Prime Agent: A Self-Improving RLM Harness
arXiv:2608.23552v1 Announce Type: new Abstract: Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context
Knowledge catalogue
arXiv:2608.23552v1 Announce Type: new Abstract: Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context
arXiv:2608.21929v1 Announce Type: cross Abstract: Agent skills extend coding agents with task-specific instructions, scripts, and resources, but they also create a trusted instruction channel that can
arXiv:2608.20622v1 Announce Type: new Abstract: Frontier models have collapsed the cost of writing custom code: a niche problem a specialist sees in their own domain now costs an afternoon. The cost o
The key skill required to make productive use of coding agents is being able to confidently instruct them on how to make changes and then confidently verify that those changes have been applied in the
Stop Making TUIs Thomas Ptacek advocates for building real native user interfaces for even the smallest of personal tools, because coding agents have reduced the cost of getting a usable-enough GUI up
Hi everyone! I’m an intern at Icosa, a startup focused on making local AI accessible. We just released the first version of Zeno, our local AI product, and I thought it might be of interest to this co
Research: A shot-scraper-style JSON API on Bun 1.4's new Bun.WebView Today saw the long awaited release of Bun 1.4, the first stable version since the infamous Rust rewrite a few months ago. Interesti
arXiv:2608.17528v1 Announce Type: new Abstract: Modern agents operate inside agent harnesses that manage tools, context, and control flow, making the harness a critical part of the agent system. Our o
Last week I recorded an episode of the Talking Postgres podcast with Claire Giordano on the subject of 'How AI is changing software development'. We had a really great conversation. Here are a couple
I am bullish on simple agent harnesses. Love this idea from Vercel. fx is a new coding agent harness for model benchmarking, sandboxing, evals, and gyms. Introducing fx, a tiny, open, native coding ag
Friday's big release was Qwen 3.8 27B, an Apache 2 licensed 27B parameter vision-capable LLM from Alibaba's Qwen research lab. I've been looking forward to this one: 27B is an excellent size for runni
arXiv:2608.13522v1 Announce Type: cross Abstract: AI agents are increasingly used for programming, but do not provide any guarantee on the correctness of generated code. Verified code generation, in w
Connect your coding agents to AI Gateway with a single command. • Auto-configure 8 popular coding harnesses • 300+ models from 30+ providers, no markup • Open-weight models with ZDR & US inference ▲ ~
arXiv:2608.11727v1 Announce Type: new Abstract: When a coding agent obeys a rule, it may simply have been going to do that anyway. Existing instruction-following benchmarks cannot tell the difference:
If you write rules in an AGENTS.md, this one is worth your time. When a coding agent follows your rule, it may have been going to do that anyway. Harness-IF separates the two by scoring 256 rules one
arXiv:2606.08348v2 Announce Type: replace Abstract: LLM agents increasingly rely on prompts, tools, memory, SOPs, skills, and harness feedback, yet current self-evolution pipelines often update these
SPACEXAI: Grok 4.6 leads on the two strongest knowledge-work / real-world productivity benchmarks (GDPVal-AA and AA-Briefcase) and on the legal benchmark, while remaining highly competitive on coding-
With the smartest person I know, @HarshSensei, we're building evsys-sdk for the community, this open-source repository allows anyone to build their own continual learning system with first-class suppo
arXiv:2608.09072v1 Announce Type: cross Abstract: Large language model-powered coding agents are increasingly used to modify existing code repositories, for example, by adding features or fixing bugs.
💡The world needs an open-source platform, and that’s exactly what we’re building to give our customers more choice and the flexibility to choose the right model for the right task. As part of this, we
Auto mode is now the default in Claude Code for Pro, Max, and Team plans Anthropic are really confident in Claude Code's auto mode, to the point that they are making it the default setting for new ses
Moonlight & Mayhem (Raccoon Heist by Codex + GPT-5.6 Sol Ultra) On Wednesday I wrote about One-shotting a Raccoon Heist game using Claude Fable 5, where I had Claude Fable 5 build a full working game
Introducing Muse Code and Muse Spark 1.2 Yet more evidence that the most important characteristic of any model these days is long-sequence agentic tool calling. Meta shipped their own coding agent as
Back in 2024 I tweeted screenshots of a game concept generated by GPT-3 and some concept 'art' created using DALL-E. Today, on the fourth anniversary of that tweet, I decided to see if Claude Fable 5
arXiv:2608.01918v1 Announce Type: cross Abstract: Harness design plays a critical role in agent performance by shaping how large language models (LLMs) perceive, reason over, and act within executable
Gas Town was intended to be reusable, but I only ever wound up using it to build itself. Gas Town fell apart at the seams with Opus 4.7. Up through 4.6 it was working brilliantly. With 4.7 we saw the
Set up a nightly cron job that executes the prompt: fetch upstream changes to the <software> and rebase all local changes on top of upstream. Check that the software works as intended and replace the
arXiv:2607.26587v1 Announce Type: cross Abstract: Automated research systems use experimental scores both to deliver artifacts and to decide which ideas to retain, transfer, and pursue. Yet one run sc
arXiv:2607.28287v1 Announce Type: cross Abstract: ARC-AGI-3 turns abstraction into an interactive problem of skill acquisition. A player must infer an unfamiliar game's rules, hidden state, and goal w
Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident Hugging Face just released this extremely detailed technical description of OpenAI's recent accidental cyberat
arXiv:2607.21672v1 Announce Type: cross Abstract: Long source-code contexts consume many text tokens, motivating the proposal to render code as images for vision-language models. Recent work asks whet
If you run local agentic coding harnesses (Aider, Claude Code, etc.), prompt evaluation usually eats up most of your execution time. Every turn re-evaluates thousands of identical prefix tokens_system
Diffusion LLMs can now handle real agentic work. LLaDA 2.2 is the first large-scale diffusion LLM built to operate as a real agent, planning, calling tools, and self-correcting across long multi-turn
Built an open-source AI coding agent that was 7%–75% cheaper than a cold 'claude -p' run on 6/6 well-localized tasks across repositories up to ~82k LOC. The biggest difference: Cold agent: 6.83, 207 t
arXiv:2607.14573v3 Announce Type: replace Abstract: Payment integration is a demanding repository-level software task: agents must select a suitable product, implement coordinated client-server flows,
// Programmatic Memory Enables Long-Horizon Reasoning // Keep the entire interaction log and search it. It works great and beats bespoke memory harnesses on long-horizon tasks. New research introduces
Progressive disclosure in agents doesn't scale. And its benefits seems agent harness dependent. (bookmark this one) Finally there is a proper study on using agent skills and the effect of progressive
Highly recommended. I've often claimed there's huge alpha in building agent harnesses. Turns out harnesses are compositional generalizers. The RLM harness is an instance of this. This could lead to in
I keep hearing anecdotes from people who used coding agents to reverse-engineer and automate devices in their homes. I think this is an interesting illustration of the impact of the reduced cost of wr
xai-org/grok-build, now open source xAI's grok CLI tool faced severe community backlash yesterday when it became apparent that running the command in a directory could upload that entire directory to
The shared language of a software project is not English or Python but it is the common understanding of what its concepts mean, where the boundaries are, which invariants matter, who owns what, and w
datasette code-frequency chart on GitHub Out of curiosity I decided to see if I could find a useful illustration of the impact of coding agents and Opus 4.5 class models on my own output. The best I'v
Directly Responsible Individuals (DRI) I went looking for a definition of 'Directly Responsible Individuals' and the best I found was in the GitLab handbook. Apparently the term originated at Apple, w
arXiv:2606.22504v1 Announce Type: cross Abstract: Coding agents often receive broad tool access for an entire task, even when a resource is needed only for one subgoal. We call this gap lingering auth
I wrote about the sqlite-utils 4.0rc1 release a couple of weeks ago. Since we only have Claude Fable on our Max subscriptions for a few more days, I decided to see if it could help me get to a 4.0 sta
Better Models: Worse Tools Armin reports on a weird problem he ran into while hacking on Pi: The short version is that newer Claude models sometimes call Pi’s edit tool with extra, invented fields in
arXiv:2607.01366v1 Announce Type: new Abstract: Federated learning (FL) research often depends on many small but consequential algorithmic choices: optimizer variants, server aggregation rules, local
One of the most interesting tips I got from the Fireside Chat I hosted with Cat Wu and Thariq Shihipar from the Claude Code team at AIE on Wednesday was to let Fable (and to a certain extent Opus) use
arXiv:2607.01211v1 Announce Type: cross Abstract: Repository-level performance-optimization benchmarks such as GSO, SWE-Perf and SWE-fficiency evaluate coding agents by applying patches to real reposi
Release: llm-coding-agent 0.1a0 Another Fable 5 experiment. Now that my LLM library has evolved into more of an agent framework it's time to see what a simple coding agent would look like built on it.
I saw Geoffrey Litt speak at AIE yesterday, and one framing he used particularly resonated with me: Understand to participate Geoffrey was talking about the challenge of collaborating with coding agen
arXiv:2606.30111v1 Announce Type: cross Abstract: Embodied agents are typically built as hand-designed compositions of perception, memory, planning, and action modules. This modularity exposes a large
shot-scraper video is a new command introduced in today's shot-scraper 1.10 release which accepts a storyboard.yml file defining a routine to run against a web application and uses Playwright to recor
Qwen publishes new work on RL coding agents. (bookmark it) The idea is to continually build a verification system that co-evolves with AI agents. LLMs suffer from all sorts of reward hacking issues. T
arXiv:2606.30573v1 Announce Type: new Abstract: We introduce SWE-Interact, a new testbed for evaluating coding agents on multi-turn, interactive, user-driven software engineering tasks. Existing front
Human Agent in the loop I dislike the phrase “human in the loop” because it cedes authority to the machines. Let’s flip the narrative. It’s our loop, we work the same way we always have, now we recrui
This morning on Hacker News I saw Moebius: 0.2B Lightweight Image Inpainting Framework with 10B-Level Performance, describing a small but effective inpainting model - a model where you can mark region
arXiv:2606.12344v1 Announce Type: cross Abstract: General-purpose agents such as OpenClaw are increasingly used as autonomous tool users, but their coding ability is difficult to measure under SWE-ben
arXiv:2606.09863v1 Announce Type: new Abstract: LLM agents can fail silently by asserting task completion when the environment state shows otherwise. We study this failure mode, false success, across
arXiv:2606.08151v1 Announce Type: new Abstract: Tool-using LLM agents often fail not because relevant text is absent, but because decisive evidence is not selected, compressed, or surfaced at action t