Hacker News Reader

mhheron OpenClaw passes React in amount of stars on GitHub

OpenClaw has a faithful base of Chinese agents behind it. The Chinese web is flooded with posts about OpenClaw users becoming millionaires in order to entice more and more people to install it due to its incredibly sloppy and easily exploitable nature.

Peter and Sam are a masterclass in deception (or incompetence).

mhheron Pentagon head to military officers: You cannot study in these 22 universities

It's good to see that these universities seem to take their oath of education serious rather than bowing down to ideological agents. If you attempt to restrict or influence education you lost before it even started.

mhheron Firefox 148 Launches with AI Kill Switch Feature and More Enhancements

Am I the only one who absolutely adores the local based translation feature? I use it regularly and its so good.

mhheron You are not supposed to install OpenClaw on your personal computer

> You are not supposed to install OpenClaw

Sentence could have ended there

mhheron 0 A.D. Release 28: Boiorix

I once randomly stumbled upon this in GNOME Software (alphabetical sorting). Was very happy to find such a quality title there.

mhheron Altman on AI energy: it also takes 20 years of eating food to train a human

To me the whole OpenClaw situation is proof enough how desperate OpenAI must be for fresh (real, non-circular) cash.

In that light Altman saying things things like that is not really surprising. Contrary it only reinforces their desperation to me.

mhheron Show HN: Late – A subagent orchestrator TUI for local LLMs (Go/Linux)

Most AI coding assistants are fundamentally broken. Tools like Claude Code and OpenCode eagerly load 10k+ tokens of monolithic system prompts into a single context window before you even type. This guarantees context amnesia and destroys local inference speeds.

I built late to fix this. It’s a single-binary Go/BubbleTea TUI built around a subagent orchestrator pattern. The main prompt is strictly ~100 lines. The orchestrator routes tasks to transient subagents with isolated context windows, explicitly preventing the "Debugging Decay" (https://arxiv.org/abs/2506.18403) and context pollution seen in monolithic agents.

Architectural notes:

* Linux Native: Strictly respects XDG base directories. No Windows support. No padded Electron GUI.

* Execution Load: The rapid context switching breaks the mainline llama.cpp scheduler. You must compile the specific PR linked in the README (or just point it at any standard API via localhost:8080).

* Licensing: BSL 1.1 to keep the core orchestrator logic out of VC-backed wrappers.

I also open-sourced pure-go-sgd (AGPLv3) today for those interested in bare-metal Go infrastructure: https://github.com/mlhher/pure-go-sgd

mhheron Keep Android Open

I need to check if Aurora Store still exists/works.

mhheron Claws are now a new layer on top of LLM agents

While I understand the premise I think this is a highly flawed way to operate these tools. I wouldn't want to have someone with my personal data (whichever part) that might give it to anyone who just asks nicely because the context window has reached a tipoff point for the models intelligence. The major issue is a prompt attack may have taken place and you will likely never find out.

mhheron Claws are now a new layer on top of LLM agents

https://arxiv.org/abs/2512.13914

Among many more of them with similar results. This one gives a 39% drop in performance.

https://arxiv.org/abs/2506.18403

This one gives 60-80% after multiple turns.

mhheron Ggml.ai joins Hugging Face to ensure the long-term progress of Local AI

It's great to see the ggml team getting proper backing. Keeping inference in bare-metal C/C++ without the Python bloat is the only way local AI is going to scale efficiently. Well deserved for Georgi, Johannes, Piotr, and the rest of the team.

mhheron Claws are now a new layer on top of LLM agents

The current hype around agentic workflows completely glosses over the fundamental security flaw in their architecture: unconstrained execution boundaries. Tools that eagerly load context and grant monolithic LLMs unrestricted shell access are trivial to compromise via indirect prompt injection.

If an agent is curling untrusted data while holding access to sensitive data or already has sensitive data loaded into its context window, arbitrary code execution isn't a theoretical risk; it's an inevitability.

As recent research on context pollution has shown, stuffing the context window with monolithic system prompts and tool schemas actively degrades the model's baseline reasoning capabilities, making it exponentially more vulnerable to these exact exploits.