Maximilian Fellner

Software Engineer and Tech Lead building full-stack products. Opinions are my own.

[Maximilian Fellner @mxfellner]

That’s some Three Body Problem shit right there

[elie @eliebakouch]

mythos will be bad ON PURPOSE on ai "frontier llm research" tasks, this is very very sad for the research community

also the fact that this is un purpose not visible to the user is crazy

[Maximilian Fellner @mxfellner]

Like this guy pointed out a while ago that game engines, especially UE, use outdated and incorrect lighting models that make everything look like plastic. DLSS5 just fixes that… https://t.co/lxp69Z8FMV

[BadRobo82 @BRobo82]

Now Digital Foundry is going to have to pretend that the plastic crap on the left looks better than the screenshot on the right. It’s a collective hallucination fueled by AI hate. The hate, which I partly understand, but I'm not going to lie to myself about what looks better.

[Maximilian Fellner @mxfellner]

@badlogicgames The Wolf’s interview skills are a bisserl overhyped though, imo…

[Maximilian Fellner @mxfellner]

Well done, Peter, congrats! 🤝 Sadly, another great one lost to the US of A. It's disappointing that the opportunities back home are not big enough for our biggest talents.

[Peter Steinberger 🦞 @steipete]

I'm joining @OpenAI to bring agents to everyone. @OpenClaw is becoming a foundation: open, independent, and just getting started.🦞
https://t.co/XOc7X4jOxq

[Christoph Janz 🕊 @chrija]

Following the awesome @openclaw meetup in Vienna yesterday, who's up for an Open Claw meetup in Berlin in a few weeks (e.g. March 10 or 12)?

1) "like" if you like it
2) + retweet if you love it
3) + add a comment if you're game so we can keep you posted

[Maximilian Fellner @mxfellner]

@chrija @andreasklinger @openclaw Do it!

[Maximilian Fellner @mxfellner]

@mitsuhiko @LXj42 The efforts for a sovereign Europe, including Ukraine, are ongoing. There are opportunities to help, for those who care.

[Sharif Shameem @sharifshameem]

I just built a *functioning* React app by describing what I wanted to GPT-3.

I'm still in awe.

[Maximilian Fellner @mxfellner]

And openclaw doesn’t have a built-in model router? I mean, I really appreciate the engineering behind it and the fact that I can now chat with my agent on Discord. However from a cost perspective, doesn’t just using Claude Code or Codex make a lot more sense?

[Maximilian Fellner @mxfellner]

Is anyone talking about how incredibly expensive clawdbot / openclaw 🦞 is in terms of model API costs? Openrouter isn’t really working for me as it either only sticks to an expensive model, a lobotomized Gemini 2.5 Flash Lite, or inexplicably times out.

[Maximilian Fellner @mxfellner]

…and it did it!

[Maximilian Fellner @mxfellner]

Opus 4.5 cannot make my LaTeX document compile with a correct layout… 😂 Maybe I’m giving AI models too hard tasks but I’m not feeling the AGI right now at all…

[Maximilian Fellner @mxfellner]

@gfodor It’s the kind of inner monologue that’s constantly going on in your mind when you know your code and the problem you’re trying to solve in and out. You already *know* what code to write and are just annoyed that it’s not already written. Most “vibe coders” don’t go that far.

[Maximilian Fellner @mxfellner]

Deploying Confluent schema-registry to Kubernetes is the final boss of LLM coding agents, for me so far at least.

[Maximilian Fellner @mxfellner]

Regarding the #chatgpt app, now there appears to only be GPT-5 left, without any other options.

[Maximilian Fellner @mxfellner]

@mitsuhiko Do you prepare and give extra context like architecture and requirements docs, agent instruction files, etc.?

[Maximilian Fellner @mxfellner]

@mitsuhiko Tired: reclaiming parent time by letting Claude write all my code. Wired: reclaiming coding time by letting Claude do all my parenting?

[Maximilian Fellner @mxfellner]

Live streaming my life to a cloud-based shoggoth owned by a big tech company seems like the most dystopian thing ever.

[Ben Geskin @BenGeskin]

WSJ on Jony Ive and Sam Altman’s OpenAI device:

• The product will be capable of being fully aware of a user's surroundings and life, will be unobtrusive, able to rest in one's pocket or on one's desk, and will be a third core device a person would put on a desk after a MacBook Pro and an iPhone.
• The Journal earlier reported that the device won't be a phone, and that Ive and Altman's intent is to help wean users from screens.
• Altman said that the device isn't a pair of glasses, and that Ive had been skeptical about building something to wear on the body.

via: https://t.co/30nJY4gL2b

[Maximilian Fellner @mxfellner]

As a company in Germany, let’s say you want to import some goods from China. You’ll need to apply for a special number in addition to the 4+ other numbers you already have. The application process only takes a speedy 2-3 weeks! 😂

[Maximilian Fellner @mxfellner]

Who’s building the first UI framework for LLMS? Like, MCP but for outputting UI. This might be the endgame for frontend development.

[Andrej Karpathy @karpathy]

"Chatting" with LLM feels like using an 80s computer terminal. The GUI hasn't been invented, yet but imo some properties of it can start to be predicted.

1 it will be visual (like GUIs of the past) because vision (pictures, charts, animations, not so much reading) is the 10-lane highway into brain. It's the highest input information bandwidth and ~1/3 of brain compute is dedicated to it.

2 it will be generative an input-conditional, i.e. the GUI is generated on-demand, specifically for your prompt, and everything is present and reconfigured with the immediate purpose in mind.

3 a little bit more of an open question - the degree of procedural. On one end of the axis you can imagine one big diffusion model dreaming up the entire output canvas. On the other, a page filled with (procedural) React components or so (think: images, charts, animations, diagrams, ...). I'd guess a mix, with the latter as the primary skeleton.

But I'm placing my bets now that some fluid, magical, ephemeral, interactive 2D canvas (GUI) written from scratch and just for you is the limit as capability goes to \infty. And I think it has already slowly started (e.g. think: code blocks / highlighting, latex blocks, markdown e.g. bold, italic, lists, tables, even emoji, and maybe more ambitiously the Artifacts tab, with Mermaid charts or fuller apps), though it's all kind of very early and primitive.

Shoutout to Iron Man in particular (and to some extent Start Trek / Minority Report) as popular science AI/UI portrayals barking up this tree.