I've sold out
mariozechner.at
mariozechner.at
This was a solid letter to the fans. I get why it’s disappointing to some, but it sounds like it was the right move for them personally. For people who’ve earned that kind of credibility, I say congrats on the move.
If I can't tell what your company is/does without clicking/interacting with the site, what are you even doing?
My other least favorite style is where stuff flies in from the side as you scroll down.
I'm not sure if this library or if the AI tool itself really matter all that much, but a lot of people seem to think that they do, or will anyways, which is probably why this post has no context. Because the author assumes that both the library and the tool are well known, but I think they're only well known by those in the AI world.
You know how it is. You only know what you know, and it's hard to remember what it was like when you didn't know. When you're around people who talk about the same things all the time, it's easy to just assume that's what's everyone's talking about.
There are so many of these little tools coming and going these days it's hard to keep track. Some of them might end up mattering, but most of them probably won't.
I switched a few months ago and have not looked back. Unfortunately Anthropic blocked access from Claude Subscription users today, but that's a different story.
> Is there any point in me even trying Pi, or should I just stick with Claude Code?
I'd say stick with Claude Code. I'm a minimalist and hate everything about Claude Code!
So what makes Pi better than CC? Is it better than OpenCode?
My (oversimplified) summary: it's like vim versus an IDE. Good for tinkerers and obsessives who like small, sharp and customisable tools.
Is it better than OpenCode? It's certainly much smaller and doesn't have a client-server architecture - already that is a big win.
I am not usually super privacy minded but if you already made it nonobvious this is happening I don’t really trust the underlying tool.
https://opencode.ai/docs/zen/#privacy
Above is the link. The front page says your privacy is important and says they don’t do training on your data except for the following exceptions which links to this page. Then even their own model is training on your data except there is no opt out. So if you pay for zen and you select one of these models in the tool you have no clue it’s auto training on your data.
One of the ideas I like about opencode is the ability to prompt and such from a web browser. So I'm curious if that is the client-server architecture you are talking about, or if it's something else.
For reference, I used replit for some vibe-ish coding for a little bit and really liked that I could easily prompt and view output on my phone when hanging out away from my computer. Or while waiting at the airport for example.
(RIP OG replit by the way. They've pretty much completely pivoted from a REPL to AI, which is pretty hilarious to me given the company name xD)
Long term I’m bullish on an open source harness “winning” the foot race, in a similar way that Linux “won” over Windows and MacOS (that is, debatably)
To put another way, how much inferior can the model be with a superior harness to achieve a similar result?
I am not a web developer, I don't need npm and I don't want it clobbering my /usr (which is immutable on many modern distro's anyway). Doesn't exactly inspire confidence in the project to me.
oh-my-pi's installer installs a bun bundled binary in my users .local folder. That's much more user-friendly.
The price of flexibility is, pi is not opinionated about adding sandboxing out-of-the-box, it gives you options on how you want to do it. You either do it with linux containers, with a dedicated VM, or just bubblewrap. It is nice that it gives you a way to hook into it in a very easy way though.
Put this into `~/.npmrc`:
```
prefix = ~/.local
```
Bam, `npm install -g` contstrained to your `~/.local`
Can you please give us concrete reasons that make Pi better than Claude Code for you? And what model are you using for it since Claude is not available.
I'm not the person you're replying to, but I value good architecture and elegance for aesthetic and maintainability reasons, the MIT license for future-proofness.
Claude Code has been steadily getting worse, it's a steaming pile of vibe-coded bloat and I don't foresee good future for the CC agent. Why would I use it?
There are many people who choose their tools using more criteria than just "how good is the immediate output now".
Claude Code is a steaming pile of vibe-coded bloat.
Pi has been elegantly designed with a minimalist mindset.
> It has about half the relevancy of "MCP", which is such a similar concept where hanger ons try to catch a steam train.
I'm not even sure what you're saying here. Are you saying MCP and Pi are similar concepts? Or that both are hanger-ons? I believe the steam train of Claude Code is heading for a wall and Pi has no intention to catch it.
Not disagreeing but I'm curious how much it really matters.
I use OpenCode which is similarly a steaming pile of vibe-coded bloat and get pretty good results!
I'm open to switching to pi, which I've read about occasionally on HN, but I'm struggling to see where it might have practical benefits versus OpenCode.
The idea that pi is now under corporate governance with a potentially sustainable business model does intrigue me, but there's risk in adopting "open core" software too due to the potential rug pull and incentives to gate important features behind more restrictive licenses.
What agent do you use then? To me it seems most are terrible, incl Claude Code, OpenCode, and the other usual ones. I never felt an urge to use one. At best I used one very reluctantly as a magical black box.
OLLAMA_CONTEXT_LENGTH=16384 ollama launch pi --model qwen3.5:9b
To get close to a happy experience with Claude Code or OpenCode with small local models I had to use 28B to 32B local models.
For those that don't know Armin, he's the creator of flask among many things and has some cool projects like around like gondolin.
As for pi, you should absolutely try it out if you've used opencode/continue-dev or even if you've used the less open cloud vendor coding harnesses like Claude/Codex.
I do agree that they write in a way that requires context.
Pi itself is a minimalist coding harness with a tiny 1500 token system message and only read, edit and bash as tools. I only discovered it a few weeks ago and it is surprisingly powerful with a local Qwen3.5-35B - especially as it allows to keep the context low.
Mario's blog posts are not easily digestible (imho) until you have read a few of them but they have plenty of profound thinking. His blog is for me the first one in years where I have spent an hour to read several posts.
Mario is deeply rooted in the OSS system and basically that is what he is talking about here in this post. That said, I have no idea what earendil is doing, except that it is based on pi.
Edit: My personal take - "I've sold out" is very much Austrian style because actually it is the opposite. To quote one thing from the post:
"Then Miguel and Nat approached us. Long story short: we sold RoboVM to Xamarin. A short while later Xamarin closed-sourced our open-source RoboVM core, quickly followed by Xamarin selling to Microsoft. Then Microsoft shut down RoboVM immediately.
While there was some monetary gain, everything about this fucking sucked."
So Mario did a lot of vetting to hopefully avoid this from happening again.
I haven't kept up with the various claw stuff, but isn't openclaw the one under MIT on github? What does "made public" mean in this context?
I just want to retire in the Shire away from this AI non-sense (no jabs, just mild burnout).
You need hyper-obsessive nerds (the kind who were hyper-obsessive in school about D&D and LotR) to bring it to fruition.
The non-LotR namers give up early.
> Despite its Tolkien-inspired name, Earendil is not a tech company with fascist tendencies. Quite the opposite. They are basically well-meaning hippies in my book
Anything else I missed?
E.g. the "Palantíri" were devices for long-distance communication and observation from an ancient civilization - which eventually got captured and corrupted by Sauron and turned into mind control / propaganda devices.
The depiction of the One Ring, is of course well known.
Tolkien himself also had some personal experiences with landscape destruction through industrialisation and experienced WWI, the "first industrialized war" firsthand, which I think gave him a profoundly negative outlook on technology.
Given all this, I find it sort of interesting that so many "Techies" are diehard fans of him, not to mention even naming startups after his works.
Torment Nexus goes again...
(There is probably some hypocrisy in his works, in the sense that only "modern" or "magical" technology gets that treatment. Everything at medieval or pre-medieval level is fine)
I was prototyping something with pi under the hood for a personal project, going to switch off it now.
For what it's worth, it's pretty straightforward to recreate it I found, at least it's base idea. Readline w/ nice output is a bit of a pain, but still, doable, and if you don't care about that part of it, then the overall agent loop that you'd build on top of? You could build it, I promise.
EDIT: here is a link of what I wrote just for my own use: https://github.com/mark-watson/coding-agent
https://github.com/girvo/girvent this is my silly one :)
Create a private repo in GitHub first, then do a bare Git clone of https://github.com/badlogic/pi-mono.git (ideally do it before the original repo gets moved to Earendil's GitHub org).
git clone --bare https://github.com/badlogic/pi-mono.git
git push --mirror <url of your private repo>
git remote add upstream https://github.com/badlogic/pi-mono.git
I can already see in https://github.com/badlogic/pi-mono/commit/6d2d03dcc9a39e60c... that the Earendil announcement will be popping up in the next released version of Pi. Even has a dumb pic of Mario, Armin, and I presume Colin, which will be displayed in Pi: https://github.com/badlogic/pi-mono/blob/6d2d03dcc9a39e60c37...
With that little how-to guide out of the way, here's what I think:
Mario is free to do whatever and not give a shit about what the internet at large thinks of him. By that metric, he's doing a hell of a job with that rambling blog post. Likewise, I'm also free to mostly concur with the internet at large (https://news.ycombinator.com/item?id=47688794) and prepare simple mitigations like above that can blunt this to a certain degree. Let's just hope that Mario and Armin don't take the "flicker company" approach (his derogatory term for Anthropic) and DMCA the shit out of any private repos.
Shameless plug, yes, but it's free, so I think it's fair :)
And if they mess it up, it’ll be easy enough to fork.
I've seen some negative sentiment in this thread which I don't really understand if you read the article.
Apparently it’s possible to give it access to much of your Google account:
I didn’t see a pricing page, but there is this:
> Lefos uses a credit-based billing system. New accounts receive a limited number of starter credits at no cost. Usage of AI features consumes credits.
> When your credits run out, you can subscribe to a paid plan to receive additional credits each billing cycle. Subscriptions are processed through Polar, our billing provider. You can manage or cancel your subscription at any time from your account settings.
Tolkien stole the name from an Anglo Saxon poem, with a somewhat unclear meaning. His version of the story is woven through lots of his work, from the very earliest days. You could view all of Middle-earth as elaborating on his attempt to assign a meaning to it.
I have been working with pi-mono locally for a few months now. Great code base to study. Much higher quality than CC. (I have posted a gist analysis before.)
Will keep an eye on the work of these talented engineers and entrepreneurs. Good luck guys!
> Before Blitzy starts any work on code generation, the platform launches collaborative agents to deeply analyze the repository – mapping dependencies, understanding conventions, and capturing domain logic. This documentation process can take hours or days. When prompted to add a feature, refactor code or fix bugs, Blitzy replies with a highly detailed technical specification.
I don't see evidence that the harness -- rather than the approach to information indexing and agent tooling -- makes much of a difference.
You can make a case "this harness bakes X in" (or in the case of pi "this harness bakes nothing in; you choose your own adventure"), but at the end of the day, skills are just markdown files and CLIs and shell scripts can be used by any harness; they are portable. CC allows override of the system prompt[0] and I would guess most harnesses have similar facilities. I don't see how the harness is going to be the bigger impact versus the configured tooling (skills, scripts, plugins).
The extraordinary claim here is that if I configured pi and CC, Codex, etc. with the same system prompt, same tools, same skills, that pi would outperform CC, Codex. That's what it means to say the harness matters. That just doesn't seem right; rather its the configuration of tools, skills, and default prompt that matters.
[0] https://code.claude.com/docs/en/cli-reference#system-prompt-...
I should've given original links instead of noisy HN threads.
[1] https://github.com/badlogic/pi-mono/blob/main/packages/codin...
[2] https://mariozechner.at/posts/2025-11-30-pi-coding-agent/
pi-conding-agent itself is not a product yet and it won't have much value to end users other than those using the products built on top of it.
Here is a session of pi analyzing coding-agent package itself.
https://ontouchstart.github.io/.pi/agent/sessions/--pi-mono-...
It was created via a non-interactive CLI command in a docker container connected to a local llama.cpp server with a very limited model. $0 token cost.
However, the worst ones are the ones who preach to others about never selling out whilst telling others that they are there for the long-term publicly, but will privately sell out the minute the offer is greater and exclusive to them.
The point here is not make such promises. People who preach one thing and then are tested to do another are the ones to avoid.
Ah haha so armin is also an Internet troll, right? For example if I said, one thing that sets the "Concorde apart from other commercial airliners, ..." I am saying the Concorde is a commercial airliner.
> Colin is pretty OK for an ex-finance dude
Making it look like the author disapproves of finance dudes but finds Colin tolerable yet still within the expected parameters of badness.
The author favors the “public benefit” part, but less so the “venture-backed for-profit” part, is my speculation. Likely the reason for the disapproval of the finance types and the title “I’ve sold out”.
tbh. it sounds a bit like he himself was somewhat of an internet troll during that time
and it might not be quite the same definition of "internet troll" I tend to us
like it sounds a lot like a definition of "toll" like
- "very vocal, convicted of their opinion, non stop discussing(/neutral) potentially for the sake of discussing(/neutral)"
while I tend to associate more something like
- "intentional annoying, non stop discussing for the sake of annoying people, often using dishonest discussion techniques, potentially outright harassment"
OpenClaw (the software project) is one of the worst messes I’ve ever seen in any git repository. Security, capabilities, simplicity, and they chose at most one.
May be overthinking this but I'm geniunely curious how that was done practically speaking? Standard git branch and merge or everyone sitting around deciding what prompt to write?
https://github.com/amantus-ai/vibetunnel/commits/main/?after...
But who am I to judge? If I made a project as popular as pi I'd sell out so fast.
If Mario had instead just written: "I'm excited to work on the future of PI at Palantir." and nothing else, clearly this is worse?
> The fact I can't understand what Earendil is
From the article:
> I learned a few things. My European brain thinks pi is just another small, mildly useful OSS project of mine with no commercial value. My peers in the space seem to think it has properties that make it stand out over the alternatives. VCs and big corps seem to think that pi has commercial value. Some demonstrated their conviction by sending term sheets or "dream job" offers.
Mario thinks there's no commercial value, but clearly some VC people do. If you have the VC money, but don't have any ideas for how to justify commercial viability, why bother? If someone just handed me a big bag of cash and told you to just "do something", I could imagine coming up with Earndil. Make a nice landing page, write some nice guiding principles, and figure out the AI stuff later.
Yeah, but mainly just because what Palantir is.
It would be better if he just told us what part of the future pi tool chain would be proprietary, but from the article it seems that they haven't even decided on that, so I don't know what to make of this.
Same, but I heard about Armin Ronacher, and the things I heard were not exactly terrible. He's the creator of the Flask microframework, he's contributed to open source more than most people. Perhaps it'll be fine!
And, as Mario has written about ten times in his post: Pi is MIT licensed and it's not hard to find the "fork" button on GitHub.
Their intention to release products that provide a return for their backers, using a “free with paid premium” model.
Congrats, and I'm rooting for this to work.
> Over coffee, Armin and I found we had more in common than we thought. Not only politically, but also in the way we think about software, and OSS specifically. In my recollection, we became actual friends that day, even if we didn't meet again for many years in real life.
This is probably the "highest value" part of that whole blog post, as this is something I see so frequently in the real world. People don't give others a chance to just be humans with conflicting views and opinions, and instead try to assign a label to the other part as quickly as possible, then they apply the same "rules of engagement" with that person, as anyone else who deserved to get that label. But once you forgo it, you realize how much you have in common even with your "worst enemies".
It goes both ways too, and the more closed up you are in that regard, others will treat you the same. But you miss out on so many interesting thoughts, ideas and conversations, when you limit yourself to labels and not realizing how diverse and interesting even the most "boring person" could be, given the right questions and the right conversation underlay.
<div class="pointer-events-none fixed inset-0 z-[30] transform-gpu mix-blend-normal" style="background-repeat: repeat; background-image: url("data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAQAAAAEACAYAAABccqhmAAAQAElEQVR4AZTa4ZocV3Nja4V/zf3fsE+/mVxVUZtZTR8/3yIQAHY2JXs0kuz/+d///d//F//999//QzftPlUHOfh4us/MtoxCFm6407wb3RQyPPkyCjvkU1nITnRlPNwpH7In6qmegg833PQ3bDZtd8bLt/Kb+jI3nm556POp7BtPG5l9+uR3p4cs3HD/pnV2Ufaksie8ldM4b7kMPE7vhg58PN2yOHc7z6e2eer+n5n50fd/forXMTP/dW+dud/M3H3dzJ2/PvBjZu7MZmb+m5mf9P6P7Hb3rzN3N3Prnd6/zryz3s3cP3/m3VnP3Dl/MjPX7+HpG2XebO+eeb+bGdGLmc+fNzMfP8O38HpwmJn5SM7tzLuf+du3T/uYG93pzP371W1m7rwdnbkzu33nZ+5+Zq4/5plRXf9z05sr+PnF..[trimmed for length]...MuWCoAAAAASUVORK5CYII="); background-size: 256px 256px;"></div>
I contribute to open source projects, but none of them to date could support me buying much more than a beer. If one took off such that I could "live" off it, I would be happy to leave my current job and dive all in. Until then, I just keep plodding along.
you can figure out the fingerprinting today, but if they change it tomorrow and wait 5 months to force update everyone, they will catch you and ban
easy to circumvent
but would fall under "circumventing security protections"/"hacking their API"/etc. And due to the sometimes very unreasonable laws the US has in that area they can use that to go after anyone providing a workaround.
Through that maybe won't work well for the EU, I'm not sure how much the laws have been undermined in recent years but we had laws which made it explicitly legal to circumvent DRM iff it's for the sake of producing compatibility (with some caveats).
The Lefos site immediately throws up a sign in page.
The Earendil site just talks about values.
Earendil seems like some startup focused on open-source founded by a bunch of famous open-source devs. Though I am not sure if they're VC funded, but various mentions of "buying out" make me think that they are.
Just take the negative reactions, (if any) as warnings. They have seen it all before like you have:
+ VC funded
+ LoTR name (really?)
+ “We are not evil I promise”
And the biggest red flag:
> None of the early-stage investors are on my naughty list, quite the opposite.
For now.
Lol, the attitude. 180 degrees from the community that gave him everything he has. Classic.
I run a tiny thing that a lot of people like. Its licence is permissive: if I 'sold out', nobody who is using it today would have to change a damned thing.
And you know what? Call me if you wanna buy it. Because here's my priority list:
1. Looking after my family.
2. Looking after yours.
There's nothing wrong with making money, I'm 100% all-in on capitalism.
It's just the sudden change in attitude, particularly when people jump to the polar opposite they've preached their whole lives. Some say it's wise to change your mind, I just find it repulsive.
The same biological impulse fuels the rampant nepotism and open corruption we seem to selectively whinge about online these days.
> Earendil is a public benefit corporation
Ah, so like OpenAI then.
> But I've also learned what I do not want. I do not want to build my own company around pi. We have a four-year-old kid. I want to watch and help him grow up as best as I can. This is, first and foremost, what I want. Everything else is secondary to that. In the past 2 months, he cried a lot because "daddy isn't here". I never ever want to experience that again.
That's completely fair. And above it, you sketch exactly what you could've done instead to solve this:
> I mostly handed over the reins to a beautiful team of core contributors in 2016, who to this day keep the project well maintained. I never commercialized libGDX, unless you consider it commercialization to build a proprietary piece of software like Spine on top of it.
Sounds like the above worked great, and this would've been the obvious option to do once again.
> part of me wants to take this further. That includes building a team. It also includes commercialization to feed the team, done in a way that doesn't repeat the shit I lived through with RoboVM.
So this is the only part that really answers the "why" - you want to earn a living from working on Pi, and presumably (?) believe you can't achieve that with OSS. I think you're wrong, and belong to the 0.001% of OSS projects that can earn a very nice living from working on Pi without taking this step. I'm not exaggerating, I would fully agree that the number of OSS projects where this is possible is exceedingly small. But this is one of them. If you don't believe that then fair enough, I guess, though I'm curious why you believe that. Because there clearly exist a good number of OSS projects that make their lead developer a very comfortable living. Which condition do those projects satisfy that Pi doesn't?
If that's not it, then the article doesn't really answer the "why" despite lots of text that appears to do so.
You obviously owe me, or anyone really, nothing. So far all you've done is contribute for free. But if you're going to write this kind of article to clearly do a little bit of soul searching, assuaging fears and "make things public" to stop them weighing on your mind, then it looks better to go all the way and state things in plain terms.
Because he wants to have tons of money in the bank and get up every day to go do anything he wants with his four-year-old kid when he wakes up. Building something in a couple months that got popular enough to be able to this is amazing. Not check GitHub issues and PRs from ClawdBot5000 and hope he'll make some money from it in 5 years.
We're entering a new world of software development that allows people with no coding experience to do things like build their own mobile apps, websites, and all sorts of other things that were previously reserved for only technical people. The gap is closing quickly. At this point, if you're a "real" developer and have an app that somebody wants to purchase, you'd be crazy not to take the offer if it's a good amount and have cash in the bank that would allow you to do your own stuff in the future without worrying about $dayjob.
>Ah, so like OpenAI then.
What kind of reasoning is this? The sort that can respond to.
"I'm a doctor"
With
"Ah, like Harold Shipman"
How is this sort of framing good for anyone?
He already has a GitHub sponsorship with lots of sponsors himself. The problem is they get to set the price and the minimum is at least $1. This tells me that he believes he is not earning enough from OSS sponsors.
So with that, how do you even start with answering this:
> It also includes commercialization to feed the team.
The author knows that in his case, OSS does not pay the bills and certainly won't pay for his staff.
It has just become impossible to believe this kind of thing unless IMO the post is even more transparent than this. That was my point, if you really do mean the best and want to show that to the world, then even this (while much more transparent than anything with investor involvement) I don't think reaches the bar. So that's something I think that could've been done to demonstrate that this was a good call.
If you can demonstrate a few years from now that my disappointment was unwarranted, then everyone's a winner :)
Mario maintains the project pro bono, which is only possible because he had an exit a couple of years ago. I believe it is a lot of work making sure that the quality of the codebase doesn't detoriate with the onslaught of slop that is hitting open source projects right now.
Additionally, I have massive respect for Armin and I don't believe Earandil is your typical VC startup that wants to grow no matter what.
Yet another bait and switch plays out - build clout on the goodwill of open source and execute the plan at the opportune moment.
I may understand the decision but I sure as hell don't respect it - cashing out and prioritizing your family, a very human thing to do, at a glance it almost even seems noble.
Palantir, Anduril industries and Mithril capital companies are not.
Well, perhaps they were right. Naturally when it comes to open source, people don't have any control over what others do with their own time. If they do not contribute financially (or their own time into improvement a project) then even less so. Still, it is a decision that has been made because of prioritising one's own personal goals. This is fine; but to expect that others share the same 1:1 opinion is not logical.
> Like me, he thinks open-source and open protocols are a necessity, not just lipstick on a corporate pig.
I think the description is also painting lipstick on a pig. I've seen too many who promote open source but then sell out suddenly. Github? And what is the influence Shopify is doing in the ruby ecosystem? But anyway, that is all their own personal thing. To assume any community needs to share those personal success stories ... it makes no sense.
> Earendil's products are built on top of pi.
Ok, so ... that is lipstick. Aka promo. I don't understand why he critisizes others but then does the same himself. Which is fine; I just don't get the assessment he is doing.
> Despite its Tolkien-inspired name, Earendil is not a tech company with fascist tendencies.
Here he refers to Palantir clearly. Thiel is abusing Tolkien IMO. Or he sees himself as Sauron or whatever. But he is more a clown Sauron, just like his mad orange king is.
> who think software, and specifically AI, should serve humans, not the other way around.
Does AI serve humans? Or does it serve those who control it?
> Finally, and most importantly to me, almost everyone on the team has kids.
So what? I mean, many people who are not so well in the head, had kids. I am not saying that refers to the blog author here, mind you - I refer to the "my criterium is that all have kids". Pity on those fools who don't have kids then?
> pi is owned by Earendil, the company.
Ultimately people will derive value from it, or not; but it is clearly a private project. Even if open source, we can see that with chrome + Google. Google makes most decisions. Yes, you can build on top of it; I use thorium right now, for instance. But I am not fooled one second who effectively controls a project here.
Best luck of success to him.
I know people hate LLM writing but at least it gets to the point and gives some background context. I have no idea what this guy is talking about. It's an article aimed at people who know him personally, it seems.