At Uber, our testing agent hit an outage in... Australia that I still think about. We had the recording. That's about it. No network trace for the run. No logs mapped to each step. No perf data. Just a video of an agent pressing a button and a team trying to reconstruct what happened underneath. We eventually traced it to a transient issue. But "eventually" is the problem. If the run had shipped with real instrumentation, the why would have taken one look, not an investigation. So when we built Revyl, that was non-negotiable. Every run comes with: > network waterfall, every request and response > device logs timestamped against each step > CPU, memory, and FPS traces for the full run The agent finds the problem. The report explains it.show more

"We have had a good run for our money... with this gas chamber story we have been putting about, but don't we run the risk that eventually we are going to be found out? And when we are found out, the collapse of that lie is going to bring the whole of our psychological warfare with it." - Victor Cavendish-Bentinckshow more

"We have had a good run for our money... with this gas chamber story we have been putting about, but don't we run the risk that eventually we are going to be found out? And when we are found out, the collapse of that lie is going to bring the whole of our psychological warfare with it." - Victor Cavendish-Bentinckshow more

Stanford researchers did it again. They just built the... agent-native version of Git. When an agent works on a longer task, the run builds up a lot of state. This includes files edited/created, a dev server, a database, installed packages, KV cache, etc. Say the agent is at step 10 and makes a mistake, maybe it misreads a traceback and rewrites a file that was actually fine. The tests start failing, and the run goes off track, although everything through step eight was correct. By default, the agent just tries to fix it, which creates more edits and tool calls. This burns more tokens and grows the context. The other options are a person stepping in to redirect it or restarting the whole run from step one. That's wasteful, because it pays for every model/tool call again and re-prefills the context. Moreover, since an agent's run is non-deterministic, it doesn't reproduce the same early steps anyway. The reason it's hard to just jump back exactly to a previous correct step and resume from there is that the trajectory is only a message log. It records what the agent said and which tools it called, but not the live state underneath. That state includes things like memory, open file handles, child processes, installed packages, /tmp, and KV cache. None of that is in the log. Git can version the files, but it doesn't snapshot the running process or the KV cache. Checking out step eight moves the files back, but the process is still sitting in step-ten memory with a cold cache. Shepherd is a runtime layer by Stanford that records the run as a trace of typed events rather than a flat log. Each agent-environment interaction becomes a commit, similar to Git, but it tracks the live run. Its commit includes the agent process and the filesystem together, copy-on-write, so a branch carries the actual state and not just the files. Going back to a previous step is then a single call that forks from that commit and continues from the exact state. The copy-on-write fork is roughly five times faster than docker commit, and because the prompt prefix through step eight is unchanged, the KV cache is reused over 95% on replay, so early steps aren't reprocessed again. Once the run can be forked, a meta-agent can sit on top and operate it. It watches the trace and reverts as soon as it looks wrong, before the bad write is committed. In practice, it's just Python calling fork, replay, and revert on the trace, rather than a separate control plane wired into the harness. Not everything is reversible though. Files and sandbox changes undo themselves, but a database write has no automatic undo, so it needs a matching undo step set up in advance. Something external, like a sent email or a real charge, can't be undone, so the supervisor's job there is to catch it before it fires. They tested this on a few public benchmarks. On CooperBench, where two agents work on the same codebase, adding a live supervisor took the pair-coding pass rate from 28.8% to 54.7%. It's still early and labeled alpha. The benefit mostly shows up when a run gets branched a lot over a heavy sandbox state, which is exactly where restarting wastes the most tokens and time. If Git was made to make file changes reversible, Shepherd is trying to do the same thing for a live agent run. Shepherd Repo: (don't forget to star it ⭐ ) That said, Shepherd reverts a bad step inside a run. The harness around it, the prompts, tools, and checks the supervisor relies on, still drifts across runs as models and dependencies change. Akshay wrote about making that harness repair itself, where a failing trace gets diagnosed, the fix is verified against the exact input that failed, and the failure is locked as a regression test so it can't recur. Read it below.show more

Visualizer of our MultiAgentRouter 🤖 The MultiAgentRouter is an... all-new multi-agent structure that leverages a hierarchical pattern to select the most specialized agent for your task. Here's how it works: Step 1. You give a task. Step 2. The Boss Agent Routes your task to the most specialized Agent Step 3. The selected agent returns your response! Get started with it now below ⬇️ Thanks to WE!SS for the visualizer!!show more

I just built a Meta Ads diagnostic in Claude... Code that tells you WHY your account broke, not just what changed 🤯 It spins up a team of agents that each investigate a different reason performance dropped, then argue against each other to kill the wrong answer before it ever reaches you. All inside Claude Code. Perfect for DTC brands and agencies who panic-kill creative the second CPA spikes. If you've watched ROAS fall off a cliff and opened Ads Manager with ten tabs going, you already know what happens next. Your gut says "creative fatigue." You kill your best-performing ad. A week later performance is still broken, because that was never the problem. Guessing wrong is the most expensive move in paid social. This workflow ends the guessing: → One agent investigates each competing theory — creative fatigue, budget and delivery changes, traffic quality, offer and seasonality → Each one is blind to the others, reasoning only from its own slice of the data so they can't bias each other → A refuter agent then attacks every surviving theory and tries to kill it → A theory only stands if the data can't disprove it → You get a ranked diagnosis: the real cause, the evidence for and against it, and the one move to make this week No anchoring on the first obvious answer. No killing winning creative on a hunch. No "here's what happened" reports that never tell you why. What you get: → Every theory tested in parallel instead of one biased guess → An adversarial pass that kills the wrong answer before you act on it → A ranked diagnosis with confidence levels and evidence both ways → A reusable workflow you drop next month's export into and re-run Built 100% in Claude Code with the new dynamic workflows. The first account I ran it on looked like textbook creative fatigue. The workflow disagreed, and traced the real cause to a budget change that had doubled spend and flooded delivery with junk traffic. I put together a full playbook with the exact workflow, the prompt, and how to run it on your own account. Want it for free? > Like this post > Comment "META" And I'll send it over (must be following so I can DM)show more

Confession: We were wrong about OpenClaw🦞. We thought that... giving each teammate a personal agent was the future. But a few weeks in, logins expired, integrations broke, and our Claws became a pain to maintain. We killed the last of our Claws last month (RIP) and pivoted to a shared company agent. The Every Agent lives in Slack, so the whole team uses it. However, if you still need a personal Claw to do recurring tasks, are willing to maintain it, and want more control than the company agent will let you, setting up a personal agent may still be worth it. We’ve just updated our OpenClaw guide to reflect where they tend to break and how to limit their authority:show more

At Uber, design source of truth was a big... problem for design and engineering teams. No one knew what the app truly looked like. So every week, a group of designers would sit in a meeting room and check that engineers had correctly implemented the Figma designs. Now coding agents have increased the throughput of changes by an order of magnitude. Keeping up manually has become impossible. Here we used the Revyl to map every state in Ubert (demo uber), navigating our mobile use agent on a cloud iOS simulator. This would have taken a team of designers tens of hours to recreate by hand. We did it in under an hour, asynchronously, with a simple prompt. Let your team see what your users actually see. Empower coding agents to ship a delightful experience with no blindspots. Get started with our free trial and create a map for your own app.show more

Created this mixed media video using the 'invideo editor',... just by chatting with an editing agent. ~ Asked the agent to break up the original video into individual frames (for this one it was 12 fps) ~ Described the mixed media style we wanted in detail to the agent (you could also give the agent reference images and videos that you feel work best for the video you're creating), ~ And then asked it to generate each frame using an image-gen model and line it back up in the right place as a separate layer. Did all this in ~30 minutes.show more

🚨 one person can now do the work of... an entire creative team. i just tested it on a real one. a friend needed an ad for his brand, so I opened the new Runway Agent 2.0 to try it out. here's how it went: → it generated the music and the key image first, so I could approve the direction → once I gave the ok, it built the full video around it → and when something was off, i changed just that one piece, without redoing the rest one prompt, and I had the ad we needed, work that used to take weeks. this is what it made 👇 if you want to try it → · 30% off 3 months with code RUNWAYAGENT — made with Runway · #MadeWithRunway · #adshow more

Ever watch an agent solve a problem with 100%... confidence and get it completely wrong? You can fix it… if you catch it. But then you need to watch every… single… run. New paper: We show that a meta-agent can do this for you, so your agents get ~2x more done on their own. 🧵show more

We share our early research on building Software World... - a "GitHub" run by agents. We deploy agents for Python packages in a dependency chain, and each agent is tasked with collaborating with others and optimizing the package it owns.show more

Everlyn isn’t just a video model. It’s a system... you can trust. 🌀 At its core is the Lyn Protocol, the onchain layer that gives AI video permanence, provenance, and ownership. Here’s what it means: 🔹Decentralized inference: Videos are not generated by a single company’s servers. They run across a distributed network of nodes. 🔹Agent ownership: When you summon your agent, it is minted onchain. It belongs to you, not to a platform. 🔹Video provenance: Every render carries a timestamp and fingerprint written to the ledger. No one can fake or steal it. The Lyn Protocol transforms AI video from content into verifiable digital memory.show more

sorry, they just did WHAT someone gave a machine... one disease name, the leading cause of blindness in the developed world with 1.5 million americans already in its path, and it came back pointing at a drug that has sat in pharmacies for years under a different label: 551 papers read in 30 minutes against the 294 hours a human would have needed, and the loop that did it is public on GitHub most agent setups answer one question at a time, so the ceiling on the work is the quality of the question you happened to think of this one was handed a single question and wrote the second one itself. turns out that follow-up is where the real find was: a target called ABCA1, upregulated threefold, in an experiment no human ordered i read the whole paper looking for the trick, and the trick is structural. that is the second question, and it is the gap between an assistant and a factory: - hand the loop a field rather than a task: it was given a disease, and choosing the mechanism was part of its job - make it rank before it spends: 151 papers in, ten candidate mechanisms out, scored against each other before anything touched a bench - split reading from judging, so the agent that forms the theory is a different agent from the one grading it - close every cycle on physical reality: the verdict was an experiment, and another model's opinion was never allowed to stand in for one - feed each result back as the next question rather than a log line, which is the step almost nobody builds - search what already passed inspection first: the winner was an approved compound with a safety file already on record - write down what the round learned before opening the next one, so round two starts where round one stopped my read, and i think it is the uncomfortable one: reading was the entire bottleneck in that field, and everybody spent the decade optimising the writing. people ran every physical experiment here, the analysis agent needs a domain expert writing its prompts, and the authors decline to call this the leap it resembles. the thinking got replaced, and the hands did not so the question i cannot answer for my own setup: which step of your loop still stops dead until you sit down and type something bookmark this one. the four parts that turn one model into a line that runs like this, the queue, the rooms, the write permissions and the gate, are built file by file in the piece below ↓show more

New open-source agent harness just landed! I got early... access to TrueForge by TrueFoundry and have been running it locally for the past few days. The harness layer deserves as much attention as the model, and open source matters here because you can inspect the loop, run it on your own infrastructure, and swap to the latest or cheaper models. TrueForge handles the runtime work that makes an agent reliable. It drives the tool-calling loop, manages context, coordinates subagents, and executes code in a sandbox, with any model you choose. Every tool call re-sends the growing context to the model, so in practice the harness controls most of what an agent costs to run. A few things stood out from my testing and their published benchmarks. Vendor-Neutral by design. It runs OpenAI, Anthropic, and Google models alongside open-weight models like Kimi, GLM, and DeepSeek. Model routing is a setting, and you can send each task to the model that fits it. On a 14-task enterprise agent benchmark, it matched the accuracy of Claude Managed Agents running the same Opus 4.8 model at roughly 30% lower cost per run (3.8M tokens vs 10M for the same answers). Routing the same tasks to GLM-5.2 held accuracy and brought cost down by about 75%, around $3 per run instead of $12. Fully self-hosted and Open Source (MIT License). I had it running locally with one command, with sandboxed code execution working out of the box. It's time to own your agent harness. Thanks to TrueFoundry for partnering on this post.show more

I MADE MY AI AGENT 10X FASTER WITHOUT CHANGING... THE MODEL not a smarter model, not a bigger context window, not another clever prompt the same kind of AI that designs vaccines for viruses we have not even met yet was spending two minutes opening the wrong files just to hand me a brief from three months ago the problem was never capability, it was the scaffolding that piled up around my agent by accident, folder by folder an agent does not think in your categories, it searches from scratch every single time, and your tidy human folders are a maze to it the fix was almost stupidly small, one index file at the root of each big folder and a few numbers in front of the folder names slowest task dropped from 2 minutes to 26 seconds, fastest ones hit 10, zero model changes capability is cheap when the scaffolding around it is broken the article breaks down the whole system in 15 minutes ↓show more

At Uber a big problem for design teams and... engineering teams was design source of truth. No one knew what the app truly looked like, so weekly a bunch of designers would get into a meeting room and check that the engineers correctly implemented the figma designs that they made Now with coding agents throughput of changes has increased an order of magnitude and it has become impossible to manually keep up. Here we used the Revyl CLI to create a flow of every state in the Uber design, by navigating our mobile use agent on an cloud iOS simulator. This is something that would have taken a team of designers tens of hours to recreate manually; All done in less than an hour asynchronously with a simple prompt. Enable your team to know what your users are actually seeing and empower coding agents to give your users a delightful experience without any blindspots Get started with our new free trial and create a map for your own app 🗺️show more

we just released a new blog "Training a coding... agent using the OpenCode harness in remote HF sandboxes with TRL and OpenEnv" you can take a real coding agent (OpenCode), let it run its own tool loop against real coding problems, and train it with RL on the exact tokens it produced and every rollout runs in its own remote HF sandbox, so rollouts scale out beyond one machine the loop: - OpenCode owns its tool loop inside an OpenEnv sandbox - an in-sandbox proxy records the real token ids + logprobs, per turn - a hidden-test verifier scores the result, and that is the reward - TRL trains with AsyncGRPO, weights sync back to vLLM over NCCL blog + runnable example:show more

Today we shipped sling, an agent-first CLI for your... GitHub Actions 💫 Right now, if you ask your agent why a CI job failed, it goes and pulls the entire set of GHA logs and eats up your context window. 🪵 > sling why StarSling (YC X25) computes the run's logs and metadata, along with the evidence lines it matched to tell you why a run failed If you want to know what your GHA usage is, you have to go to the GH UI and click around to find it. 💸 > sling bill tells you how many minutes, how much you're spending, and across which runner labels Trying to understand what your slowest job is? You'll need to hunt around the GHA UI and even then you only get an average. 📊 > sling top and sling usage tell you runner minutes and cost across all workflows and repos, with p50, p95, p99, queue wait and a week over week trendshow more

Zuckerberg built his own AI agent to run Meta.... this man is literally becoming Tony Stark. it pulls data from every team inside the company so he can skip meetings, skip the chain of command, and make decisions faster than any human process allows. 78,000 employees have their own AI agents now too. one messages coworkers on your behalf. another acts as your AI chief of staff. their agents talk to each other in an internal network. humans optional. Meta also bought an entire social media platform built for AI agents to interact with each other. read that again. Zuck said he wants every person at Meta to have a personal AI agent. then every person outside Meta. the Jarvis era started.show more

Today America roots for Norway, and I need you... to understand exactly why. We love an underdog. It is not a preference. It is a genetic condition. Two hundred and fifty-three years ago, a handful of colonials with no navy, no army, and no chance decided to throw a tea party in Boston Harbor to spite the largest empire the world had ever seen. An empire the sun never set on. An empire that could have ended us in an afternoon. And when the greatest military on earth said you cannot possibly do that, our answer was, and remains, hold my beer. Then we picked a fight we had no business winning. Then we declared ourselves free before we had actually won anything. Then we went out and made it true. We turned 250 last week. Still here. Still loud. That is the whole American operating system. The impossible odds. The comeback. The redemption. We do not just like that story, we are that story, and we will root for it anywhere we find it. It is why a bunch of college kids beating a Soviet hockey machine in 1980 still makes grown men cry in bars. So today. Norway. Five and a half million people. A country with more sheep than starting lineups. Playing England, a founding nation of the sport, on the biggest stage there is. Now let me be fair. We love England. Truly. Best ally we have got, and I say that with a full heart. Though the Japanese are gaining fast, so do not get comfortable. We love the Scots. We love the Northern Irish. We love our drunk cousins in Ireland, and yes I am lumping you all in, because half of you moved here anyway and you hate the English for reasons we are all familiar with. Nobody loves Wales. Nobody knows Wales exists. Nobody can understand a word of it. But I digress. Because today is not about any of that. Today is about five million Vikings walking into a stadium against an empire, with everyone on earth telling them they cannot possibly do this. You know exactly what we say to that. Let’s row!!! NORWAY!!!!!!show more