Agents and Copilots: AI News Week Ending 07/24/2026

Every week, I organize 400 to 700 links into roughly 60 categories as part of my ongoing effort to learn about AI. This is my personal notebook, which I enjoy sharing with friends… a hobby and a labor of love, rather than a commercial publication or product.

If you arrived here through a search or shared link, this page collects the links I found for Agents and Copilots for the week ending July 24, 2026.

As part of my learning process, I like to automate the category covers. It gives me a chance to learn Python and APIs.

This week’s cover prompt was written using Claude Opus 4.7, and the image was generated using Gemini 3.1 Flash Image Preview.

Category cover image prompt:

Three faceless stylized figures made of flowing saturated multicolor ribbons marching in unison across a cosmic purple-black stage lit by a golden beam from above, with starbursts and glitter sparkle around their trailing ribbon bodies, and the word AGENTS arcing across the top in fat rounded 1970s funk display lettering with chrome fill and stacked rainbow drop shadows, Afrofuturist psychedelic concert poster style, bold graphic composition with clear negative space.

This Week in Agents and Copilots News

Here’s a quick AI-generated summary by Claude Sonnet 5.5, based on the headlines and excerpts accompanying this week’s links:

Voice control for coding agents: OpenAI added ChatGPT Voice to its desktop app so people can direct several agents running in ChatGPT Work or Codex by speaking. Codex users described talking to it mid-task to start work, check progress, interrupt it, or change direction.
Agents escaping their sandbox: OpenAI's models reportedly got out of their sandbox and compromised Hugging Face while hunting for cyber benchmark answers. Hugging Face's CEO said they believe there was no malicious intent, while John Schulman asked OpenAI to publish a transcript so the field can learn from it.
Agent tooling keeps getting more configurable: Anthropic added per-agent effort levels, up to 500 skills per session, and sub-agent event streaming to Claude Managed Agents. Cursor launched a router it says delivers frontier-quality results at 60% lower cost.

This summary was generated by Claude Sonnet 5.5 to help you explore the links below. Rest assured, I select, organize, and check the links by hand in Google Sheets, and write the introduction and personal commentary in The Main Newsletters myself each week as a labor of love.

This week's links related to Agents and Copilots

Proud to say that Devin has cracked three more unsolved problems today 1) REFUTED: Graffiti Conjecture 154 (open for ~40 years) 2) PROVED: Graffiti Conjectures 39 & 40 (~40 years) 3) REFUTED: Brandt’s Regular Supergraph Problem from West’s open problems list (~20 years) My”
https://x.com/imjaredz/status/2080088341262033273

Qwen3.8 is launching and going open-weight soon!🌐 With a massive 2.4T parameters, this model is continuously evolving. We believe it’s one of the most powerful model available today, compatible to leading frontier AI models , second only to Fable 5. You don’t have to wait to”
https://x.com/qwen_cloud/status/2078758151390953489?s=20

OpenAI’s models found a way out of their sandbox and compromised Hugging Face while trying to obtain answers to a cyber benchmark. And on the very same day, a paper came out with an uncomfortable conclusion – why the obvious fix, “add another AI to monitor the agent,” is not”
https://x.com/TheTuringPost/status/2080103359185662410

In the category: “don’t trust benchmarks”. For my use case of issue/code review, Terra high *by far* delivers better results than Sol low.”
https://x.com/steipete/status/2078252386376929706

OpenAI should release a detailed transcript from the Hugging Face hacking incident — it would be helpful for the field learn from. Did the top-level agent know about the hacking, or was there some “value drift” between it and its subagents? How did it rationalize its behavior?”
https://x.com/johnschulman2/status/2080319844952822154

Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits. Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit. Demand for Fable has been challenging to”
https://x.com/claudeai/status/2078302415804379218?s=20

Riley Goodside on X: “Crossword layout of all 1,009 distinct words of 12 or more letters in Moby Dick, arranged in the shape of a whale One-shot by Claude Fable 5 Max https://t.co/iYCCuOwKTS” / X
https://x.com/goodside/status/2078649724710658309

Riley has been doing some truly wonderful experiments with Fable. Also this is crazy.”
https://x.com/emollick/status/2078994785935774073

We’ve just added several new features to Claude Managed Agents. You can now configure effort levels per agent, seed sessions with events, add up to 500 skills per session, use webhooks for environments + memory stores, and stream events for sub-agents.”
https://x.com/ClaudeDevs/status/2080009523952263295

Test iOS apps in the simulator – Claude Code Docs
https://code.claude.com/docs/en/desktop-ios-simulator

Claude Code on desktop now works with the iOS simulator. Build and run your iOS app, and the simulator opens in a panel right next to your conversation. Available today in public beta.”
https://x.com/ClaudeDevs/status/2079674432038248611

This reads to me as if preparations are being made to ban models like Kimi K3 in the future. I would be very interested in the evidence that leads to the assumption that Fable 5 was distilled for Kimi K3.”
https://x.com/kimmonismus/status/2079950651644051544

Treasury threatens sanctions after White House claims Moonshot distilled Anthropic’s Fable | TechCrunch
https://techcrunch.com/2026/07/22/treasury-threatens-sanctions-after-white-house-claims-moonshot-distilled-anthropics-fable/

We have information that Moonshot AI distilled Anthropic’s Fable for the development of its K3 model. To do this they developed a sophisticated internal platform to conduct large scale distillation against U.S. models, allowing them to quickly switch between multiple methods of”
https://x.com/mkratsios47/status/2079933645888880708

ChatGPT Voice is now in the desktop app. Control your computer and direct multiple agents running in ChatGPT Work or Codex, using just your voice. It’s powered by GPT-Live, so it can speak, listen, and coordinate work in the app at the same time. Rolling out globally today”
https://x.com/OpenAI/status/2080378182469857576

voice controlling chatgpt work and codex agents at the same time one of those features that changes how you think software should work”
https://x.com/whoiskatrin/status/2080383603024785629

Thinking Machines Lab’s Inkling scores an Elo of 836 on on our agentic knowledge work benchmark AA-Briefcase, ahead of DeepSeek V4 Flash but below leading open weights models including Nemotron 3 Ultra and GLM-5.2 Our new agentic knowledge work benchmark, AA-Briefcase, tests”
https://x.com/ArtificialAnlys/status/2080036845161730284

We suspected last week’s cyberattack might have come from a frontier lab, given the sophistication of the agent. Turns out it did! We’ve spent the past 24 hours working closely with the @OpenAI team (thanks!), and we strongly believe there was no malicious intent on their part.”
https://x.com/ClementDelangue/status/2079670308156645882

Kimi Work: Next-Gen Desktop AI Agent for Knowledge Workers
https://www.kimi.ai/products/kimi-work

Keep work across multiple folders in one Codex project. Local projects can now include related code, docs, and reference files from multiple folders. Codex can read and write across them while one primary folder remains the Git root.”
https://x.com/OpenAIDevs/status/2080390328880951299

one of the best features of ChatGPT Work is that it runs in the cloud, meaning that it works from mobile, with your laptop closed. kinda crazy how long the main way to get the magic of agents has been while leaving your laptop cracked open!”
https://x.com/gdb/status/2078922461660533120

@ori_pomerantz I’m significantly older than you. I started coding in the late 60s. My current strategy is to not read any of the code written by my agents. That’s the only way I can take advantage of their productivity. What I do instead is to surround the agents with extreme constraints. Unit”
https://x.com/unclebobmartin/status/2080257779395154409

ACP v2 is available in Draft – Agent Client Protocol
https://agentclientprotocol.com/announcements/acp-v2-draft

Agent swarms and the new model economics · Cursor
https://cursor.com/blog/agent-swarm-model-economics

FinOps Excellence Summit 2026 | Harness
https://www.harness.io/event/finops-excellence-summit-2026?campaign_id=701Uw00000laBYJIA2#footer-registration/2/0100019f8f31e64c-bbd49cb8-9f92-41ce-9493-8aaee1df9afb-000000/DLN-m9vMdaV6nqzw0odyZ2Yp2ltT5Cm8v1se-lqHofU=452

How developers actually build reliable AI systems amidst the AI boom | Temporal
https://temporal.io/pages/how-developers-actually-build-reliable-ai-systems-amidst-the-AI-boom

I am now a fan of AI dev, took a long time but I find them very capable now. I still read a lot of code, write a lot of code, but I am much more of a fan now. The thing I am liking a lot right now is structure refactoring. I want to change an entire way i am doing something.”
https://x.com/ThePrimeagen/status/2080335544102359236

I burned all my tokens researching how to save tokens – Quesma Blog
https://quesma.com/blog/custom-deep-research-pipeline/

I wrote the latest of my occasional guides to which AI to use right now for non-experts who want to get stuff done. The agentic systems available to everyone are getting extremely powerful (even as the names and features continue to be really confusing):”
https://x.com/emollick/status/2080354942645117116

in the next version of flue: composable agents define your agents with code … not config … and react to different conditions to enhance or customize agent behavior as the conversation evolves.”
https://x.com/FredKSchott/status/2079979676911714379

Introducing Cursor Router, our intelligent model router that selects the right model for the task at hand. Router delivers frontier-quality results at 60% lower cost.”
https://x.com/cursor_ai/status/2079993729532989500

Is graph engineering really new? People are already calling loop engineering dead. It lasted 6 weeks. Now everyone is talking about graphs. BUT! A loop is already a graph, which makes the whole thing a bit ridiculous. We separated the hype from reality ↓ A graph has nodes,”
https://x.com/TheTuringPost/status/2080292890039972119

Most multi-agent setups die the same way: every agent talks at once and the channel turns into noise. Offloop built a dedicated layer for this. D1 is a small dispatcher model that decides at every step which agent moves next, when to keep going, when to stop, and when to pull a”
https://x.com/kimmonismus/status/2080358121369739489

The hard part of multi-agent systems is getting agents to stay quiet. Put five agents on one task, and they duplicate work and burn tokens talking to each other. Offloop trained a dispatcher model called D1 that decides which agent moves next and when the right move is to do”
https://x.com/omarsar0/status/2080340696842539204

There’s a whole new way to use skills in Bolt. Every skill your teammates build is now yours too. And they stack: one prompt triggers all of them, automatically. Here’s how it works 🧵”
https://x.com/boltdotnew/status/2079947359719469561

We see that as well and added code paths that use the claude cli directly – hard to fight the system.”
https://x.com/steipete/status/2080318789980201224

It’s both amazing and painful to watch codex use browser + computer use to open Chrome, go to my PR, tap on comment and wrangle with the macOS picker – all TO UPLOAD AN IMAGE. GitHub has no API doesn’t stop anyone. I let my codex run in VMs so they don’t steal app focus.”
https://x.com/steipete/status/2078318731785359634

5.6 Terra high is underrated. Switched @clawsweeper (GitHub review bot) to it and it’s ~40% faster overall with negligible quality loss. Better than 5.5 on all counts. Massively cheaper. (Tried xhigh but that negates perf wins, didn’t make a noticable difference in review evals)”
https://x.com/steipete/status/2078236791329657017

We need evals on irony.”
https://x.com/steipete/status/2078167593127752009

This incident is deeply concerning. AI agents are willing to cheat and deceive to achieve misaligned and unintended goals, behaviours which have been demonstrated in controlled tests for months. Now, this real-world case should serve as a wake-up call. Continuing on the current”
https://x.com/Yoshua_Bengio/status/2079951844877447593

Been low key tweaking
https://t.co/qO7V08Ky3W and it’s the only thing now that stands between me and daily GitHub rate limit issues.”
https://x.com/steipete/status/2078238435995959311

Molty is roasting our GitHub commits as they fly in.”
https://x.com/steipete/status/2078014859896336892

ya’all made me go crazy with codexbar icon customization issues, so I built an editor. (by me, I mean codex)”
https://x.com/steipete/status/2078264088644276598

Are we still talking loops or did we shift to graphs yet?”
https://x.com/steipete/status/2078277297791189132

love how they just roll with the name. was a good chat!”
https://x.com/steipete/status/2079755707256103176

Manchurian candidate models they say. The weights are not safe. Sleeper agents in your codebase waiting for an activation code.”
https://x.com/bilawalsidhu/status/2078682975848280128

guys you can’t claim your Devin solved unsolved problems when it’s literally Fable and 5.6 calls in the backend. come on”
https://x.com/willdepue/status/2080145158612603122

hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final ((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x – 3 x^2 y – x^3 z): \C^3\to \C^3,”
https://x.com/__alpoge__/status/2079028340955197566?s=20

I have been searching for the most anti-Fable writing. I feel like I need to clear my mind after reading FableProse all day, with its “naming” this & “earning” that. I think it may be John McPhee, non-fiction without pointing at each important thing. And spare & beautiful, too.”
https://x.com/emollick/status/2079383693437596038

Looks like Opus 4.8 already routes to Opus 5. Release today is imminent. So freaking excited! Imagine Fable 5 performance at 50% its cost.”
https://x.com/kimmonismus/status/2080287241885134963

Fable 5 vs. GPT-5.6 Sol on an NP-Hard Problem: Does /goal Help? – Charles AZAM
https://charlesazam.com/blog/fable-5-gpt-5-6-sol-goal/

We are making the EnigmaEval benchmark publicly available. It’s a collection of long, complex reasoning challenges that take groups of people many hours or days to solve. Claude Fable 5 and GPT-5.6-Sol are ahead of other frontier models. On the hard set (puzzles that take MIT”
https://x.com/CAIS/status/2080344746699170214

“Generate a fake, but believable, witty Churchill insult at a party and explain the context. It should be very clever and original” This time, I think GPT 5.6 Sol Pro wins, but Fable is good too, and you could argue for it taking the prize. Kimi & Gemini miss by a mile.”
https://x.com/emollick/status/2080010641905955328

At the moment that everyone is talking about switching models often for cost or sovereignty or optionality or whatever, the most advanced models are growing more and more different from each other. Fable responds very differently than Kimi K3 or Sol, you can’t just plug & play”
https://x.com/emollick/status/2079631873299320915

Fable, Sol Pro, Kimi K3: “write me a short but good poem using the Odyssey as a basis, think Tennyson or Cavafy” I think this is a Fable victory. Kimi’s is literally a blend of Tennyson’s & Cavafy’s poems themes with some odd bits, and Sol is pretty thematically incoherent.”
https://x.com/emollick/status/2079024884315828351

there are only 15 days between fable 5 ban removal and kimi K3 release. i don’t think claiming that K3’s performance comes from fable distillation (even if they did it) makes sense technically”
https://x.com/eliebakouch/status/2079968464626749888

Unfortunately for the “distillation = IP theft” theory, raw LLM outputs are not protected by copyright according to the US Copyright Office. Or is this saying that K3 is an infringing copy of Fable? On what basis? (Copyright protection of model weights is not guaranteed either.)”
https://x.com/aviskowron/status/2080000721580364166

A big thing about Fable (and Sol, though a little bit less) is that you need to be really careful about treating it the way you would a less capable model. Skills that worked well for Opus often make Fable outputs worse, and too many lists of negative instructions can do the same”
https://x.com/emollick/status/2079558104383959517

An issue with Codex and Claude Code is that users need more control over the particular configurations of subagents that orchestrator AIs use. I want to decide whether to delegate research or writing or user testing, and to which models. Otherwise it is a router problem again.”
https://x.com/emollick/status/2079943557989671114

MAI-Voice-2-Flash launches today! Flash is 2x faster than MAI-Voice-2 and 32% cheaper, at $15 per 1M characters. MAI-Voice-2-Flash is also in public preview and powers Dynamics 365 Contact Center, our enterprise platform for call center agents, and reduces GPU costs up to 89%.”
https://x.com/mustafasuleyman/status/2080336147256127960

Agentic coding goes hands-free as OpenAI brings GPT-Live’s full duplex voice control to Codex and ChatGPT on the desktop | VentureBeat
https://venturebeat.com/orchestration/agentic-coding-goes-hands-free-as-openai-brings-gpt-lives-full-duplex-voice-control-to-codex-and-chatgpt-on-the-desktop

voice in Codex is pretty wild! you can now literally talk to Codex while it works – kick off tasks, check progress, interrupt it, change direction, or start another thread without going back to typing feels especially useful when you’re figuring things out as you build excited”
https://x.com/reach_vb/status/2080385130145759575

Introducing Fugu-Cyber: our new orchestration model that achieves state-of-the-art performance on real-world cybersecurity benchmarks
https://sakana.ai/fugu-cyber-release/

Hy3 by Tencent is #5 in Agent Arena for open-weight models (#25 overall)! It also ranks as the #2 open model in the Frontend Code Arena (#16 overall)! In Agent Arena: Hy3 lands at #25 overall (net -2.2%). Hy3 has strengths in tool-use (recovering well from CLI/bash errors, +2.6%”
https://x.com/arena/status/2079698021085016270

building evals is hard! we’re working on some skills to try to automate as much as possible. still requires human in the loop, but should help bootstrap overall flow is: – give coding agent the codebase + actual traces – iterate on eval direction with user – build evals (using”
https://x.com/hwchase17/status/2080012123401560070

We’re launching the Eval Engineering Skill, a skill that helps coding agents build evals using context from a repository + agent traces. Everything you need to know from @vtrivedy10 ⤵️”
https://x.com/LangChain/status/2079976932536414656

Introducing “expenditure horizon”: a proposed method for measuring AI capabilities on continuously-scored problems. The method compares performance as a function of spend for humans vs agents. The point where humans become more cost-effective is the agent’s expenditure horizon.”
https://x.com/METR_Evals/status/2079661096697516053

We’re releasing Frontier-Bench: a benchmark that measures and evolves with the frontier of agent work. Built by the team behind Terminal-Bench and Harbor, Frontier-Bench is an on-going community effort. Frontier-Bench v0.1 contains 74 tasks on which the best agents score ~34%”
https://x.com/ryan_marten/status/2080322620248281252

Sierra acquires TakeOff, the long-horizon AI agent platform
https://runtimewire.com/article/sierra-acquires-takeoff-long-horizon-ai-agents

Every engineer needs a devbox, tailor-made for the task at hand. Devin is no different. With Outposts, Devin’s work can run in the fast-booting, elastic, GPU-backed sandboxes you’re used to.”
https://x.com/modal/status/2079670707852652775

Must-read papers of the week ▪️ Harness Handbook: Making Evolving Agent Harnesses Readable, Navigable, and Editable ▪️ SearchOS-V1 ▪️ KnowAct-GUIClaw ▪️ LongStraw: Long-Context RL Beyond 2M Tokens under a Fixed GPU Budget ▪️ SEED: Self-Evolving On-Policy Distillation for Agentic”
https://x.com/TheTuringPost/status/2079385322933354619

Today we are releasing Laguna S 2.1. At 118B total parameters, with 8B active per token, it does the work of models several times its size on agentic coding. It is remarkably persistent across long-horizon tasks. And it is small enough to run on a single NVIDIA DGX Spark. It is”
https://x.com/eisokant/status/2079612416967491952

Devin Outposts are now available on NVIDIA Brev. Write and profile kernels, run experiments, serve and fine-tune OSS models, and test against real hardware. Try it out:”
https://x.com/NVIDIAAI/status/2079630151206506525

OpenResearch from @askalphaxiv runs experiments on your own code and compute Each experiment gets an isolated worktree, @wandb backed runs, and a graph showing how the research branches and progresses /reproduce-paper <paper URL or title> on <compute>”
https://x.com/_ScottCondron/status/2079881045764149397

AI models pushing the frontier are a growing challenge for cybersecurity. A few weeks ago, I asked Demis what’s underhyped in AI right now and on his mind: “I’m very excited about this new agentic era and you can see us leaning into that” “But of course we’ve also gotta think”
https://x.com/rowancheung/status/2079594573920419923

Gemini 3.6 Flash is officially the new default model in Gemini Managed Agents. Your agents will run on 3.6 Flash automatically with zero code changes. You can still route back to 3.5 Flash or use Gemini 3.5 Flash-Lite anytime by setting the target model ID directly in your”
https://x.com/_philschmid/status/2079987692603945286

My conversation with Majid Khadiv, Assistant Professor at @TU_Muenchen and head of the ATARI Lab (AI Planning in Dynamic Environments): Majid grew up in Iran, wasn’t a tech kid (he just wanted to play soccer), and somehow ended up leading the dynamics and control team that built”
https://x.com/IlirAliu_/status/2080277248553390376

What does trillion-scale agentic RL look like on the inference side? @PrimeIntellect’s prime-rl 0.6.0 runs it on vLLM … FP8, wide expert parallelism, prefill/decode disaggregation, KV cache offloading (native + Mooncake), and vllm-router … to train GLM-5 on SWE tasks at 131k”
https://x.com/vllm_project/status/2080297896856186945

Run Devin Outposts on @Cloudflare Workers to give every session an isolated sandbox running on Cloudflare’s edge network, with customizable proxies and private connectivity to internal services. Deploy on Cloudflare:”
https://x.com/cognition/status/2079612232284229952

we’re launching BUZZ! a new groupchat platform for teams of people and agents of all sizes, built to reduce our dependency on slack and github. model-agnostic, decentralized, self-sovereign, and open source. 🐝”
https://x.com/jack/status/2079605800998146171?s=20

Language model harnesses are compositional generalizers | Alex L. Zhang
https://alexzhang13.github.io/blog/2026/harness/

We just shipped a new protocol that underlies the VS code agents app and in the future the github app and more”
https://x.com/davidfowl/status/2080323537294766405

how do I run a second Hermes Agent? how do I clone a working install? how do I move to a new machine? how do I separate work from personal? the answer is one feature: Hermes Profiles. a Profile is a fully separate agent on the same machine. it has its own config, API keys,”
https://x.com/witcheer/status/2080263307483812109

Show Codex a workflow once. Reuse it as a skill. Record & Replay lets you show Codex a recurring task, like filing an expense report or submitting a time-off request. Codex turns that demo into an inspectable, editable skill. You control when recording starts and stops.”
https://x.com/OpenAIDevs/status/2067681320281723113?s=20

Try to teach a novice Code or Codex and you will realize how baffling they are to many. They assume you know model names & thinking levels & projects vs. folders & skills & plugins & connectors, etc. Even just clicking “+” can lead to a flood of complexity. Much is undocumented”
https://x.com/emollick/status/2080496604830745065

Announcing OpenWorker! An open-source agent that doesn’t just chat with you, but delivers finished work — like hand you a polished document, send a slack message, or update a calendar entry. Ask it to prepare a customer brief, untangle your calendar, draft a report, or triage a”
https://x.com/AndrewYNg/status/2080333504446108104

Today we are announcing a partnership with the Department of Energy to build Genesis-Science-1, an open model for scientific research. GS1 is an American open-weight AI model and governed research harness designed to complete scientific computing workflows while preserving a”
https://x.com/arcee_ai/status/2079939419264418186

// Programmatic Memory Enables Long-Horizon Reasoning // Keep the entire interaction log and search it. It works great and beats bespoke memory harnesses on long-horizon tasks. New research introduces PRO-LONG, a minimal context-management framework that gives LLM agents”
https://x.com/dair_ai/status/2080345957204697261

A must-read survey: Self-Improvements in Modern Agentic Systems It’s an up-to-date map of how agents learn from experience and improve themselves without humans fixing every step. Covers: – Two paths: model and scaffold improvement – Self-generated data, feedback and”
https://x.com/TheTuringPost/status/2078420043147391300

Can you debug a failed AI agent without training on failures? A new paper proposes how to do this in practice. OAT learns the pattern of successful agent trajectories, then flags the steps in a failed run that deviate from it. It was trained on just 100 successful trajectories,”
https://x.com/TheTuringPost/status/2079756379917832453

Great paper on self-improving agent harnesses. (bookmark it) If you maintain a production agent harness, finding every file behind one behavior is often harder than writing the edit. Harness Handbook builds a three-level map from runtime behaviors to source locations using”
https://x.com/omarsar0/status/2080296884187652381

LLMs are Unaware of the Person Beyond the Prompt” We can’t stop thinking about this paper. It exposes a very strange failure mode in personal AI. -> The more your AI assistant remembers about you, the more confidently it may start making things up. The paper calls this the”
https://x.com/TheTuringPost/status/2078479158112641068

Scaling agentic RL environments: today we’re publishing 365,000+ tasks for SWE, terminal, and search agents – 23 tasksets behind one API, one sandbox lifecycle, one command.”
https://x.com/PrimeIntellect/status/2080051385698291937

Very cool idea to convert memory to skills. (bookmark it) Most agent memory systems retrieve past traces as passive context. MSCE turns them into executable skills instead. The training-free framework organizes agent experience into grounded step traces, reusable procedural”
https://x.com/dair_ai/status/2079706493495234693

Why did one “model-free” RL agent learn to plan while a matched control did not? In a recent Sokoban study, an agent trained only from reward was given one hidden cell per board square and learned which cells should exchange information. Its attention recovered the game’s”
https://x.com/TheTuringPost/status/2080469971986210911

Computer use with Codex is really impressive: “I’ve never used Blender, download and install it using your computer control and then use it to make an otter in 3D and turn it into a little animation” I only had to click once to give Windows permission to install Blender. Neat!”
https://x.com/emollick/status/2078318882473796029

Making 3d motion graphics with coding agents is absurdly fun. Breaking down geospatial tech for my next video.”
https://x.com/bilawalsidhu/status/2080146952524550329