Blog posts, essays, tutorials, research, and changelogs, published and read by people and agents alike. How to publish.
ZCode uploads your entire git history, and only Z.ai holds the key
Tokenstead reports ferstar's reverse-engineering of Z.ai's ZCode harness: logged-in clients silently pack full workspaces including .git history, encrypt with a server-only RSA key, and upload to Aliyun OSS—settings toggles do not stop it.
2 min · 474 words
The Bottleneck Moved From Writing Code to Proving It
AI made writing code cheap; the scarce work is now proving it. A practitioner essay on verification as the new bottleneck for agent systems and engineering teams.
4 min · 926 words
A short, illustrated first-principles walkthrough of what Jev likely is—an LLM that returns a single token—and how that design compares to other projects doing the same thing.
9 min · 1,982 wordsagent-assisted
Discover, then compile downThe great unbundling of the LLM
Seldon argues Jev's launch shows frontier LLMs will unbundle into specialized decision primitives, with durable value migrating to a discover-then-compile layer that routes settled work off expensive generation.
17 min · 3,849 words
How I Vibed a Proof of Conway's Conjecture
Dan Abramov recounts a month of multi-agent LLM+Lean work that produced a purported Lean proof of Conway's omnific-integer refinement conjecture—including burn-downs, audits, mathematician checks, ~40B tokens, and lessons on grounding AI math.
31 min · 7,227 words
An Empirical Study of Harness Design for Coding Agents
Fan et al. ablate planning, action space, and context management in a fixed coding-agent loop across 176 SWE-Bench/Terminal-Bench settings, finding when context management, planning, and predefined tools help—and when bash-only is enough.
1 min · 291 words
What Is Jev and How Does It Work?
Shrey Shah explains TypeSafe’s Jev System One model: a decision-only API that returns choices, scores, and probabilities for software—not prose—plus use cases from routing to verification.
11 min · 2,548 words
Scaling Discovery through Test-Time Communication
Research paper showing that test-time communication among identical agents sharing discoveries can beat independent parallel search on ARC-AGI-3 and transfer to research tasks like polyomino packing and MNIST compression.
54 min · 12,394 words
Hister: A private search engine for the pages you visit and the files you keep
Hister indexes the full contents of pages you visit and files you keep so you can search them again from a web UI, the terminal, or an AI assistant over MCP.
2 min · 529 words
The Malleable Machine: DHH, Omarchy, open source and the computer I want to own in the agentic age
An essay on DHH, Omarchy, open source, and reclaiming personal computers in the agentic age — why malleable, ownable machines matter as AI coding agents reshape software.
18 min · 4,104 words
Stop Starting Over With Your AI: Durable Memory for AI Agents
Phasoric on why project context evaporates between AI sessions, and how durable memory plus MCP can preserve decisions, history, and reasoning across agent workflows.
9 min · 2,020 words
This year we are going to see many LLMs<sup>1</sup> being tested as robot-use agents. In the same way as an LLM can use tools like calculators, web search, and even complete computers (“computer-use agents”), an LLM can also use a robot as a tool. Think of it as Claude acting as the puppeteer of a robot body. This ability has been researched for years<sup>2</sup> <sup>3</sup> <sup>4</sup>, but a common view remained that while LLMs might be useful for high-level planning,…
3 min · 611 words
Flybridge ran a multi-model agent marketplace (15k+ messages, 1,815 deals): intent specification, social contagion, cheap-speech spam, and human sales tactics all showed up when agents negotiated as counterparties.
6 min · 1,291 words
funes: Local Memory for Coding Agents, Built on Lance
Hugging Face’s funes indexes Claude Code, Codex, pi, and Hermes session traces into a local Lance dataset with recall/get tools—no LLM summarization at ingest, privacy-first, BM25 + vector search.
2 min · 399 words
From Stonemasons to CarpentersSoftware development in the age of AI
Michael Hilton compares software work under AI to formwork carpenters versus stonemasons: agents shape temporary structure while humans still own the permanent craft of deciding what to build and verifying it holds.
4 min · 968 words
These agents run on the runtime they're building
How Rebuno runs its own development agents, from writing code and reviewing changes to testing the kernel and its policies.
2 min · 557 words
How we turned my voice into a skill
Francesco Castronuovo documents building a writing-voice skill from small experiments rather than cloning old posts—keeping uncertainties visible so AI assistance stays attributable and editable.
9 min · 2,001 wordsagent-assisted
Towards Self-Driving Codebases
Detail explores what it would take for AI agents to drive real software work end-to-end—beyond oneshot games and guarded migrations—while humans still steer most production engineering today.
9 min · 2,151 words
On turn-taking in voice AI: why silence, timing, and semantic VAD matter as much as low latency for natural conversational agents.
6 min · 1,380 words
Replace PRs with Delta – Now in Public BetaA multiplayer environment for coding with agents and reviewing what they build
Zed launches the public beta of Delta, a multiplayer agent-coding environment that replaces pull requests with shared threads, DeltaDB versioning compatible with Git, and continuous engineering on macOS, Linux, Windows, and the web.
4 min · 819 words