J Toolman @j_toolman

Recursive Language Models (RLMs) let agents manage 10M+ tokens by delegating tasks recursively. This Google Cloud Community Article explains why ADK was the perfect choice for re-implementing the original RLM codebase in a more enterprise-ready format →goo.gle/4kjT12E

A few random notes from claude coding quite a bit last few weeks. Coding workflow. Given the latest lift in LLM coding capability, like many others I rapidly went from about 80% manual+autocomplete coding and 20% agents in November to 80% agent coding and 20% edits+touchups in December. i.e. I really am mostly programming in English now, a bit sheepishly telling the LLM what code to write... in words. It hurts the ego a bit but the power to operate over software in large "code actions" is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do. This is easily the biggest change to my basic coding workflow in ~2 decades of programming and it happened over the course of a few weeks. I'd expect something similar to be happening to well into double digit percent of engineers out there, while the awareness of it in the general population feels well into low single digit percent. IDEs/agent swarms/fallability. Both the "no need for IDE anymore" hype and the "agent swarm" hype is imo too much for right now. The models definitely still make mistakes and if you have any code you actually care about I would watch them like a hawk, in a nice large IDE on the side. The mistakes have changed a lot - they are not simple syntax errors anymore, they are subtle conceptual errors that a slightly sloppy, hasty junior dev might do. The most common category is that the models make wrong assumptions on your behalf and just run along with them without checking. They also don't manage their confusion, they don't seek clarifications, they don't surface inconsistencies, they don't present tradeoffs, they don't push back when they should, and they are still a little too sycophantic. Things get better in plan mode, but there is some need for a lightweight inline plan mode. They also really like to overcomplicate code and APIs, they bloat abstractions, they don't clean up dead code after themselves, etc. They will implement an inefficient, bloated, brittle construction over 1000 lines of code and it's up to you to be like "umm couldn't you just do this instead?" and they will be like "of course!" and immediately cut it down to 100 lines. They still sometimes change/remove comments and code they don't like or don't sufficiently understand as side effects, even if it is orthogonal to the task at hand. All of this happens despite a few simple attempts to fix it via instructions in CLAUDE . md. Despite all these issues, it is still a net huge improvement and it's very difficult to imagine going back to manual coding. TLDR everyone has their developing flow, my current is a small few CC sessions on the left in ghostty windows/tabs and an IDE on the right for viewing the code + manual edits. Tenacity. It's so interesting to watch an agent relentlessly work at something. They never get tired, they never get demoralized, they just keep going and trying things where a person would have given up long ago to fight another day. It's a "feel the AGI" moment to watch it struggle with something for a long time just to come out victorious 30 minutes later. You realize that stamina is a core bottleneck to work and that with LLMs in hand it has been dramatically increased. Speedups. It's not clear how to measure the "speedup" of LLM assistance. Certainly I feel net way faster at what I was going to do, but the main effect is that I do a lot more than I was going to do because 1) I can code up all kinds of things that just wouldn't have been worth coding before and 2) I can approach code that I couldn't work on before because of knowledge/skill issue. So certainly it's speedup, but it's possibly a lot more an expansion. Leverage. LLMs are exceptionally good at looping until they meet specific goals and this is where most of the "feel the AGI" magic is to be found. Don't tell it what to do, give it success criteria and watch it go. Get it to write tests first and then pass them. Put it in the loop with a browser MCP. Write the naive algorithm that is very likely correct first, then ask it to optimize it while preserving correctness. Change your approach from imperative to declarative to get the agents looping longer and gain leverage. Fun. I didn't anticipate that with agents programming feels *more* fun because a lot of the fill in the blanks drudgery is removed and what remains is the creative part. I also feel less blocked/stuck (which is not fun) and I experience a lot more courage because there's almost always a way to work hand in hand with it to make some positive progress. I have seen the opposite sentiment from other people too; LLM coding will split up engineers based on those who primarily liked coding and those who primarily liked building. Atrophy. I've already noticed that I am slowly starting to atrophy my ability to write code manually. Generation (writing code) and discrimination (reading code) are different capabilities in the brain. Largely due to all the little mostly syntactic details involved in programming, you can review code just fine even if you struggle to write it. Slopacolypse. I am bracing for 2026 as the year of the slopacolypse across all of github, substack, arxiv, X/instagram, and generally all digital media. We're also going to see a lot more AI hype productivity theater (is that even possible?), on the side of actual, real improvements. Questions. A few of the questions on my mind: - What happens to the "10X engineer" - the ratio of productivity between the mean and the max engineer? It's quite possible that this grows *a lot*. - Armed with LLMs, do generalists increasingly outperform specialists? LLMs are a lot better at fill in the blanks (the micro) than grand strategy (the macro). - What does LLM coding feel like in the future? Is it like playing StarCraft? Playing Factorio? Playing music? - How much of society is bottlenecked by digital knowledge work? TLDR Where does this leave us? LLM agent capabilities (Claude & Codex especially) have crossed some kind of threshold of coherence around December 2025 and caused a phase shift in software engineering and closely related. The intelligence part suddenly feels quite a bit ahead of all the rest of it - integrations (tools, knowledge), the necessity for new organizational workflows, processes, diffusion more generally. 2026 is going to be a high energy year as the industry metabolizes the new capability.

Full episode here: youtube.com/watch?v=v2l0HZ…

Your @github Copilot subscription can now officially be used with @opencode

x.com/i/article/2008…

x.com/i/article/2005…

x.com/i/article/2003…

🕰️📚 Event Deep Research Made by the LangChain Community Researches historical figures and generates JSON timelines. Built with LangGraph's supervisor pattern for multi-agent coordination and Studio visualization. Supports multiple LLM providers. Check out the repository: github.com/bernatsampera/…

Evaluation for deep agents looks different than evaluating single LLMs calls or chains As we’ve built out deep agents internally, we’ve gotten a sense of what this future looks like and written some initial thoughts on it

LangChain @LangChain

6 months ago

Evaluating Deep Agents: Here's what we learned Deep agents can't be evaluated like simple LLM tasks. After building and testing 4 production agents over the past few months, we learned that evaluating deep agents requires: 1. Bespoke test logic for each datapoint — each test

17 64 398 55K 407

In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from. In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit like what you'd see on Stack Overflow / Quora, or etc., but geared towards LLM use cases. Neither of the two above are going away (imo), but in this era of reinforcement learning, it is now environments. Unlike the above, they give the LLM an opportunity to actually interact - take actions, see outcomes, etc. This means you can hope to do a lot better than statistical expert imitation. And they can be used both for model training and evaluation. But just like before, the core problem now is needing a large, diverse, high quality set of environments, as exercises for the LLM to practice against. In some ways, I'm reminded of OpenAI's very first project (gym), which was exactly a framework hoping to build a large collection of environments in the same schema, but this was way before LLMs. So the environments were simple academic control tasks of the time, like cartpole, ATARI, etc. The @PrimeIntellect environments hub (and the `verifiers` repo on GitHub) builds the modernized version specifically targeting LLMs, and it's a great effort/idea. I pitched that someone build something like it earlier this year: x.com/karpathy/statu… Environments have the property that once the skeleton of the framework is in place, in principle the community / industry can parallelize across many different domains, which is exciting. Final thought - personally and long-term, I am bullish on environments and agentic interactions but I am bearish on reinforcement learning specifically. I think that reward functions are super sus, and I think humans don't use RL to learn (maybe they do for some motor tasks etc, but not intellectual problem solving tasks). Humans use different learning paradigms that are significantly more powerful and sample efficient and that haven't been properly invented and scaled yet, though early sketches and ideas exist (as just one example, the idea of "system prompt learning", moving the update to tokens/contexts not weights and optionally distilling to weights as a separate process a bit like sleep does).

Prime Intellect @PrimeIntellect

10 months ago

Introducing the Environments Hub RL environments are the key bottleneck to the next wave of AI progress, but big labs are locking them down We built a community platform for crowdsourcing open environments, so anyone can contribute to open-source AGI

128 403 3K 1.8M 2K

🧠🤖Deep Agents Simple tool calling loops fail on longer time horizon or more complex tasks Deep Agents like Deep Research, Claude Code & Manus succeed by using a number of tools and tricks We created a new Python package which makes it easy to build your own Deep Agents!

@jxnlco @ebadgio_ @ExtendHQ Extend has an amazing platform

Document automation is one of those deceptively complex problems that engineers often underestimate. We think "I'll just feed this PDF into GPT-4 and extract all the data!" only to discover a world of pain awaits. @ebadgio_ , CTO of @ExtendHQ , has seen it all when it comes to automating document workflows across industries, and his advice is refreshingly practical. Don't try to automate everything at once, invest heavily in domain-specific evals, and for heaven's sake, involve the people who actually understand the current process. This talk is packed with real-world examples that will save you months of frustration if you're tackling document automation. A s go the documents, so goes the business world. link below

Announcing AG-UI Launch Week 📆 7 days of BIG releases for the AG-UI protocol including partnerships with new frameworks, handshakes to other prominent protocols, new AG-UI clients, and more! Stay tuned 👀 Starting today with the AG-UI CLI 📟 ➡

Here are 5 practical tips for Context Engineering, which apply to @GoogleDeepMind Gemini 2.5 as well from @manusai! 1. Context Ordering Matters: Try to use "append-only" context, adding new information to the end. This maximizes cache hits reducing cost (4x) and latency. 2. Manage Tools Statically: Avoid changing tool order or availability mid-task, if not explicitly needed. This might break context caching and can/will confuse models if used tools in the history are no longer defined. 3. Use External Memory: Write explicitly or implicitly context/goals to external storage to. Preventing information loss. A typical task in Manus requires around 50 tool calls on average. 4. Recite Goals to not get lost: Prevent the model from "getting lost" by having it periodically restate its objectives. This keeps the primary goal in its recent attention span. 5. Embrace Errors: Keep error messages in the context. This allows the model to learn from its mistakes and avoid repeating them.

Our latest LangChain Academy course – Building Ambient Agents with LangGraph – is now available! Most agents today handle one request at a time through chat interfaces. But as models have improved, agents can now run in the background – and take on long-running, complex tasks. LangGraph is built for these “ambient agents,” with support for human-in-the-loop workflows and memory. LangGraph Platform provides the infrastructure to run these agents at scale, and LangSmith helps you observe, evaluate, and improve your agents. Together, they make it easier to build reliable, production-ready agents. In this course, you’ll: 📩 Learn the fundamentals of LangGraph and how to build your own email agent 📊 Evaluate your agent using LangSmith ✅ Add human-in-the-loop for reviews, and memory so your agent adapts over time 🚀 Deploy your agent with LangGraph Platform and connect it to Gmail

the new hot topic is "context engineering" we think LangGraph is really great for enabling completely custom context engineering - but we want to make it even better see our proposal (s/o @sydneyrunkle) for streamlining context management: github.com/langchain-ai/l…

🤖How and when to build multi-agent systems Two really good blog posts (from Anthropic & Cognition) came out last week on building (or NOT building) multi-agent systems ❌Cognition: Don’t Build Multi-Agents 📃Anthropic: How we built our multi-agent research system Despite seeming different on the surface, they actually have a lot of common guidelines and advice. I tried to summarize some of that (and relate it to LangChain efforts) Blog: blog.langchain.dev/how-and-when-t…

We’re working on making it easier for devs wanting to use Azure with LangChain or LangGraph! Doing some work internally and also in public here if you want to check it out ♥️ Feel free to tag me in a PR if you want to contribute too🤗 github.com/langchain-ai/l…

New post in the Vibe Enginering series: “Field notes from shipping real code with Claude” diwank.space/field-notes-fr…