01Agents
Tools
Agents and dev tools worth using, the releases that change the work, and status-page incidents worth watching.
Other
Flash-dLLM: IO-Aware KV Caching and Parallel Decoding for Fast, Memory-Efficient Diffusion LLMs
IO-aware KV caching and parallel decoding for diffusion LLMs — cuts memory footprint and speeds up token generation, relevant for agent runtimes that need fast, efficient inference.
PACT: From Credit Assignment to Critic Alignment
Formalizes token-level credit assignment via three conditions — directly relevant to RL-trained agents where sparse rewards make critic alignment the bottleneck.
How we made claude.ai 3x faster in two weeks
Anthropic used Claude to measure and optimize claude.ai, cutting latency 3x in two weeks — a case study in self-improving agent workflows, with HN discussion on reward hacking in kernel optimization.
Advancing Private AI Compute with secure, server-side memory
DeepMind details secure, server-side memory for private AI compute — relevant if you're running agents on sensitive data and need enclave-backed memory.
Ringg’s AI agents resolve up to 65% of customer calls with OpenAI
Ringg reports its OpenAI-powered agents resolve up to 65% of customer calls autonomously — a concrete production benchmark for customer-service agent deployments.
Can Agents Design Better Chips with a Higher Level Abstraction?
Asks whether agents can design better chips when given higher-level abstractions — early evidence on agent capability in hardware design workflows.
When LLM Agents Fail to Read the Room: ReAdapt for Relational Social Reasoning
Proposes ReAdapt, a method for improving LLM agents' relational social reasoning — targeting failures where agents miss conversational context and social cues.
The Tasteful Agent: Measuring and Improving Taste in Long-Horizon Tasks
Introduces a benchmark for 'taste' in long-horizon agent tasks — measuring whether agents produce outputs that match human quality preferences, not just complete the task.
SF October 14th: A Birds of a Feather Session on Agentic Engineering
Simon Willison and Jesse Vincent host a Birds of a Feather on agentic engineering in SF Oct 14 — an evening for practitioners to compare notes on building and operating agents.
Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war
Simon Willison on the day's model drops — Claude Opus 5.5, GPT-6 Sol/Luna, Grok 4.7, MiMo v2.6 — and the price war they're triggering. Worth reading for how the cost curve shifts for agent workloads.
Recognition, Simulation, and Refusal: A Contamination-Aware Study of Classic Psychological Effects in LLM Agents
Contamination-aware study of classic psychological effects in LLM agents — checks whether agents actually exhibit effects or just parrot contaminated benchmarks. Useful for anyone evaluating agent behavior.
Memory That Looks Forward: A Zero-Inference Prospective Term for Personal Memory Retrieval
Proposes a zero-inference prospective term for personal memory retrieval — memory that anticipates future queries without extra inference cost. Relevant to agent memory design.
RRSI: Regularized Recursive Self-Improvement of Agent Harnesses
RRSI formalizes regularized recursive self-improvement for agent harnesses — adds a regularization term to keep self-modifying loops from drifting. Directly relevant to anyone running agents that rewrite their own prompts or tools.
onPanda: Efficient Annotation of On-Policy Alignment Data for LLMs and Agents via Token-Level Correction
Open-sourced onPanda — token-level correction for on-policy alignment data, aimed at cheaper RLHF/RLVR annotation for LLMs and agents. Useful if you're building agent-specific alignment pipelines.
MCP was always a bad idea?
Simon Willison pushes back on 'MCP was always a bad idea': MCP's value is for contexts without a full terminal, not for terminal-based agents. His comment clarifies the protocol's actual niche.
One to More, More to One: Category-Aware Iterative Expert Training for Software Engineering Agents
Paper proposes category-aware iterative expert training for SWE agents — trains specialized experts per task category, then merges them. A concrete recipe for improving coding agent performance.
ProgramDistill: From Interactive Web Apps to Verifiable Reference-Guided SWE Tasks
ProgramDistill converts interactive web apps into verifiable, reference-guided SWE tasks — a new pipeline for generating grounded coding-agent benchmarks.
Self-Compact Pi Agent: ZERO HYPE Agentic Coding Devlog
IndyDevDan builds a self-compacting agent on a Pi — zero-hype devlog on keeping agent context lean on constrained hardware.
How V7 gives AI agents institutional memory
V7 gives agents institutional memory by persisting organizational context — a case study in grounding agents in enterprise data.
MintAct: A Unified Visual Agent for Digital Environments
MintAct unifies visual agents across digital environments — one model for GUI, web, and desktop control instead of per-environment harnesses.
BI-Agent and BI-Bench: Towards Automating End-to-End Business Intelligence
BI-Agent + BI-Bench propose an end-to-end BI automation agent and benchmark — enterprise decision workflows as an eval target, not just a demo.
RecreationWorld: Scalable and Verifiable Environments for Hybrid Computer-Use Agents
Scalable, verifiable environments for hybrid computer-use agents — moves GUI+API agent evals beyond toy sandboxes with reproducible tasks.
LimiX-2: A Contextual Mechanism Network Towards General Structured-Data Intelligence
Contextual mechanism network for structured-data reasoning — candidate architecture for agents that need to query and reason over tables/DBs.
HarnessTax: How Much Does the Harness Matter for Coding Agents?
Quantifies how much the harness (vs. the model) drives coding-agent performance — direct evidence for the Claude Code vs. Codex vs. open-harness debate.
Verifiable Social Reasoning for LLM Assistants
A paper proposing a method to evaluate social reasoning in LLM assistants.
Rethinking Critic Learning in PPO: Understanding and Mitigating Value Flattening
A paper addressing value flattening issues in PPO critic learning.
What Does Privileged Information Add to On-Policy Self-Distillation?
A paper auditing if providing teachers with more solution data improves self-distillation.
Agora: Git as Shared Memory for Collective AutoResearch
Uses Git as the shared memory substrate for multi-agent research collaboration — agents read/write to a repo as their coordination mechanism. Relevant for anyone building multi-agent orchestration that needs persistent, versioned shared state.
Can Jev Be a Better Agent Evaluator?
LangChain's case for Jev as an agent evaluator, positioned against their LangSmith eval stack. Worth reading for how they frame scoring agent performance beyond single-turn accuracy.
PACT: Can Enterprise AI Assistants Be Trusted Under Pressure?
Benchmark for enterprise AI assistants under pressure, with an interactive leaderboard — stress-tests whether assistants stay trustworthy when stakes rise.
Building an Agent Harness for Life Sciences: Introducing Deep Life Sci
LangChain ships Deep Life Sci, a life-sciences agent harness built on deepagents — sandboxes, evals, and observability tuned for domain-specific agents.
CERA-MoA: Co-Evolving Routing Mechanisms with Continually Learning LLM Agents
MoA routing that co-evolves with continually learning agents instead of freezing router weights — targets the drift where a static router misroutes as agent capabilities shift.
AI for Games in the Foundation Model Era
Survey of foundation-model game agents — where LLM-driven NPCs, planners, and world models stand, and what's still missing.
Agentgit: a Git host for AI agents with no accounts
Git host where the repo name is the identity — no accounts, tokens, or keys; agents push to a name and hand off the URL. Open source and self-hostable.
HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harness
Training-free embodied navigation via a harness that unifies VLN methods — no fine-tuning, the harness coordinates perception and action. Relevant if you're building embodied agents.
Characterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and Responses
Empirical study of how conversational LLM agents decide when to search, which strategies they pick, and what they return — useful signal for tuning your own agent's search policy.
Closed-World Resolution Against Tool Hallucination in LLM Agents
Proposes closed-world resolution to stop LLM agents from hallucinating tool calls — a top production failure mode when agents invoke tools that don't exist or misuse schemas.
MAGS: Multi-agent Auto-formalization Guarantees Safety for Agentic Outputs
Multi-agent auto-formalization claims formal safety guarantees for agentic outputs — worth a look if you need verifiable agent behavior beyond prompt-level guardrails.
Do AI Agents Understand Computer Architecture?
Benchmark probing whether AI agents actually understand computer architecture — useful signal for coding-agent evals and where they break on low-level reasoning.
SoL-Pi: Recursively Scaling Auto-Research Loops for Efficient Agent Harness
Proposes recursively scaling auto-research loops for agent harnesses — targeting unattended coding agents that move beyond supervised code completion.