We run scheduled news sweeps and publish only the items that clear the editorial threshold.
The piece opens by drawing a line: "An AI application needs a harness. An enterprise fleet of agents needs an operating system." It takes its definition of a harness from a talk it credits to Charles Holloway, "Build Your Own Harness",…
Larson gives the pattern a one-sentence definition: "The software factory pattern is looping on a broad goal, and then relying on the harness to drive progress towards that goal." His first implementation is an agent skill called…
jevchat generates text out of a typed decision model by asking it one question per symbol: given the question and the reply so far, which symbol comes next. The options are an alphabet plus a stop option, the model returns a probability for…
Patel's case is that MCP was built for models that could not yet write and run their own code, that most remote-service MCP servers "ultimately wrap APIs that already exist," and that models have now learned to discover a CLI by running…
Laya is the open implementation of the typed-decision primitive that TypeSafe ships as jev: a model that returns probabilities over a fixed set of questions rather than generating text. This port compiles it to Apple Core ML and runs it…
The author builds agents "that consumers are meant to use, to perform tasks on their behalf", which he separates from chatbots with subjective outputs. His argument is that prompt text is the wrong unit of work.
Kraishan followed 37,623 provenance-labelled pull requests from OpenAI Codex, Devin, GitHub Copilot, Cursor and Claude Code across 2,807 GitHub repositories between December 2024 and July 2025, alongside 4,027 human pull requests drawn from…
Khramtsov sets out seven defensive layers he has seen hold defect rates flat while output rises. Six of them are ordinary quality practice with agents attached: requirements review, unit tests at high coverage, manual testing, end-to-end…
Toub reports that roughly 430,000 lines of production TypeScript passed through the port, and that by August 21 the runtime stood at 832,378 lines of production Rust plus 468,689 lines of Rust unit tests. It was done in place on the main…
Spotify says it added two questions to its monthly major-incident retrospective during its AI ramp-up: whether AI-authored code contributed directly, and whether the increased volume of change pressured review, testing, rollout or…
The piece carries no personal byline. It describes one long-lived session that assigns work, re-runs every command an executing session claims to have run, and reads diffs rather than transcripts, while short-lived sessions do the…
Swerdlow built a version of Brood War playable only through an agent, then ran a round robin across model and effort configurations, 18 games per configuration, saving game engine data and both agents' harness logs for each match. Codex…
In auto mode a classifier checks actions such as shell commands and network requests before they run. Anthropic's documentation says that from 2.1.278 Claude Code asks the server to perform those checks as part of the session's own model…
ferstar says he found the archive while clearing disk space: a 313,070,842 byte encrypted file in ~/.zcode/v2/checkpoints, beside a state file recording a 345,549,173 byte workspace and 564 failed upload attempts. The packing manifest is…
Section 2.3 lists what a customer will not do. Clause (f) is that customers will not "publish benchmarks or performance information about the Services." Clause (c) separately bars reverse engineering or deriving the underlying ideas,…
Claude Code now reads AGENTS.md in a project that has no CLAUDE.md, changeable under Project instructions in /config, and not yet available on Bedrock, Vertex or Foundry. Claude Code uses AGENTS.md only when no CLAUDE.md exists, so a team…
Heumann drew the measurements from session logs for the month-long run he described in August. Across 47 subagent runs, he compared every read of a file at least 5kB in size.
Osmani walks through his own 82-minute build of a TMDB clone and tracks where the agent's output required human judgment before it could be trusted. He calls that time a "verification budget," borrowing the concept from web performance…
Martin Fowler's site, August 19, 2026. Laycock argues that production trust now limits AI-accelerated software development.
Anthropic built an internal index that rates every category of its AI research and development work on a six-level automation scale from Epoch AI, running from no AI involvement (AL0) to full autonomy with no human in the loop (AL5). As…