Keturakis @laulau.land · 20/08/2025

my favorite place on the internet for at least 15 years straight

010

Laurynas Keturakis @laulau.land · 15/08/2025

tbh I also could probably just as easily work in git as I have for years. What converted me: - fearless rebase and context switching between branches (no stage/stash/pop dance) - fake-merge unmerged branches to test if multiple changes work together well - jj undo and op log if I f things up

140