Nice writeup here about how coding agents can score similarly on benchmarks but operate very differently behind the scenes blog.jetbrains.com/junie/2026/0...

With Codex/Claude the bottleneck in software development is now expressing your intent. Writing the code is no longer the rate limiting step. This completely upends software development. Take issue trackers. The key problem they were solving was allocation of scarce coding resources to features.

An interesting paper with implications for skill extraction, continual learning, and token efficiency - it provides even more incentive & value from running user-specific task evals because you can extract skills from them. arxiv.org/abs/2608.07885

Reason Wide, Not Deep: Amortizing the Reasoning Premium into Distilled Skills

Reasoning modes of language models outperform their non-reasoning counterparts on multi-step agentic tasks, but pay a 3-6x premium in output tokens on every episode -- much of it spent re-deriving pro...

arxiv.org

This week's #PaperILike is "People construct simplified mental representations to plan" (Ho et al., Nature 2022). People generate ("construe") simplified task representations on-the-fly. (Robots should too!) Also: awesome first paragraph. PDF: arxiv.org/abs/2105.06948

People construct simplified mental representations to plan

One of the most striking features of human cognition is the capacity to plan. Two aspects of human planning stand out: its efficiency and flexibility. Efficiency is especially impressive because plans...

arxiv.org

After The Rain - Overmono

really enjoying this setup for syntopic reading with Fable

My transparent Creator Micro keys arrived, so now I can make atelier.dev decision points map to physical buttons

"the phoenix must be reborn with its empty spaces intact — intent includes the intent to remain open, and “immortality of intent” is only half-achieved by a system that can’t carry a dodge through the fire." wtf you talking about fable? 😂

Just realized that my portable table lamp would make the perfect vessel for a voice-powered AI assistant

What's new isn't the loop. It's that the controller is now expressive enough to handle unstructured goals: natural language specs, ambiguous tasks, open-ended iteration. That's the actual innovation. Not the feedback loop itself.

How to create reading experiences that "go beyond information transmission and toward reader transformation.” Great paper by @blue-phia.bsky.social @lepidopterane.bsky.social @yijunliu.bsky.social Sarah Sterman @sh1m.bsky.social @maxkreminski.bsky.social arxiv.org/abs/2606.04308 >

Every time an agent explores your code base it builds up its own mental model of how the code works, and then throws that model away when the session ends. Code walkthroughs turn that model it into a step-by-step guide to bring you up to speed on any part of the system.

In 1985 Peter Naur argued that a program is more than just its source code. "Programming As Theory Building" explained how we build theories of the code which help us debug and refactor it but those theories rely on knowledge from outside of the code. pages.cs.wisc.edu/~remzi/Naur....

finally a dedicated device for the 20X per day you need to open MS authenticator 😄

What are users thinking during their interactions with LLMs? Introducing ThoughtTrace — the first dataset capturing what users think during real-world human-AI conversations. These thoughts improve user behavior prediction and model alignment, opening a new paradigm of user-centric LLM research.

Most engineers stopped reading docs ages ago. But now, agents are reading them. What if you had a way to measure freshness on every PR? Is it as easy as implementing three signals, plus a Claude Code layer? Let's find out together!

How Fresh Are Your Docs? Score Documentation Freshness in CI

Learn how to score documentation freshness in CI pipelines to keep your docs accurate and up to date. Automate quality checks and catch outdated content before it reaches users.

go.dosu.dev

I sketched this out a few years ago. The HTML vs Markdown debate is conflating substrate with information density. The real question is what kind of feedback an artifact actually invites. Hi-fi invites parameter critique. Lo-fi invites paradigm critique.

One of the most famous power-user tools in the world is switching to an AI chat interface 😬 "This will be the new Terminal. This will be the primary way most interactions are happening..." www.wired.com/story/the-bl...

The Bloomberg Terminal Is Getting an AI Makeover, Like It or Not

WIRED spoke with Bloomberg’s chief technology officer about the big, chatbot-style changes coming to the iconic platform for traders.

wired.com

For the past few years, humans have been doing “prompt engineering” to coax the best performance out of different LLMs. In this work, we explored what happens if we train an AI to do that job instead. Link to our #ICLR2026 paper: arxiv.org/abs/2512.04388 Thread:

Sakana AI@sakanaai.bsky.social · 4mo ago

Introducing our new work: “Learning to Orchestrate Agents in Natural Language with the Conductor” accepted at #ICLR2026 arxiv.org/abs/2512.04388 What if we trained an AI not to solve problems directly, but to act as a manager that delegates tasks to a diverse team of other AIs? Thread:

A graph you can act on, not just stare at 🌴 @obsidian.md

Still great advice 43 years later

What if Anthropic is intentionally making Claude sound like the most annoying TED talk presenter ever so that we know it's not a real person 🤔

Got to talk at @aidotengineer.bsky.social conf last week about the need for collaborative AI engineering. All our current coding agents are single player. We're trying to scale up individual productivity, but creating tons of alignment problems in the process. We have no good tools for...

Every year, I think, “This must be the most interesting year to work on AI coding agents,” and every year, I am correct.

AI can create code fast. But it does not mean you cannot and should not understand it at the speed at which it creates it.

This is how I use AI to augment my design process. Instead of having the agent show me a giant diff of changes, Atelier links each edit back to the original thread of feedback so I can review them in context

There comes a point in every man’s life where he releases an open source frontend framework. That day, for me, is today. Tutuca: a zero-dependency, batteries-included SPA framework marianoguerra.github.io/tutuca/ Would love your feedback on the framework, the landing page and the tutorial

The CLI is great for a lot of things but reviewing Claude Code plans is not one of them. So much scrolling back and forward just to figure out what the agent changed

I just released v0.3 of atelier.dev with the new inline revisions feature available for everyone to try

Shawn Simister@narphorium.com · 5mo ago

Revising a document with an agent should feel like collaborating with someone sitting next to you reading the same page. You highlight what needs to change, the agent suggests revisions, and you decide. That's what I'm building in Atelier

Revising a document with an agent should feel like collaborating with someone sitting next to you reading the same page. You highlight what needs to change, the agent suggests revisions, and you decide. That's what I'm building in Atelier