Image created with Gemini. Image prompt: {“category”:”Agents”,”short_title”:”The Errand-Runner’s Many Faces”,”concept_description”:”A small delivery courier in an apron hurries down a cobbled seaside street carrying a stack of grinning porcelain masks strung on a pole, each mask slightly different, suggesting one agent wearing many personas as it runs errands across town. A bakery, florist, and post office line the street, hinting at the many tasks an agent performs on behalf of its user. Above a shop awning, painted lettering spells the category name. The warm golden-hour light contrasts with the uncanny row of identical smiles bobbing behind the courier.”/,”image_prompt”:”Hand-painted Studio Ghibli style wide square illustration of a young courier in an apron running down a sunlit cobbled seaside street, carrying a tall pole strung with a row of grinning porcelain carnival masks that trail behind like lanterns, passing a bakery, florist, and post office with terracotta rooftops and harbor views, a wooden shop sign overhead painted with bold clean legible letters spelling ‘Agents’, soft gouache watercolor backgrounds, cel-shaded characters, warm golden-hour light with deep plum shadows, cozy but quietly uncanny mood.”}
“Loop engineering” is a hot buzzphrase after mentions of it by Boris Cherny (Claude Code’s creator) and Peter Steinberger (OpenClaw’s creator) went viral on social media. Loops are now a key part of how we get AI agents to iterate at length to build software. In this letter, I’d”
https://x.com/AndrewYNg/status/2071988145667928442
OpenClaw is now on iOS + Android 🦞 📱 Native mobile apps, finally 💬 Agents in your pocket 🔔 Channels, tasks, replies on the go Run agents from wherever your thumbs are. iOS:
https://t.co/7LHHc9htgM Android:”
https://x.com/openclaw/status/2071688039114342592
Apparently we didn’t talk enough about w̶o̶r̶k̶f̶l̶o̶w̶s̶ loops yet! See ya there!”
https://x.com/steipete/status/2072143124496097302
Good morning @aiDotEngineer ! Are we talking loops today?”
https://x.com/steipete/status/2071972449277898924
Finally, we’ve added a new Managed Agents Observability tab in Console. This provides session-level metrics such as input/output token use and tool usage.”
https://x.com/ClaudeDevs/status/2072058433097122145
I’ve been getting a TON done with Fable today and I’m not hitting rate limits. Wanted to share some tips on how I’m doing that 1. I only use Fable on “high” effort for now. xhigh is token hungry. max/extra is a furnace with worse outputs than lower options imo 2. I taught”
https://x.com/theo/status/2072481845363822914
Introducing Claude Sonnet 5, our most agentic Sonnet yet. It makes plans, uses tools like browsers and terminals, and runs autonomously at a level that just a few months ago required larger and more expensive models.”
https://x.com/claudeai/status/2072017450611142835
It’s been 18 days since Fable 5 was banned. Kind of insane. I genuinely thought it would be back after a few days.”
https://x.com/theo/status/2072058513669693608
My one serious piece of advice having used Fable a bunch before release is that, unless you are careful it develops its own internal bizarre cadence & dialogue over long tasks. If you aren’t asking it report in plain language, this starts to creep into everything, including menus”
https://x.com/emollick/status/2072543365124481045
New Claude app strings suggest Anthropic is preparing to put Fable 5 behind a separate usage-credit system billed outside existing plans, with credits added only after identity verification. Anthropic previously said identity verification was unrelated to Fable and limited to”
https://x.com/kimmonismus/status/2071868011804266828
Redeploying Claude Fable 5 \ Anthropic
https://www.anthropic.com/news/redeploying-fable-5
We’ve added a few updates to Claude Managed Agents: Streaming session event deltas, per-session agent overrides, new webhook event types, reverse pagination, and credential injection scoping.”
https://x.com/ClaudeDevs/status/2072058428424589412
Since June 12, we’ve been working closely with the US government to restore access to Claude Mythos 5 and Fable 5. Today, the government notified us that Mythos 5, our strongest cybersecurity model, can be redeployed to a set of US organizations that operate and defend critical”
https://x.com/AnthropicAI/status/2070665903440871779
We’ve received notice that the Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5. We’ll begin restoring access tomorrow, and will share an update soon. We’re grateful to our users for their patience, and to everyone who worked with us on”
https://x.com/anthropicai/status/2072106151890809341?s=46
This is crazier than you might think: Fable-5 now scores 16.10% on the Remote Labor Index What is RLI? The Remote Labor Index uses 240 real remote-work projects from professional freelancers, covering 23 domains and more than $140,000 of human work. Each task comes with the”
https://x.com/kimmonismus/status/2072376968729817531
Hark is building the next generation of computer-use agents Our goal is simple: use any computer as well as a human. Think of it as a digital humanoid that can navigate the entire internet We’re hiring – consider joining the CUA team with Tanmay”
https://x.com/adcock_brett/status/2070674685575209066
surprised more people aren’t doing something like this Codex now creates a “newspaper” for me every morning Unread messages, calendar, surf report, news Anything I can do to stay off my phone until later in the day is a priority”
https://x.com/doooyle/status/2072351019913171442
This is a fascinating and important set of data which shows us where things are going, using OpenAI as a canary in the coal mine. The chatbot era is over, and agentic systems are coming to tasks beyond engineering. And skills show promise as a way to standardize AI use in firms.”
https://x.com/emollick/status/2070171580030656744
Work at OpenAI is being transformed by agents, in every department. Across our entire company, people are using Codex to do work that is more complex, longer-running, and increasingly cross-functional. Our internal usage offers an early look at how agentic tools may reshape”
https://x.com/OpenAI/status/2070196105745518913
A humanoid robot autonomously executes a full long-horizon task, from ONE natural language command… for the first time. It goes downstairs to get a snack package, rides the elevator upstairs, opens the box, and puts the snacks into a drawer. The platform integrates”
https://x.com/IlirAliu_/status/2071652920135921913
Fable: “Last and First Men is out of copyright. I want you to make a movie that features a reading of it with appropriate mixes of animation and images using access to the APIs you have (elevenlabs, hugging face) . Give me the first 10-15 minutes, ending at an appropriate break.
https://x.com/emollick/status/2072872373758382497
Agents as Webs of Beliefs — LessWrong
https://www.lesswrong.com/posts/M39Z2CvyfaxZdaxR4/agents-as-webs-of-beliefs
before @geoffreylitt ‘s talk I was 90% given up on sloppy work around me, now i am more hopeful. the future will be very polarized: those who understand will keep having the next big idea. those who delegate understanding will be replaced by the agent.”
https://x.com/lily_gpupoor/status/2072469046000496963
everything is a ralph loop
https://ghuntley.com/loop/
Excited to release OpenWiki! OpenWiki is the easiest way to generate and maintain documentation in your codebase, built specifically for agents to consume Get started with a single command: `openwiki –init`”
https://x.com/BraceSproul/status/2072375499125596262
Factory | Agent-Native Software Development
https://factory.ai/
Formal organizational structures are a useful way to think about the challenges of agents. They provide a template to thinking about how work gets delegated up and down between smart expensive agents & cheaper weaker ones, as well as between narrow specialists & generalists.”
https://x.com/emollick/status/2072410869426733211
Framer: AI design agent
https://www.framer.com/?dub_id=byUuuLJlwJZq56Qe
if i were building an agent from scratch with memory i would use a wiki structure, it’s simple and extensible”
https://x.com/sydneyrunkle/status/2072311589072486879
If you’ve ever wondered why we will need 100X more AI inference in the future, and what it’s going to be driven by, this is another good example. Devin pushes forward an idea of agentic mapreduce, which means we’ll now have swarms of agents that are processing large amounts of”
https://x.com/levie/status/2072519377371459836
Imagine if, as an engineer, your working memory of a codebase was wiped every time you started a new ticket Sounds ridiculous, but this is functionally what happens to agents whenever we start a new thread Maybe your company is ahead of the curve and has committed context like”
https://x.com/caspar_br/status/2072420582717858292
Impeccable: The missing design vocabulary for agents
https://impeccable.style/
Introducing CoreWeave ARIA, the first AI research agent that runs autoresearch in your W&B dashboard. It reads your runs, finds what’s working, and launches the next experiment itself. See it on @karpathy’s nanochat, proposing configs and launching real training runs. Watch👇”
https://x.com/wandb/status/2071603727585448025
Memory has somehow consistently been the most exciting area of agent development over the last 3 years (imo), and it’s still a largely unsolved problem!! Wiki’s are the biggest advancement I’ve seen in a long time, and I’m super excited to push more on research here. Great”
https://x.com/BraceSproul/status/2071982037276475502
Oz Cloud Agent Platform: Every agent at your command | Warp
https://www.warp.dev/oz
Stay in the loop with Live Activities, and get notified when an agent finishes or needs your input. Review demos and diffs before merging PRs from your phone.”
https://x.com/cursor_ai/status/2071641104869691671
The longer the task, the more orchestration carries it. The OpenHands SDK just added three primitives for building robust, long-horizon agent workflows, usable by the model and by you: Thanks to: @gneubig @juanontech Vasco Alona @OpenHandsDev”
https://x.com/rajistics/status/2070555095725457494
This is a neat idea around doing model routing and sub-agent delegation while ensuring cache hits on accumulated context for all agents. It makes a lot of sense: you want to ensure that all subagents can also make use of accumulated context in their cache”
https://x.com/jerryjliu0/status/2071737452323303750
This is exciting and all, but the most interesting part for me is this: Auto Research, powered by @cursor_ai agent. The AI agent independently wrote code, ran experiments, analyzed the results, and improved the model’s Word Error Rate by up to 19.8%, vastly outperforming”
https://x.com/stalkermustang/status/2071590526965502027
Warp — The Open Platform for Automating Development | Warp
https://www.warp.dev/
we just released dynamic subagents, which let your agent programmatically orchestrate subagents in a code interpreter. this lets agents do work at scale that tool calls can’t reliably handle, like processing hundreds of documents or analyzing thousands of datapoints!”
https://x.com/sydneyrunkle/status/2071632107026174364
[2606.26907] Qwen-Image-Agent: Bridging the Context Gap in Real-World Image Generation
https://arxiv.org/abs/2606.26907
Qwen-Image-Agent Alibaba’s agentic framework bridges the context gap in real-world image generation It plans, reasons, searches, and remembers to build the precise context that text-to-image models need”
https://x.com/HuggingPapers/status/2070489753573548365
Got excited about the 52′ Dell, but it requires BetterDisplay hacking to get a usable resolution, but then I get weird flickers and also startup wake issues. [on M4 Max MBP] So, Apple wins this round, 2xXDR it is.”
https://x.com/steipete/status/2071034256051097799
Thank you, @colinsolvely, for helping move the needed iOS and Android UI improvements forward with this post. We want to meet this feedback in the same spirit. Please share what is not working here or in Discord so the team can listen, respond, and turn that feedback into”
https://x.com/openclaw/status/2072774467538043177
Was thinking if I should highlight this tweet or not, but it’s a masterclass in the amount of vitriol people face when working on open source. Is the app great yet? No. It’s a start. It was built by the community. Getting the iOS and Android apps working with secure pairing and”
https://x.com/steipete/status/2071972239734616146
I love how Apple notarization breaks multiple times a year until I manually log in and accept some new legal agreements.”
https://x.com/steipete/status/2070626638887555227
Pointed codex at some Twitter feedback on the OpenClaw iOS app and it did a first improvement pass. It’s still not good, but for two prompts it aint bad. Especially cool how it uses computer use to add before/after screenshots, as there’s no GitHub API.”
https://x.com/steipete/status/2072439279520039380
We getting Apple Car after all”
https://x.com/steipete/status/2071017970953052160
Episode 2 of OpenClaw’s official podcast @steipete and @somalley108 join @hrudolph and @Pat_Erichsen to discuss OpenClaw stability, security, the future of OpenClaw, and the community’s response to the recently released OpenClaw mobile app.
https://x.com/openclaw/status/2072422934761910276
Will be talking about the “app situation” and roadmap plans with @hrudolph live TODAY at 11:30 AM PST on the official OpenClaw podcast, live on Discord.
https://x.com/steipete/status/2072373527970562495
WF2026: Software Factories & Keynotes ft. Microsoft, OpenAI, OpenClaw, Z.ai (GLM), MiniMax, HF – YouTube
https://www.youtube.com/live/htM02KMNZnk?t=18296s
The AI Agent stack in 2026 ▪️ OpenClaw → local agent as personal control plane ▪️ Hermes Agent → local agent with self-improving skills ▪️ Gemma 4 → efficient models for local agent workflows ▪️ Skill Engineering → reusable agent abilities ▪️ VLA Models → interface for”
https://x.com/TheTuringPost/status/2071059725257732309
Word of the day so far at AIEWF is Loop. @swyx talked about “loopcraft” in his opening address, and the word was used constantly by the following speakers from Microsoft and OpenAI, and then “the clawfather” Peter Steinberger.”
https://x.com/latentspacepod/status/2072003484120203362
Asked codex to download+transcribe all sessions from @aiDotEngineer and tailor them to my interests.”
https://x.com/steipete/status/2072433013624033282
As agents take on longer-running work, engineering shifts to setting direction, reviewing work, and designing better systems around the models. @steipete at @aiDotEngineer”
https://x.com/OpenAIDevs/status/2072036305442406772
Folks I found the reset button. @thsottiaux”
https://x.com/steipete/status/2072061089177539003
History teaches us that access blockage rarely stops determined users.
https://x.com/steipete/status/2071063588329193551
Honored to be part of @aiDotEngineer’s keynote today!”
https://x.com/steipete/status/2072037247122284825
How did I ever function without AI? cc chefcook @theo”
https://x.com/steipete/status/2072447453622882338
Just going to leave this here for the haters.”
https://x.com/openclaw/status/2072479970484744288
Never thought I give @Steve_Yegge a shoutout. He was just early, like most visionaries. Now everyone is building factories.”
https://x.com/steipete/status/2072532278476148881
Price per token != cost per task”
https://x.com/steipete/status/2072144627474579925
This would have been amazing a few years ago, but in the age of AI, what’s the point?”
https://x.com/steipete/status/2071769993151398074
v2026.6.11 has dropped. This release focuses on the rough edges that make OpenClaw feel less dependable: misplaced replies, stuck sends, reconnects, model setup failures, and more. Beware, this release is boring.”
https://x.com/openclaw/status/2072137139937325558
was this vibed”
https://x.com/steipete/status/2071770560875671831
We moved the changelog to our docs to be easier readable.”
https://x.com/steipete/status/2072137754683789515
Well this is happening..”
https://x.com/openclaw/status/2072375549721469325
wouldn’t that also make the tools better for humans”
https://x.com/steipete/status/2071017724495774049
sneaky, but also clever.”
https://x.com/steipete/status/2072233293035213052
I’m looking for a semi-private hack space for a few days in SF for me and some of the OpenClaw maintainers. We need to cook. DM leads my way 🙏🦞”
https://x.com/steipete/status/2072475858435276840
Been reading all sorts of posts about the best ways to develop workflows for Fable and it reminds me of how little we actually know about the best ways to organize work for long-running agents. Nobody has enough experience or has done enough testing to reach any real conclusions.”
https://x.com/emollick/status/2072539615836213387
Claude 5 has so far been the worst launch by Anthropic Fable 5 isn’t available and Sonnet 5 was nerfed to death”
https://x.com/scaling01/status/2072044421634281636
Claude Fable 5 is available in Devin. You can use Fable 5 in Devin Cloud’s Ultra agent, our smartest and most capable agent, which excels at long-horizon tasks and debugging. Claude Fable 5 is also available in Devin Desktop and Devin CLI.”
https://x.com/cognition/status/2072405137117548601
Claude Sonnet 5 is here. Top-tier performance on coding and tool use at Sonnet pricing, with a 1M context window. It’s the new default in Claude Code for Pro users, and available everywhere on the Claude Platform, including the API and Managed Agents.”
https://x.com/ClaudeDevs/status/2072018504392601762
Claude Sonnet 5 is now available in Devin Desktop and Devin CLI. Sonnet 5 pairs frontier-level coding performance with a more affordable price point and outperforms Opus 4.8 on FrontierCode Extended.”
https://x.com/cognition/status/2072022778144821292
Claude Sonnet 5 is now available in Droid. Available 1/3 off for all users until August 31.”
https://x.com/FactoryAI/status/2072021755619864778
Computer Use isn’t included in this release, more coming soon. Read more in the docs:”
https://x.com/ClaudeDevs/status/2071988883802444125
Conventional model routing sucks. It passes benchmarks but fails to write code you’d actually merge. Introducing Devin Fusion, a new hybrid-model harness for agentic coding. In testing, it reduces the cost of Fable-level intelligence by 35% and still feels good to use.”
https://x.com/cognition/status/2071624568465490170
Excited to finally share “sidekick” and “mid-session routing” with the world! Our latest methods for multi-model routing. Headlines: – 35% cost savings – Still frontier-level on day-to-day – Works very well with Fable 5 Hope this helps other agent builders out there!”
https://x.com/walden_yan/status/2071627241818399181
Fable 5 is back.”
https://x.com/claudeai/status/2072402636813607381
Fable in Claude Code is capable of really amazing things, including for non-coders, but the interface is not really designed for managing 5+ hour long autonomous tasks. Really hard to observe what is happening and intervene in real time, you often have to wait until the outputs.”
https://x.com/emollick/status/2072722575126319366
Fable: “you have Unity and access to MCP. I want you to build a game that is a unique twist on a FPS. You want the player to say “wow” & “so clever” and to enjoy the core gameplay loop” WebGL:
https://t.co/xB707DCpBp * It had no assets so the graphics are procedurally generated”
https://x.com/emollick/status/2072749816690483301
Great experiment testing how good AIs are getting at very ambitious end-to-end coding tasks. Opus 4.7, in 14 hours, was able to build a software package that would take 2-17 weeks of human engineering work. It cost $251. The models are still not perfect, but are improving fast.”
https://x.com/emollick/status/2070537422027342085
Here we go: Sonnet 5 is live: The tl;dr • Anthropic calls it the most agentic Sonnet yet • Near Opus 4.8-level performance, but cheaper • Strong gains in reasoning, tool use, coding, and knowledge work • Default model for Free and Pro users • Available in Claude Code and”
https://x.com/kimmonismus/status/2072019015577333804
I have no idea if this is a legitimate leak. But a *Sonnet 5* release today would surely go hand in hand with a *Fable 5* re-release.”
https://x.com/kimmonismus/status/2071941904636531167
Just tweeting while waiting for my Fable runs to finish.”
https://x.com/emollick/status/2072406024158650744
Now that Fable 5 is ready to build (again), we’ve reset everyone’s 5-hour and weekly rate limits.”
https://x.com/ClaudeDevs/status/2072429181565288665
Person familiar with the discussions between the admin and Anthropic says conversations are expected to continue over the weekend on Fable, which was not restored with this latest development. Also says that conversations have been focused not just on resolving this export”
https://x.com/cheyennehaslett/status/2070670490494976491
Rumors about Sonnet 5 have been circulating for a long time. Now it looks like Sonnet 5 will be released alongside the re-release of Fable 5 (presumably with ID verification). This leaves a bad taste in my mouth, because it raises the possibility that those who might not be”
https://x.com/kimmonismus/status/2071899142616408377
Since its back, here were my impressions I posted a couple weeks ago of Fable after my time as an early access user (yes, it really is very impressive, but that shows off best in longer, harder tasks)”
https://x.com/emollick/status/2072450372602311145
Sonnet 5 incoming today per leaker @synthwavedd – Knowledge cutoff January 2026 – Launching at $2/$10 per Mtok promotional rate* – 1M context variant available I assume we at least (!) get also fable 5 news.”
https://x.com/kimmonismus/status/2071953298169778636
Sonnet 5 is here! This is going to support better long-running agents. Previous Sonnet models were unreliable, so it’s great to see the improved version that can complete agentic tasks more reliably. It also seems to have improved substantially in computer use.”
https://x.com/omarsar0/status/2072022542521438300
Super interesting workshop from @rubrikInc → Tame Your AI Monsters: Claude Edition Rubrik is the first organization to put Rubrik Agent Cloud to work governing a live Claude implementation Now, they’ll open that experience up in a hands-on lab. You will: • Watch a Claude”
https://x.com/TheTuringPost/status/2070657528925704438
Through Aug 31, Sonnet 5 will use roughly 30% less quota than Sonnet 4.6 in Devin Desktop/CLI. After that, it will use the same quota as Sonnet 4.6. Try it:”
https://x.com/cognition/status/2072022784084000810
Dynamic workflows (generating harnesses on the fly) are a new form of test-time compute. But LLMs aren’t great at building them. I often have to steer agents to generate complex patterns. Curious how effective Mythos/GPT-5.6 is at dynamically generating complex workflows.”
https://x.com/omarsar0/status/2070596184339562946
It would be good to have an official government statement about the risks they saw in Fable, how they are viewing defensive preparations in light of coming open weights Mythos-class models, & whether the worries are state actors, independent hackers, etc It’s important to know”
https://x.com/emollick/status/2072392345128239475
The new Sonnet 5 achieves Opus 4.8 level performance on Terminal-Bench for less than half the cost. Importantly for –yolo coders, this model is also significantly better at refusing hijack attempts in prompt injection attacks. Available in Cline now!”
https://x.com/cline/status/2072051144436928727
omg omg GLM-5.2 is beating fable. china is catching up” chill out and listen to Lisan: > slightly ahead of Opus 4.5 > behind GPT-5.2, Gemini 3 Pro and Opus 4.6″
https://x.com/scaling01/status/2072346101068238946
Microsoft just released a new GUI agent on Hugging Face Sico-Evolution jumps from 39.8% to 82.9% Task Success Rate Outperforming GPT-5.4, Claude Opus 4.6, and Claude Opus 4.7 All from a 4B parameter model”
https://x.com/HuggingPapers/status/2071951218889339131
best case: OSS surpasses Mythos, and the gov stops banning GPT-5.6/Fable. worst case: OSS surpasses Mythos, and then decides to stop being open source.”
https://x.com/Yuchenj_UW/status/2070554908139659400
Claude Fable 5 is available again in Cursor. It leads all models on CursorBench, but is the most expensive per task.”
https://x.com/cursor_ai/status/2072403323844428217
Claude Sonnet 5 is ready for you in the Agent Arena! In Agent Arena, we measure models on millions of real-world, long-horizon agentic tasks from a global community of users. Models can access web search, filesystem, and terminal tools to complete complex workflows. The”
https://x.com/arena/status/2072035566829568111
Claude Sonnet 5 ranks second only to Fable 5 on AA-Briefcase, our new agentic knowledge work benchmark, with a ~17x cost per task range across its five effort settings @AnthropicAI has released Claude Sonnet 5, the latest addition to the Claude Sonnet family. On AA-Briefcase,”
https://x.com/ArtificialAnlys/status/2072427328689619241
Even though I made this graph, it is also kind of wrong. Fable is guardrailed Mythos. If we use the Mythos date”
https://x.com/emollick/status/2071459383868117371
Fable 5 is back in the Arena! When it first debuted, Fable 5 ranked #1 in Agent Arena: our benchmark for real-world, long-horizon agentic performance. Agent Arena evaluates models on millions of real tasks submitted by a global community of users, with access to web search,”
https://x.com/arena/status/2072423538641031372
Oh my god, Sonnet 5 was MORE EXPENSIVE THAN FABLE to run the whole bench 💀”
https://x.com/theo/status/2072068395529576912
The talk about Mythos and cybersecurity was not, in fact, hype. (As anyone using Fable to do autonomous work has probably recognized)”
https://x.com/emollick/status/2072778376494895139
Claude Code turned every engineer into three. Now companies need more product thinkers | VentureBeat
https://venturebeat.com/ai/claude-code-turned-every-engineer-into-three-now-companies-need-more-product-thinkers
Open source AI is gaining traction due to the White House’s clampdown on Anthropic & OpenAI. “A lot of founders and developers have been burned by trying to build apps on top of these really advanced closed source models, like Fable for instance…” “… and then seeing those”
https://x.com/theinformation/status/2071700452605829433
An open-source OSINT agent for security research – OpenOSINT It’s like an AI investigation layer that lets you: examine emails, domains, IPs, phone numbers, GitHub profiles, etc. It exposes its OSINT tools through MCP. This agent decides what to run, chains findings, and”
https://x.com/TheTuringPost/status/2072126171270295892
Fable is very good, and having it taken away after getting early access delayed a few things I was working on and made me defer others until it returned But in my early access I also tripped the security guardrails a lot for non-obvious reasons. We will see how the new ones are”
https://x.com/emollick/status/2072165332370985026
Following conversations with the US government, we’ve updated our cybersecurity safeguards. The vast majority of coding work is unaffected. In the near term, the new safeguards will flag a slightly higher fraction of harmless requests than the previous Fable safeguards; we’re”
https://x.com/claudeai/status/2072402638247968855
“Agentic kernel optimization is the future of on-device inference” @xenovacom used Fable 5 to write kernels that pushed Gemma 4 to a massive 255 tok/s on WebGPU with M4. He shared the demo, so you can try in your browser!!”
https://x.com/googlegemma/status/2072416614188974274
I can personally attest: OpenClaude using GLM 5.2 is now performing on par with Claude Code powered by Opus 4.8.”
https://x.com/kevincodex/status/2070354383158861955
Fable, one prompt: “build an elaborate game that makes it feel like I’m a brilliant chess player without knowing anything at all about chess. It should make me feel like a grand master. Feel free to go as meta as you want but the more chess-y the better.”
https://x.com/emollick/status/2072547557968925093
Introducing Mixture of Agents 2.0 in Hermes Agent. Combine any provider’s models into a mixture of your own. Access your presets as if it were a normal model in Hermes. Big improvement in our soon-to-release HermesBench against opus and gpt-5.5 with MoA using Opus & GPT”
https://x.com/Teknium/status/2070615003674366277
This is the relevant section of my CLAUDE.md I’ll be real – I haven’t read it much. Just vibed out what I was looking for with Fable, and had it confirm it can use Codex for the things I care about I still find Codex to be WAY better at computer use, verification of UI/UX work,”
https://x.com/theo/status/2072482460122964067
while we cannot use fable 5 and gpt 5.6 will be screened before allowed to use the model, open source model builders are hard at work to close the gap. Ornith-1.0-397B from @ornith_, the best coding open source model.”
https://x.com/nathanhabib1011/status/2070469918475116750
Claude Fable 5 is once again available in Computer as an orchestrator model.”
https://x.com/perplexity_ai/status/2072433125104505226
New orchestrator model for Computer users (Pro and Max): Sonnet 5.”
https://x.com/AravSrinivas/status/2072031649693675810
Announcing the hosted X MCP – Announcements – X Developers
https://devcommunity.x.com/t/announcing-the-hosted-x-mcp/269558
Introducing Cursor for iOS. Build from anywhere by launching always-on cloud agents. Or remotely control agents running on your computer from the app. Composer 2.5 is 75% off in the app now through July 5.”
https://x.com/cursor_ai/status/2071641103191998810
Introducing ZCode, the official development environment for GLM-5.2 – GLM Coding Plan subscribers: now 1.5x usage quota in ZCode – BYOK supported: works with your existing subscriptions and APIs – Available on macOS, Windows, and Linux Download now:”
https://x.com/Zai_org/status/2072349453361557898
Build from anywhere with Cursor for iOS · Cursor
https://cursor.com/blog/ios-mobile-app
Arena has crossed $100M in annualized revenue run rate, eight months after launching our evaluation product. With our recent release of Agent Mode, millions of users on Arena are doing real work with agents, from coding to document analysis, in long-running, multi-turn sessions”
https://x.com/ml_angelopoulos/status/2071629882057228680
Arena reached a $100M annual revenue run rate just 8 months after launching our evaluation product. We started as a research project at UC Berkeley with a simple mission: measure AI progress through real-world use. As AI shifts from chatbots to agents taking on longer,”
https://x.com/arena/status/2071630464583151727
A free practical guide – Total Cost of Ownership for Evaluating Agents from @fiddler_ai It explains why agent evals at scale cost more than just API bills – with hidden costs like sampling risk, overhead, and the “Evaluation Trust Tax” Covers how to: – Evaluate every trace”
https://x.com/TheTuringPost/status/2071315398046412928
Today’s YouTube video is a longer look at MoA in Hermes Agent, trying to answer some of the common questions: How does using MoA impact cost, speed, and quality? I created an open weights MoA using GLM-5.2, Kimi K2.6 and Minimax M3 to find out, then made a three-headed Grok”
https://x.com/tonbistudio/status/2071595794147250540
OpenAI evaluated its new models on PostTrainBench-Lite, a shortened version of our original benchmark that gives agents 5 hours instead of 10 to improve an open-source base model. GPT-5.6 Sol and Terra outperform GPT-5.5, but still rely on narrow strategies and sometimes”
https://x.com/karinanguyen/status/2070577740022231232
We’re introducing GeneBench-Pro, a research-level benchmark for a harder kind of AI progress: how well agents can navigate messy biological data, choose the right analysis path, and make judgment calls that real computational research depends on.”
https://x.com/OpenAI/status/2072004836674167294
harbor is a great framework for running evals for long running, stateful agents its becoming industry standard, powering benchmarks like terminal bench 2 we’ve integrated deeply with harbor: across LangSmith, Sandboxes, and Deep Agents”
https://x.com/hwchase17/status/2071974139926294897
Introducing Agent Evals: Score your agents on real outcomes – Inngest Blog
https://www.inngest.com/blog/introducing-agent-evals
[2605.15846] RoadmapBench: Evaluating Long-Horizon Agentic Software Development Across Version Upgrades
https://arxiv.org/abs/2605.15846
[Token efficiency in Agent Arena] Agent Arena measures agent performance across a range of real-world tasks from our global community. Models get search, filesystem, and terminal tools to complete complex workflows: writing code, creating slide deck, researching the web, building”
https://x.com/arena/status/2070531800603238634
📢 1) We have a few papers that advance the state of the art of AI agent evaluation. Details and links in Stephan’s post. 2) AI agent evaluation has quickly become a distinct discipline. We’re working on a paper titled “Emerging trends in AI agent evaluation” that extracts best”
https://x.com/random_walker/status/2072375245969719374
Congratulations to @Etched on announcing their first inference system! We have seen early performance demos, and we’re looking forward to kicking off benchmarking. Our new system performance benchmark, AA-AgentPerf, measures how many agents a system can run per megawatt of”
https://x.com/ArtificialAnlys/status/2072254061244825981
Cursor · CursorBench
https://cursor.com/evals
Two years ago, we built OSWorld 1.0 — the benchmark that became the standard for computer-use agents. Agents now score 83.5% on it. Problem solved? Not even close. 🚀Today we introduce OSWorld 2.0: Benchmarking Computer Use Agents on Long-Horizon Real-World Tasks. What’s new:”
https://x.com/XLangNLP/status/2070517498974253269
What are the largest software engineering tasks AI can perform? To answer this, we built MirrorCode, our long-horizon SWE benchmark that lets AI code autonomously for days at a time. The best models complete some tasks we estimate would take human engineers several weeks.”
https://x.com/EpochAIResearch/status/2070528800941920263
11 sources to master AI agents and agentic reasoning ▪️ Agentic Reasoning for LLLMs ▪️ Agent Skills for LLMs: Architecture, Acquisition, Security, and the Path Forward ▪️ The Landscape of Agentic Reinforcement Learning for LLMs ▪️ Toward Efficient Agents: Memory, Tool learning,”
https://x.com/TheTuringPost/status/2071309540654743935
𝗚𝗟𝗠-𝟱.𝟮 (the latest open weights model) is having an Enterprise moment, and it is not an exaggeration.🚀 🔥 We have been impressed by how strongly GLM-5.2 is pushing long-horizon performance .. not just in coding, but also in 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗽𝗹𝗮𝗻𝗻𝗶𝗻𝗴, 𝘁𝗼𝗼𝗹”
https://x.com/RajeswarSai/status/2072006835444347390
@hwchase17 s “wiki memory” really resonates with how we’ve been thinking about Enterprise Knowledgebases. Agent memory is still early, but one pattern we observe consistently: enterprises don’t need agents to retrieve raw documents, Slack threads or tickets at query time. They”
https://x.com/bpalit/status/2072378273343082537
Today, we give robots a /skills library that self-evolves and compounds indefinitely! Introducing ASPIRE: a robot solving its 100th task is no longer as clueless as solving its first. Coding agents observe multimodal sensory traces from simulation and real robots, launch an”
https://x.com/DrJimFan/status/2072004190856212902
Excited to see @cohere open-source how they use AI coding agents to maintain their vLLM fork. 🙌 Keeping a long-lived fork in sync is tricky. Their agent treats it as a control loop: rebase onto each upstream release, run tests, diagnose, fix, repeat until green. Weeks of work”
https://x.com/vllm_project/status/2070364532296536346
A list of definitely must-read papers of the week ▪️ World Action Models: A Survey ▪️ ImageWAM ▪️ EvoEmbedding ▪️ Training Open Models for Agentic Phone Use ▪️ Learning from Your Own Mistakes ▪️ Learning from the Self-future ▪️ Causal Discovery in the Era of Agents ▪️ Zone of”
https://x.com/TheTuringPost/status/2070100698252083547
Calling it now. Agentic MapReduce will be one of the most important agent architectures of the decade. Exhibit A: Devin Security Swarm finds vulnerabilities cheaper than the next-most-accurate tool and fixes them too.”
https://x.com/jakejluo/status/2072380678419705949
Devin for Security is now public And it’s the most comprehensive tool on the market A Fortune 500 company we’ve been piloting this with has already found & fixed over a thousand vulnerabilities in their production repos”
https://x.com/walden_yan/status/2072377406267273248
Introducing Devin Security Swarm A more cost effective and accurate way to find security vulnerabilities in complex codebases, based on a new architecture: Agentic MapReduce.”
https://x.com/cognition/status/2072368168182432109
New paper on giving LLM agents experience that improves the weights and stays readable at the same time. Agent-experience methods split into two camps. Externalized natural-language rules stay interpretable but drift out of sync with the policy. Parameter updates generalize but”
https://x.com/dair_ai/status/2070589168837947693
Got GLM-5.2 running on my Mac Studio via llama.cpp, the reasoning behind all my medical agentic workflows. It orchestrates a swarm of tiny on-device OpenMed experts: oncology, meds, labs. No cloud, no rate limits, nobody can take it away. AI must be owned, not rented.”
https://x.com/MaziyarPanahi/status/2070503452178796704
If I had to choose just one metric, I’d argue that the KV-cache hit rate is the single most important metric for a production-stage AI agent.” – Manus AI prompt caching is important! read about how we do it with deep agents”
https://x.com/hwchase17/status/2070577381392482732
More and more developers turning to glm5.2 as a daily driver for coding Here’s how to use it with deepagents code”
https://x.com/hwchase17/status/2072344890755977571
Built a new agent app for DevRel content research powered by @Zai_org GLM-5.2. Takes your product + audience and turns developer demand into practical blog and talk ideas. Chat-first. Say something like: “I’m working on a Chrome extension for frontend devs that extracts SVGs,”
https://x.com/Astrodevil_/status/2071572680470655253
You can start building with GLM 5.2 in minutes: 1️⃣ Download dcode 2️⃣ Select your model (GLM 5.2) 3️⃣ Enter your API key …and you’re ready to go with frontier performance on open weights. Great demo from @its_ao”
https://x.com/LangChain/status/2072334663457067064
Zai going strong: they officially launched ZCode 3.0. The new AI-native coding IDE is deeply optimized for GLM-5.2 and supports agentic software development from planning and coding to code review and deployment. • Deep GLM-5.2 integration with multi-agent collaboration •”
https://x.com/kimmonismus/status/2072378141041991702
Here’s a question I find confusing and interesting and which actually tells us a lot about the nature of current AI progress: Why has progress on computer use been so slow? Computer use is so clearly verifiable. I think the answer is that it is not enough for a domain to be”
https://x.com/dwarkesh_sp/status/2070672008946589922
We are starting to role out our Trace Judge model to early partners today Designed to detect errors in agent trajectories at 1/100th of the cost compared to closed models If interested in early access, sign up on the form below:
https://x.com/hwchase17/status/2071630837976822237
Semantic search alone doesn’t cut it. Neither does brute-force grep. Agents need both. Today we’re shipping the Retrieval Harness in LlamaParse Index: semantic search, server-side grep, and file-level navigation working together in a single agent reasoning loop. 🦙🌤️ Grep a”
https://x.com/llama_index/status/2071656315210826006
First, Mark was clearly talking about the industry’s progress on agentic capabilities on the whole. But, while we’re on the topic: Our next Muse Spark update is coming soon. Big improvements in coding and agentic capabilities to be more competitive with other leading models.”
https://x.com/alexandr_wang/status/2072848108342677597
Check out Tonbi’s latest video to learn about Hermes Agents’ subagent delegation tool!”
https://x.com/Teknium/status/2070557376726634526
Agents are being adopted very quickly and accelerating work. How this looks across OpenAI itself:”
https://x.com/gdb/status/2070199649823297653
btw if u thought ur custom fanned-out data processing multi-agent orchestration harness was a particularly relevant long term moat, sorry, openai just added it as an api parameter”
https://x.com/tenobrus/status/2070573483319521423
digital ocean for running a codex remote session:”
https://x.com/gdb/status/2070281121594827025
Okay I owe my @OpenAI friends an apology for sleeping on Codex. I was not aware how strong your game was. This is… really quite something.”
https://x.com/QuinnyPig/status/2072132613196898714
We are giving all Codex users a usage reset on the house. Should be showing in your accounts in the next few hours. We have applied some mitigations, but our investigation hasn’t shown users being impacted at large. We are continuing to monitor the situation.”
https://x.com/thsottiaux/status/2070653282440405046?s=20
Your favorite Codex shortcuts are getting an upgrade. July 15th.”
https://x.com/OpenAIDevs/status/2071639953927438440
.@harborframework can now integrate directly with Deep Agents, LangSmith Sandboxes, and LangSmith Observability. You need to run agents in a real, reproducible, isolated environment, many times in parallel, with a deterministic check at the end. Harbor solves this problem.”
https://x.com/LangChain/status/2071978566691049559
.@mondaydotcom had one agent trying to handle 200+ tools. Context pollution everywhere. The LLM was confused, costs rising, and it still wasn’t working. @omribruchim on rebuilding Sidekick with Deep Agents:
https://x.com/LangChain/status/2070507927798993352
a common pattern for (a part of) memory that we see: Wiki Memory Examples: – DeepWiki (@cognition) – AutoWiki (@FactoryAI) – LLM Wiki (@karpathy)”
https://x.com/hwchase17/status/2071963841009942671
agentic map reduce” is a nice name spawning subagents programatically allows you to do these advanced agent patterns, where you want some determinism to control how agents are created/run check out dynamic subagents in deepagents for a way to do this:”
https://x.com/hwchase17/status/2072377816780624266
Deep Agents now supports dynamic subagents. Instead of invoking subagents with tool calls, the main agent writes orchestration code to coordinate work at scale. This enables workflows like processing hundreds of documents or analyzing thousands of datapoints with deterministic”
https://x.com/LangChain/status/2071631563897377010
Dynamic subagents in deepagents! Lets you spin up subagents programmatically We highlight 6 diff use cases for this”
https://x.com/hwchase17/status/2071633874736804066
The AI SDK Harness API now supports @OpenCode and @LangChain Deep Agents through a single unified interface. Use 𝙷𝚊𝚛𝚗𝚎𝚜𝚜𝙰𝚐𝚎𝚗𝚝 to run any supported runtime without changing your app code.”
https://x.com/vercel_dev/status/2070559261399339432
Warp Open-Sources Its Agentic Development Environment, Introducing a New Model for Building Software with Cloud Agents | Warp
https://www.warp.dev/newsroom/2026/4/28/warp-open-sources-its-agentic-development-environment
OpenWiki is the easiest way to document your codebase, built specifically for agents to consume. OpenWiki can generate repo docs, automatically update them as your codebase evolves, and do Q&A over your docs and codebase!”
https://x.com/LangChain/status/2072376975545798792
Devin Fusion | Cognition
https://cognition.com/blog/devin-fusion
From more details, you can see my blog on agentic RL (https://t.co/zg3LTAn35F) or the DeepSWE blog (https://t.co/RFurIkCLyJ), which provide a lot of info on this topic. For a great practical example, check out Prime Intellect environment hub / RL. This is the best open framework”
https://x.com/cwolferesearch/status/2070500073679552604
Agent-Assisted SGLang Development: An Initial Exploration – LMSYS Org
https://www.lmsys.org/blog/2026-07-02-agent-assisted-sglang-development
Micro-Agent: Beat Frontier Models with Collaboration inside Model API What if the next capability jump comes from the router, not a bigger model? Read our blog to checkout the exciting early results:
https://t.co/ohitadkuOd #vLLM #SemanticRouter #AIInfrastructure”
https://x.com/agenticin/status/2071494912277938398
Somewhere in my agent’s memory, two versions of the same fact are still arguing. That’s the whole problem with agent memory in one sentence. A bigger context window doesn’t fix it. It just delays when you notice. More tokens, more cost, and eventually worse retrieval as the”
https://x.com/PrajjwalYd/status/2072291317695324410
We’re excited to introduce the Retrieval Harness in LlamaParse – which is the 2026 version of RAG over documents Generalized agents need the right set of tools to scalably search and read through an arbitrary corpus of data (from 10 docs to 1m+ docs). They can already”
https://x.com/jerryjliu0/status/2071729856900215261
Great paper on managing agent skills. Skill libraries keep growing, and picking the right skills has become a bottleneck for coding agents. The defaults are to expose the agent to the whole skill collection, or retrieve skills with embeddings and rerankers. Both treat the”
https://x.com/omarsar0/status/2072430551446032847
AI went from predicting biological processes to controlling them, and now it is becoming an active participant in science. Agents’ architectures perfectly fit the scientific loop at the execution level: read literature, write and run code, generate hypotheses, test them,”
https://x.com/TheTuringPost/status/2072130350042534390
Katalyze – Agentic Infrastructure for Life Sciences
https://www.katalyzeai.com/
you can now build RLMs with Deep Agents! the more context agents accumulate, the worse they perform, a phenomenon called context rot. RLMs, proposed by @a1zhang from MIT, help: instead of working model turn by turn, the model runs code in a REPL that dynamically dispatches”
https://x.com/sydneyrunkle/status/2072348322526810594
Confidence-Aware Tool Orchestration for Robust Video Understanding”
https://x.com/_akhaliq/status/2070478699019804872
Most evaluations of world models ask, “Does this look right?” 🌎 We built WorldModelGym to ask a different question: if an agent actually uses this model to make decisions, does it still choose well? 🏋️🏃♂️🏁🦾 A simulation can look physically plausible and still lead an agent”
https://x.com/RekaAILabs/status/2072325792558956573
REST3D: Reconstructing Physically Stable 3D Scenes from a Single Image” TLDR: reconstructs simulation-ready 3D scenes from a single image by combining agentic physical scene understanding with physics-constrained optimization.”
https://x.com/Almorgand/status/2071984726512906605
Leave a Reply