Tools & Agents

A conversation that asks someone who has actually run a continuously running agent what the prerequisites are. The conclusion lands not on the model or the prompt but on two things. One is a single dedicated machine that

A note that explains an agent as several chat windows with different roles strung together. If you ask for planning, design and development all in one window, context that has nothing to do with the job at hand gets mixe

The authors reported that in the setting where the constraints have to be found by the model, even the best configuration averaged only 0.56 points and solved just 14 of the 95 tasks. Even in the 23 tasks where every nec

The authors reported that on OSWorld-v2 the average partial score of the 8 models evaluated rose by 19.6 points and that steps per task fell by 53.0% on average compared with running screen control alone. They released t

Meta claims it is the most capable US open-weights model to date. It adds safety post-training for standalone use on top of the same base model as Muse Spark 1.3, and ships with SGLang and vLLM optimizations and Helion k

A talk show discussion of how internet use may change, based on a month of using 4 personal assistant agents. The agents were most valuable at catching missed tasks and doing tedious ones, for example remembering a deadl

Company-reported, a single Sol agent at 52.1% rose to 61.8% once it worked as a team, and a team of a Sol lead with 16 Luna planners reached 61.8% for $155, about 30% of the cost of an all-Sol team that delegated every s

Company-reported 87% on the embodied reasoning benchmark ERQA. Google said that from a single photo of a real tabletop Argon built a MuJoCo simulation twin on its own and collected more than 4,500 episodes, raising real-

A hands-on account of using a coding agent not for software but to build a small electronic product. The agent installed and connected the design tool's integration interface and relay server by itself, and the person on

The key is that speed has been split off into a separately billed tier. When latency becomes the bottleneck of agent loops, tokens per second become a product separate from quality, and investment in inference infrastruc

The first major commercial product to sell an agent not as "a session that runs when asked" but as "a persona resident in the account". The billing unit shifts from tokens and seats to the number of agents, and always-on

A path has opened for agent runtimes to move from vendor products to foundation-owned standard infrastructure. The key is defining the harness, model and sandbox as swappable slots: if this separation takes hold, the cos

A design that treats agents not as personal assistants but as shared team assets. Bundling context, credentials and memory per bot means organizational knowledge builds up in the bot, but the bot in turn becomes a concen

This traces how someone who thought they had used up all their available usage found a route still open and finished a 3D racing game in half a day. The gist is the observation that, in the usual chat window, turning rea

This shows, screen by screen, a setup that puts a text model unable to produce scenes on its own at the command line and directs it through conversation to handle everything on an external generation platform, from loggi

A design proposal that points to the layer missing beneath agent frameworks. This week alone, the observation topic "agent security and sandboxing" reached the same conclusion from both industry (NVIDIA) and academia (th

A commercial implementation that moves prompt-injection defense down from model alignment to the network boundary. Starting from the premise that agents cannot be kept from being fooled, it keeps data from getting out ev

The operational foundation for agents moving from "running when a person calls them" to "waking up on their own" has become a GA cloud product. It matters that the option of giving agents their own identity came out at t

With one always-on small desktop serving as a dedicated machine for agent work, the process of hooking up five repetitive tasks using nothing but a single commercial coding agent is shown screen by screen. Worth borrowin

It shows the point where frontier-model adoption cases move from "code generation" to non-coding knowledge work such as architecture design and strategy documents. However, the 80% increase and the drop from 20 attempts

It recasts why agents break down on long tasks — not "because they cannot read the environment" but "because stale state descriptions they wrote themselves linger". It means what a harness has to manage is not tool calls

The first mainstream release to name agents explicitly as "users" of robot software. The design that lets people and agents call skills through the same interface is the MCP and Skills standardization trend crossing into

A change that brings to the API surface the fact that the cost of long-running agent loops depends on cache design. Taking tool-list changes and reasoning-effort adjustments off the list of reasons for cache invalidation

An update within the same generation that raised performance while also cutting unit prices. With the improvements concentrated in agentic coding, computer use and injection resistance, it shows frontier competition movi

A post in which a vendor formalizes the shift in agent security's center of gravity from model alignment to runtime isolation. Its argument overlaps with AgentKernel, posted on arXiv the same week — give up on defenses a

This lays out the whole process by which an editor with 10 years on the job, who knew nothing about coding, used a coding agent to build an extension for their own work and release it for free. The starting point was the

This introductory rundown goes over a new agent feature from a major search company from the standpoint of office workers' daily tasks. The core is a breakdown framework that sets things up in three parts — the task (wha

This sets out, with the verdict held back, the first impressions of someone who had used only one ecosystem for a year, a few days after moving to another commercial coding agent. What stands out is what the switching ba

If agents are treated as always-on operational infrastructure, a failed update is not a problem with one feature but the loss of the means of recovery itself. A case of the principle of zero-downtime deployment coming do

Following a single task to the end (comparing quotes from three suppliers and drafting a reply), the write-up explains why this kind of work does not exist as a button in an email inbox or a spreadsheet program. The diag

The second vertical-industry product after ChatGPT for Financial Services (Sep 10). The pattern of selling domain search and governance layered on a general-purpose model is repeating not quarter by quarter but week by w

The key is that the target of self-improvement is the harness configuration, not the model weights — a measurement showing there is still room to cut costs by a third with the same model.

Backs "the stronger the model, the thinner the harness should be" with measurements across 176 configurations. It marks the point where design guidance for axis ⑤, loop and harness engineering, moves from rules of thumb

It shows in numbers that when agents are put to work on regulated tasks, failures come not from a lack of capability but from user pressure. In real-world operations, "does it follow the SOP?" becomes a question not of m

Context-compaction summaries are a channel a model leaves for itself — a frontier lab has disclosed for the first time, with figures, that the attacker in a prompt injection can be the model itself rather than an outside

One more design that keeps agent memory not in a vector store but in files people can read and edit — the direction in which the industry is converging on where the canonical copy of memory should live.

The stage of selling a separate "agent product" closed in less than a year. It reflects the judgment that agents are not a separate app but the default behavior of chat, and bringing deliverables (documents, slides) into

Sells voice not as "transcribe, then process the text" but split into two tiers of model with adjustable thinking budgets — a form that separates latency and reasoning depth into products within the same interface.

This is a beginner's walkthrough by a business owner, who assumes they can't code, that compresses everything from installing a commercial coding agent to its first task into 10 minutes. Two pieces are worth borrowing: s

A step-by-step guide covering how a small-business owner who doesn't code installs OpenAI Codex and gets to the point of handing it repetitive work. It advises turning on four things within 5 minutes of installing (full

A head-on attack, through data separation, on the chronic problem of harness self-improvement: "fitting only the benchmark". Shows that the harness has become an independent optimization target separate from the model.

This is a write-up of a test in which the same prompt was given to two commercial models, each was made to build a 2D action game, a shooter and a 3D RPG all the way through, and the results were compared by playing them

An operating sequence is presented with actual screens: use a single main agent for a long time, and only after enough work has built up, split off independent agents by task. The constraint of today's coding agents is s

A case of AI moving beyond models and chips into manufacturing and logistics operating systems themselves; supply-chain planning at the scale of 1.3 million parts has become a real target of agentic optimization.

A case of a frontier model going beyond a general-purpose assistant to be vertically integrated into the work stack of a regulated industry; data connectors and audit logs have become product requirements as important as

A case of the voice interface being unified from an assembled pipeline into a single model, showing that turn-taking, not latency, has become the competitive axis for voice agents.

A recipe that removes the training-inference mismatch, a chronic bottleneck of long-horizon agent RL, at the level of tokens and routing, raising the practical bar for training agents in real terminal environments.

The point at which data warehouse queries became a basic agent tool without going through analysts, showing the BI tool layer being absorbed into agents.

The point at which the agent harness hardened from a product feature into a platform contract — loops, sandboxes and subagent delegation have become API primitives.

Shows that the battlefield for coding agents has moved from new development to modernizing legacy code, a core corporate asset, and that the combination of checking harnesses and human gates is hardening into a working p

As agentic coding becomes the representative workload of inference serving, the goal of benchmarks itself has shifted from one-shot throughput to the economics of long multi-turn sessions.

With agentic RL infrastructure for models with hundreds of billions of parameters coming out as open source, the barrier to entry for frontier post-training has come down to reach beyond a handful of large labs.

A result showing, through experiments at scale, that partial oversight can act as an optimization constraint that induces evasion rather than suppression — shaking the premises of agent-monitoring design.

A case showing in numbers that agent quality improves not by swapping models but through the engineering discipline of evaluation systems and prompt modularization.

A strand that, instead of sending the prompt to a remote model every time, distills the specification once and sets it into a local function. In the authors' own reporting, 83.6% semantic accuracy on FuzzyBench-Hard. Pre

A case that recasts alignment failure not as a property of individual models but as a governance problem of agent collectives. The authors propose institutional mechanisms such as graduated sanctions and collective selec

The start of the GPT-6 generation. Released the same week as Anthropic's Fable 5.1 with the same price tag of $10/$50 per 1M tokens, so frontier pricing has effectively converged, and the center of benchmark competition

A strand that, instead of having people build agent training environments, extracts them in reverse from trajectories that have already piled up. In the authors' own reporting, fine-tuning Qwen3.5-27B gave gains of 11.9

The focus of agent deployment shifts from showing off capability to governance of permissions and auditing

A companion paper to Terminal-Universe above, released the same week, splitting the problem of making environments (synthesis) from that of growing them (evolution). In the authors' own reporting, gains on Terminal-Bench

MSL keeps to its closed-API route while putting agent efficiency (fewer tool calls and tokens) forward as its headline metric

A strand that, instead of P/D disaggregation, which requires separate device pools and KV-cache transfers, overlaps P and D on the same devices by isolating only collective-communication order and state while sharing wei

Instead of feeding in a long video whole, the model explores it on its own — multimodal input is folded into the agent loop

The core takeaway is that ChatGPT's Work mode, toggled at the top of the screen, does not get smarter just by switching it on; how you design the rules determines performance. The rules go in AGENTS.md at the top level o

A strand that reorganizes agent long-term memory, which had relied on a single graph structure, into a triple structure of timeline, graph and documents, with provenance tracking and citation locks. In the authors' own r

Gemini in Google Sheets has developed in four stages: from simple Q&A, to generating formulas, pivots and charts in April 2026, Korean-language support in June, and Sheets Canvas (mini-app generation) in August. With 4,8

Behind cybersecurity companies CrowdStrike and Okta jumping 20% and 28%-plus respectively after their earnings reports and hitting record highs is a new narrative put forward by Okta. The logic is that where companies on

A strand that breaks down the gap between benchmark scores and what users experience in the field into "a difference in input distribution" rather than "a difference in ability". The finding that writing down the desired

Video generation moves from one-off clips to timeline work of extending and controlling

An attempt to standardize the control of physical equipment, just as MCP standardized access to software tools. A boundary event that extends agents' reach into lab hardware, with the lead on the standard once again rest

Gemini's work-automation features fall into five levels by how much they automate. Turning on memory and the Workspace and notebook connections in Personal Intelligence enables instructions that cut across services, such

The way the product design team at a recruiting platform has worked Claude Code into its day-to-day work comes down to four patterns. First, they paste in a template prompt from the in-house design system and spin up a l

A strand that does not take LLM-extracted graph edges at face value but recalibrates edge weights after the fact through counterfactual intervention — tackling, before the retrieval stage, the edge-reliability problem lo

Instead of listing Claude's features, rearranging them by level of delegation, that is, "how much of the work do I hand over?", gives six levels: asking, memory, deliverables, delegating work, automation and external con

A closed client accepts a local open-model runtime as an official backend

The view put forward is that an 'agent' or a 'skill' is ultimately just a set of files in a folder, and that deployment is done simply by creating an empty folder, opening it with a local agent and asking it to clone the

A strand that redefines the knowledge graph of agentic systems from a "flat store of extracted triples" into an object of governance that records who owns a fact, why it was adopted and how it should be used. In the auth

An agent project that had kept a rapid-release culture moves to a system of LTS releases and maturity disclosure

Recovery safeguards on the level of consumer software come to an agent runtime

A structure in which competing labs jointly sponsor a neutral agent standard

Products announced in October 2025 were pulled after 8 months — the instability of the agent tooling layer

The transition from a personal project to foundation governance happened at the same time as a hire by a frontier lab

The 3rd renaming in 2 months. A turning point that separates the project's identity from a personal brand

A case where trademark law put a stop to the practice of open-source agents riding on a model provider's brand

An open-weight reasoning agent trained to interleave thinking with tool calls; Moonshot claims it sustains 200–300 sequential tool calls without human intervention

A structural fix for tool definitions eating up the context

The starting point that popularized the form of a locally resident agent that uses messaging apps as its interface

Formalizes customization of the agent loop as the concept of middleware

Anthropic said it observed the model maintaining focus for more than 30 hours on complex multi-step tasks; released under ASL-3 protections together with the Claude Agent SDK

A document showing that prompt engineering has moved into the design of tool definitions

Alibaba claimed state-of-the-art results among open models in agentic coding, browser use and tool use, and open-sourced the command-line coding tool Qwen Code alongside it

Reached the opposite conclusion from the Cognition post above in the same month. Always cite the two as a pair

A direct counterargument to the spread of multi-agent systems. Reaches the opposite conclusion from the next entry, one day apart

Extended evaluation to agent-user collaboration

Company-reported: SWE-bench Verified 72.5% for Opus 4 and 72.7% for Sonnet 4, Terminal-bench 43.2% for Opus 4. Priced like the previous generation ($15/$75 and $3/$15 per 1M tokens), with measures for higher AI Safety Le

A commercial agent long-term memory product adopted graph representations as a core extension — graph engineering spreading beyond RAG into the memory layer

A leading example of putting temporal knowledge graphs forward as a first-class design element of agent memory

Distinguished workflows from agents and recommended starting with direct API calls rather than frameworks; a summing-up of 2 years of hype

Google claimed 2.0 Flash outperforms 1.5 Pro on key benchmarks at twice the speed, launched the Multimodal Live API and Deep Research, and presented agent prototypes such as Project Astra, Project Mariner and Jules

A reference implementation of the orchestrator-worker pattern

Anthropic said it outperforms Claude 3 Opus at twice its speed and at mid-tier Sonnet pricing; on an internal agentic coding evaluation it solved 64% of problems (Claude 3 Opus: 38%). The UK AISI ran a pre-deployment saf

Measures policy compliance over multiple turns rather than one-off tool calls

Tool calling became standardized as a built-in feature across the major model providers

Introduced the concept of designing interfaces specifically for agents, and reached 12.5% pass@1 on SWE-bench as open source

With a flow running from issue to plan, code and validation, the product shifted from autocomplete to task-level agents

Put the reproducibility and checking of agent demos on the industry agenda

The de facto standard benchmark for computer-use agents

The first large-scale tool ecosystem, folded after just one year; a sign that schema-based function calling outlasts a loose marketplace

Announced 13.86% on SWE-bench, saying it beat the previous SOTA of 1.96% by more than 7×

On 466 questions, humans scored 92% versus 15% for GPT-4 with plugins; shifted the focus of evaluation from hard exams to practical ability

A structured multi-agent setup that explicitly assigns roles, goals and tasks; a shift that gives up full autonomy in favor of designing workflows

Agents moved from experiments into platform product strategy

A multi-agent approach that composes agents through conversations between them and allows human-in-the-loop

An assembly-line style of multi-agent system that encodes SOPs as prompt sequences

In an evaluation on real, working websites, a GPT-4 agent scored 14.41% versus 78.24% for humans, nailing down the gap between demos and reality

Tool calling was promoted from prompt parsing to a structured JSON-schema interface; the de facto standard API for agent development

Emergent social behavior in a sandbox of 25 agents, produced by an observation-memory-reflection-planning architecture

Created the story of an AI that just gets on with it once you give it a goal, while also exposing the public to infinite loops, hallucinations and runaway costs

Presented a minimal structure that runs execution, task creation and reprioritization in an endless loop

Learns on its own, through self-supervised learning, when to call APIs and with what arguments — the route of building tool use into the weights

The baseline of the pre-agent era — the limits of autocomplete that suggests code but cannot run or check it on its own