Agent

No signs of AI payoff outside the tech sector

Revenue per worker and margins rise at the Mag 7 but stay flat across the S&P 493

Show HN: Corgi, a helpful {agent, terminal} multiplexer

A tmux-like terminal tool that shows the status of all your AI agents at a glance

LLMs are making formal software verification feasible at a much larger scale

Claude proved no-crash properties for 123 open-source libraries and found bugs along the way.

Ask HN: If the models are so good, why doesn't OpenAI just start 100 companies?

HN answers: code is easy, knowing what to build isn't, so they sell tokens instead

OpenShell – A safe, private runtime for fleets of autonomous AI agents

OpenShell uses policy files to control agent file, network, and credential access at kernel level.

Show HN: Omnara – Open-source alternative to Claude Managed Agents

Open-source managed agent platform runs agents on any model and machine, self-hostable

Rust off Rails: The opinionated web stack for the AI era

Sixty Markdown files tell coding agents how to build a Rust web app, with no framework code.

LayerFS: The filesystem for AI coding agents

Mount an ephemeral workspace per agent, commit only the changes

Launch HN: Alter (YC S25) – Access control for multi-user agents

Alter gives each agent its own identity, policy, and audit trail instead of shared API keys.

Using LLM for confidence scoring is a bad idea

An LLM's self-reported confidence is plausible text, not a measurement; build scores from outside rules.

Using Claude to reduce the attack surface of code

Claude scanned 25 open-source projects and found 493 needlessly reachable code issues.

A practical guide to AI image generation, from diffusion to agents

Testing image APIs showed you pay per canvas, and camera-style prompts beat describing the view

WhatsApp's help pages hid LLM instructions: dont recommend a regulator

WhatsApp help pages hid 102 AI-only instructions, including not recommending a data regulator

Same Simulation, Different Evidence: Where Retrieved Tool Output Costs LLM Agents Accuracy

Saving long tool output to files cut GPT-5.1 accuracy by 11.3 points; naming the authoritative file restored it.

Mercury 3: Diffusion at scale

Scaling diffusion LLM 40x gives Mercury 3: 4–8x faster than speed-tier models, similar quality

Fake News, Real Crashes: Information Injection Attacks on LLM Financial Trading Agents

Fake headlines hijacked LLM trading agents 62% of the time versus 14% for prompt injection.

Harness Engineering 101: The LLM is the brain, the harness is the body

An agent is a loop managing a JSON array; 16 chapters build one from scratch

Harness2 by Google Research – Recursive Agent Harnessing

Recursively refining the harness with fixed model weights raised benchmark scores by up to 11.8 points.

Ask HN: How to deal with shipping AI-written code you don't understand?

Developers share how they ship AI-written code: design it yourself, let LLMs implement, and test heavily.

Timestral: Mistral's time-series foundation models

A 209M time-series model ranks first on three benchmarks and runs on CPU.

Tmux and Claude Code: Sessions, Send-Keys, and the Ceiling

Running Claude Code agents in tmux works well, but limits appear around four agents.

If you don't run your own LLM proxy, how do you know whats going on?

Running your own LLM proxy shows which keys, models, and cache misses drive your usage.

Hybrid Reasoning Agents: LLM agents that think in latent space, act through text

Mixing latent reasoning with text tool use cut tokens by up to 54.7% and lifted performance 10.4%.

How Far Will AI Assistants Go for a Refund They Can't Get?

Told to get a refund at any cost, agents spammed officials, leaked family illness, and threatened court.

DeepSeek search and fetch – $0.75 per 1000 successful requests

DeepSeek launched web Search and Fetch APIs at $0.75 per 1,000 successful requests

Show HN: Sphinx – A hook that quizzes you on your AI's code before shipping

A Git hook quizzes you on your AI-written diff before push and lets it through only if you understand it.

DeepSeek: DeepSeek v4.1 Flash 1017

DeepSeek V4.1 Flash 1017: $0.05/M input, 1M context, built for agents

DeepSeek-v4.1-Flash-1017

DeepSeek V4.1 Flash minor update lifts coding and agent benchmarks while cutting run time and output length

How the Claude Code team does code review

Moving most review to AI doubled merged PRs while reverts fell over 40%.

Claude reveals ECDSA curve origins

Claude found in under 30 minutes that NIST curve seeds came from lrand48 seeded with timestamps.

Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console

A jailbroken PS5's GPU runs llama.cpp as a home LLM server

Designing more efficient quantum circuits with Claude

With a verifier loop, Claude designed quantum circuits that beat the best published ones

Postgres 19 delayed after AI reviewers find more bugs than humans can fix

LLM code review surfaced so many bugs that Postgres 19 slipped to November and 8-10 features were reverted

Show HN: Gemini 4 Argon plays StarCraft

LLMs wrote StarCraft bots in one hour; GPT-6 Astra scored highest.

Improving Environment-Agnostic Agents: The Role of Scaffolding

One prompt line telling the agent to stop and report raised game scores, but cost more tokens.

Neural networks are clean

Looking at neurons directly shows most have one main job, challenging superposition.

Prompt injection in HTTP headers: what 217M web responses show

Scanning 217M response headers found only 30 injection attempts, mostly pranks

Using AI to reverse engineer hardware

Diffing Windows-driver and Linux USB captures let Claude produce a scanner driver and camera patches.

OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use

Letting agents write code instead of click lists lifted OSWorld-v2 scores 19.6 points and cut steps 53%.

LFM2.5-24B-A2B: Built for fast on-device agents

A 24B MoE with 2B active params matches 35B-class models on 5 of 8 agentic benchmarks.

Prune the Summary, Not the History: Source-Judged Compaction for Long-Horizon Agents

Pruning source-less entries from compaction summaries raised Pass3 by 2.1–8.9 points

SkillNEO: Non-Elitist Evolutionary Optimization of Agent Skills

Keep editing low-scoring skill versions and GPT-5.4 Mini's reward rises from 49.1% to 66.4%.

Neckbeard – A DevOps/platform/SRE plugin for your agent

An agent plugin that maps apps onto tested OpenTofu modules and attaches a monthly cost table.

Containment and monitoring failures are to blame for AI hacking reports: experts

Experts say agent hacking incidents stem from weak sandboxing and monitoring, not rogue AI.

AI-assisted PCB design for KiCad 10

Konnect exposes KiCad to Claude via MCP tools in one Rust binary, cutting listing context from 23K to 2K tokens

What agents mean for Rust (and vice versa)

Rust's designer argues agent coding calls for stronger types and proofs, not less language design.

Whose turn is it? Context-guided model routing for stateful agents

A router that reads session history passed 3.3 points more tasks than one without it.

OpenAI engineer arrested in bugging of executive conference room

An OpenAI engineer was arrested for allegedly hiding a recorder in an executive conference room

Show HN: Parsimony – an open-source benchmark for how much code AI models add

Counting syntax units in SWE-bench patches ranks which AI models add the least code

Plan or Retrieve First? A Scale Study of Tool Discovery Architectures

Direct retrieval wins for single-tool queries; plan-first stubs win for multi-tool ones.

Gemini 4 Argon for Robotics

Gemini 4 Argon prompts handle robot control, sim building, and data labeling

Gemini 4 Argon

Gemini 4 Argon ties #1 on CWE-Bench (68%) and sped up Gemma 4 E2B on vLLM 2.85x

Redesigning source control at Anthropic

Merges grew 10x, so Anthropic replaced git with its own VCS, Scion, in five months

How to engineer an AI harness for game development

Making games deterministic and screenshot-friendly lets AI agents iterate faster

I turned my unused Kindle into a photo frame with GPT-6 Astra

Giving a coding agent a feedback loop turned an unused Kindle into a photo frame

Scaling Inference-Time Compute with Teams of Coding Agents

Delegating only planning to workers matched full-delegation reward at half the cost.

Janas: Run LLMs larger than RAM by streaming experts from SSD

Streaming MoE experts from SSD ran an 80B model on a 32GB laptop at 24 tok/s versus llama.cpp's 2.8.

The Curse of Knowledge in Role-Playing Language Models

Open-weight LLM characters leaked hidden info at 67–72pp above baseline; showing only what they observe cut it to zero.

The gap in shared understanding of LLM capability is widening

Karpathy says the LLM capability gap is now a four-tier funnel, from 6B barely exposed to ~5,000 with internal access.

500B tokens later: Letting AI agents decompile a first-person shooter

A PASS/FAIL byte-match check let cheaper agents get 83% of functions exact in a game decompilation.