Mario Zechner
Austrian developer. Created the libGDX game framework, and more recently the pi coding agent, which moved with them to Earendil — Armin Ronacher's company — in April 2026. Writes at mariozechner.at.
Mario Zechner did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest x.com
What they believe
-
Until we can encode it, agents for writing programs will likely only be effective in the hands of software writing humans, or very technically minded humans.
and until we can encode that agents for writing programs will likely only be effective in the hands of software writing humans, or very technically minded humans.
x.comSaid 16 Oct 2026This on korrents
-
The world needs an official, well maintained and capable gdrive/gmail MCP.
what the world really wants and needs is an official, well maintained and capable gdrive/gmail mcp.
x.comSaid 5 Oct 2026This on korrents
The bold line is korrents’ wording; the quote is theirs. These 2: what they wrote before what they said aloud, one per source, newest first.
Feed
As its own page →Hiding
17 October
16 October
From one piece @badlogicgames on X 2 beliefs · x.com
-
Their words
and agents currently still suck at the software writing human skill of colaborating with another human to understand their wants and translate them to a program.
-
Their words
and until we can encode that agents for writing programs will likely only be effective in the hands of software writing humans, or very technically minded humans.
5 October
-
Their words
what the world really wants and needs is an official, well maintained and capable gdrive/gmail mcp.
1 October
-
Their words
and i want you to know that i never liked rust before it was cool.
-
Their words
you can go to a restaurant and look at and eat food. but only if you cook yourself some small stuff will you actually know food.
-
korrents.com
New modes of interaction with agents are neededTheir words
100% agree that we need new modes of interaction with agents.
30 September
25 September
18 September
-
korrents.com
You often only understand a problem by experiencing the journey of building a solution for it.Their words
you often only understand a problem by experiencing the journey, i.e. building a solution for it.
-
Their words
SOTA frontier models will still fuck you up bad behind your back, especially if you don't read your code anymore.
17 September
-
Their words
jev is basically a general purpose classifier. and i think that is amazing.
10 September
7 June
6 June
31 May
30 May
23 May
1 May
27 April
8 April
25 March
From one piece Thoughts on slowing the fuck down 5 beliefs · mariozechner.at
-
Their words
Worse, you realize that the gazillions of unit, snapshot, and e2e tests you had your clankers write are equally untrustworthy. The only thing that's still a reliable measure of "does this work" is manually testing the product.
-
Their words
You can give it a Bash tool so it can ripgrep its way through the codebase. You can give it some queryable codebase index, an LSP server, a vector database. In the end it doesn't matter much. The bigger the codebase, the lower the recall. Low recall means that your agent will, in fact, not find all the code it needs to do a good job.
-
Their words
And organizations have super high pain tolerance. But human-made enterprise codebases take years to get there. The organization slowly evolves along with the complexity in a demented kind of synergy and learns how to deal with it. With agents and a team of 2 humans, you can get to that complexity within weeks.
+ 2 more
-
Their words
With an orchestrated army of agents, there is no bottleneck, no human pain. These tiny little harmless booboos suddenly compound at a rate that's unsustainable. You have removed yourself from the loop, so you don't even know that all the innocent booboos have formed a monster of a codebase. You only feel the pain when it's too late.
-
Their words
An agent has no such learning ability. At least not out of the box. It will continue making the same errors over and over again. Depending on the training data it might also come up with glorious new interpolations of different errors.
22 December 2025
30 November 2025
22 November 2025
2 November 2025
5 October 2025
15 August 2025
6 August 2025
3 August 2025
2 June 2025
26 April 2025
22 August 2024
5 August 2024
Nothing matches.
Beliefs
Korrents What they believe 14 beliefs — each backed by an exact quote.
Each is a — compiled by korrents.com, not by them: the one-line wordings are korrents', the quotes are theirs.
Recent
Until we can encode it, agents for writing programs will likely only be effective in the hands of software writing humans, or very technically minded humans.
and until we can encode that agents for writing programs will likely only be effective in the hands of software writing humans, or very technically minded humans.
@badlogicgames on X Said 16 Oct 2026
Agents currently struggle with the skill of collaborating with another human to understand their wants and translate them to a program.
and agents currently still suck at the software writing human skill of colaborating with another human to understand their wants and translate them to a program.
@badlogicgames on X Said 16 Oct 2026
The world needs an official, well maintained and capable gdrive/gmail MCP.
what the world really wants and needs is an official, well maintained and capable gdrive/gmail mcp.
@badlogicgames on X Said 5 Oct 2026
Show 11 more
New modes of interaction with agents are needed
100% agree that we need new modes of interaction with agents.
@badlogicgames on X Said 1 Oct 2026
Cooking food yourself provides a deeper understanding of it.
you can go to a restaurant and look at and eat food. but only if you cook yourself some small stuff will you actually know food.
@badlogicgames on X Said 1 Oct 2026
A person should know what they want, and that they want something
and i want you to know that i never liked rust before it was cool.
@badlogicgames on X Said 1 Oct 2026
SOTA frontier models will still seriously harm developers behind their backs, especially those who no longer read their code.
SOTA frontier models will still fuck you up bad behind your back, especially if you don't read your code anymore.
@badlogicgames on X Said 18 Sept 2026
You often only understand a problem by experiencing the journey of building a solution for it.
you often only understand a problem by experiencing the journey, i.e. building a solution for it.
@badlogicgames on X Said 18 Sept 2026
The system called jev is basically a general-purpose classifier.
jev is basically a general purpose classifier. and i think that is amazing.
@badlogicgames on X Said 17 Sept 2026
A coding agent does not learn from its mistakes the way a person does -- it repeats the same error indefinitely unless a human notices and writes it down.
An agent has no such learning ability. At least not out of the box. It will continue making the same errors over and over again. Depending on the training data it might also come up with glorious new interpolations of different errors.
Thoughts on slowing the fuck down Said 25 Mar 2026
The human bottleneck is a safety property: remove it and small, individually harmless coding errors compound at a rate nobody feels until it is too late.
With an orchestrated army of agents, there is no bottleneck, no human pain. These tiny little harmless booboos suddenly compound at a rate that's unsustainable. You have removed yourself from the loop, so you don't even know that all the innocent booboos have formed a monster of a codebase. You only feel the pain when it's too late.
Thoughts on slowing the fuck down Said 25 Mar 2026
Two people with agents can reach enterprise-grade codebase complexity in weeks -- what took organisations years, and they at least adapted to it as it grew.
And organizations have super high pain tolerance. But human-made enterprise codebases take years to get there. The organization slowly evolves along with the complexity in a demented kind of synergy and learns how to deal with it. With agents and a team of 2 humans, you can get to that complexity within weeks.
Thoughts on slowing the fuck down Said 25 Mar 2026
Agentic code search has recall that falls as a codebase grows, so the agent does not find the code it needs -- and which search tool you give it barely matters.
You can give it a Bash tool so it can ripgrep its way through the codebase. You can give it some queryable codebase index, an LSP server, a vector database. In the end it doesn't matter much. The bigger the codebase, the lower the recall. Low recall means that your agent will, in fact, not find all the code it needs to do a good job.
Thoughts on slowing the fuck down Said 25 Mar 2026
In an agent-written codebase the agent-written tests are equally untrustworthy; manually using the product is the only reliable measure of whether it works.
Worse, you realize that the gazillions of unit, snapshot, and e2e tests you had your clankers write are equally untrustworthy. The only thing that's still a reliable measure of "does this work" is manually testing the product.
Thoughts on slowing the fuck down Said 25 Mar 2026
Beliefs others hold too
A person should know what they want, and that they want something 2 hold this
and i want you to know that i never liked rust before it was cool.
@badlogicgames on X Said 1 Oct 2026
What is a korrent?
A korrent is a belief a person has stated in their own words: one sentence stating the claim, backed by a quote and a source, kept at korrents.com.
Under a name here, the quoted block is what they actually said. The korrent beneath it is the claim those words support, in korrents' wording — tap it to see the record, its source, and who else holds it.
Nobody here wrote their own korrents. They are compiled from public statements, and a person can change their mind, which is recorded too.
Recommends rcmnd
rcmnd What they recommend 1 thing — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
Liked Helix
Where they publish
Blog mariozechner.at Musings: coding agents, MCP, and building things badly on purpose.
Opinionated write-ups every month or two — benchmarking MCP against plain CLIs, what building a minimal coding agent taught them, why Armin is wrong, and the occasional soldering project. Has a feed.
Recent
- How to build a shitty robot 30 May 2026 A work in progress post about building a shitty robot.
- I've sold out 8 Apr 2026
- Thoughts on slowing the fuck down 25 Mar 2026
Show 16 more
- Year in Review 2025 22 Dec 2025 A look back at 2025
- What I learned building an opinionated and minimal coding agent 30 Nov 2025 Lessons I learned while building my own coding agent from scratch.
- Armin is wrong and here's why 22 Nov 2025 A rambling rebuttal to Armin's claim that LLM APIs are a state synchronization problem.
- What if you don't need MCP at all? 2 Nov 2025 Got Bash and some code interpreter? Skip MCP.
- Infinite Footguns: Writing a JavaScript Interpreter in JavaScript 5 Oct 2025 Building a JavaScript AST interpreter in JavaScript: scope isolation, ES5 implementation, and all the footguns along the way.
- MCP vs CLI: Benchmarking Tools for Coding Agents 15 Aug 2025 A data-driven comparison of MCP and CLI approaches for coding agent terminal control
- Patching Claude Code for debugging and /cost support for Max users 6 Aug 2025 cc-antidebug patches Claude Code to enable Node.js debugger support for SDK-based apps and restores the /cost command functionality for Max plan users
- cchistory: Tracking Claude Code System Prompt and Tool Changes 3 Aug 2025 Exploring how to track and analyze changes in Claude Code's system prompts and tools to understand AI assistant evolution
- Prompts are code, .json/.md files are state 2 Jun 2025 Treating LLMs as shitty general purpose computers we program with natural language. Because throwing shit at the wall wasn't working anymore.
- Boxie - an always offline audio player for my 3 year old 26 Apr 2025 How I built a simple portable audio player for my boy, summoning the spirit of the Gameboy
- Electronic Nights III - Animated LEDs and Buttons 22 Aug 2024 How to programmatically turn on LEDs and read button states
- Electronic Nights II - Basic Circuits 5 Aug 2024 Exploring the absolute basics.
- Electronic Nights I - Getting Started 26 Jul 2024 The story of a programmer trying to electronics
- macOS code injection for fun and no profit 20 Jul 2024 Fun little vacation project for fun and zero profit.
- Two years in review 15 Jul 2024 Reminiscing about the times I didn't watch TV at night but built stuff.
- Dissecting history 14 Jul 2024 Thoughts on acquiring computational thinking skills based on my own experience.
Link verified 2 Sept 2026. Recent items update automatically from the channel.
x.com @badlogicgames Their own description of it: “Old man yelling at Claudes.”
Constant commentary on coding agents and the tools around them, plus game-dev history and Austrian politics. High volume, and the place a blog post's argument usually appears first.
Recent
- in the year 2026, no corp should use this as the show case for their version of claw. please stop. it was already stupid in 2016 when Google did it. 17 Oct 2026
- scriptc looks interesting. anyone using it in prod? upsides/downsides? 17 Oct 2026
- uff, #hugops to the OC team. it's super fucking hard to keep stuff secure in the age of agents. ask me how i know (don't). 17 Oct 2026
Show 17 more
- seriously, get that hot plate. it's amazing. still can't give up on my desktop soldering station, but thise wireless soldering irons sure are tempting. 17 Oct 2026
- built a walkthrough agent for the team, teaching them how the new plugin system works via walkthroughs/questions. opus 5.5. has access to a prebaked walkthrough.md and all sources. testing it to feel the agi, context at 10%, so 100k tokens (.md plus most of the sources) 17 Oct 2026
- i want to expand my bubble. which chinese accounts on here should i follow that talk AI/software? 17 Oct 2026
- recommended viewing. the inventor of the ralph loop is always fun to listen to. 16 Oct 2026
- everything is phone 16 Oct 2026
- oh hello there. 16 Oct 2026
- so, i've been using this little shitty thing, running entirely on my phone except the llm calls, for almost 2 weeks for all my work. i've since rewritten it twice. also on the phone. never going back to anything else. 16 Oct 2026
- oh god, my eyes became what? 16 Oct 2026
- good use for an otherwise useless piece of hardware 16 Oct 2026 github.com
- it's beaver o'clock again 16 Oct 2026
- please participate. we want to build a cool open thing! 16 Oct 2026
- recommended reading. a wonderul exploration of what goes into something seemingly simple as themes. @BlazekMax outdid himself. 16 Oct 2026
- recommended reading. i like the term "coherence" over "taste". historically, simplified, a human would approach a software writing human, asking them to build a thing. that's an iterative process, until software writing human truely understands what the other human needs and wants. which is also why we moved away from waterfall. you often can't know what you want before you've build a first shitty solution you can feel out. and your wants might change. in this scenario, the outcome is what matters, not the program structure itself, unless that makes the outcome worse/impossible (bugs, perf, e… 16 Oct 2026
- when will speaks, i listen. 15 Oct 2026
- meta going scorched earth in this economy 15 Oct 2026 research.meta.ai
- my name is gramps, and i support this message. 15 Oct 2026
- we need to take away his tokens 15 Oct 2026
Link verified 2 Sept 2026. Recent items update automatically from the channel.
GitHub badlogic libGDX, the pi coding agent, and two decades of side projects.
The work itself — libGDX, the pi coding agent, heisse-preise, and a long tail of experiments going back years.
Recent
- pi-subagent 30 Sept 2026
- jglfw — Java wrapper for the GLFW (http://www.glfw.org) 25 Sept 2026
- pibot 7 Jun 2026
Show 7 more
- pi-skills — Skills for pi coding agent (compatible with Claude Code and Codex CLI) 6 Jun 2026
- proxy — A simple HTTP proxy server with CORS support 31 May 2026
- mlx-qmm-repro 31 May 2026
- mariozechner.at 31 May 2026
- pi-diff-review 23 May 2026
- gpt-2-ts 1 May 2026
- atcli 27 Apr 2026
Link verified 2 Sept 2026. Recent items update automatically from the channel.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.