We run scheduled news sweeps and publish only the items that clear the editorial threshold.
GPU Mode ran a contest to implement batched compact-Householder QR factorization on B200 hardware. sankalp placed 12th of 183, taking the reference torch.geqrf path from roughly 419,000 microseconds to a final tracked 1,805, and made over…
Voting runs from 00:00 UTC today to 23:59 UTC on August 28, after a discussion period that opened July 23 and was extended. Eight options are on the ballot, drafted independently by eight different proposers, which is why they do not read…
Chakrabarti tracked individual instructions, not file sizes, across 1,867 GitHub repositories and 247,694 instruction lifetimes. At its last tracked version the median agentic context file carries 39 instructions, with a 90th percentile…
The experiment holds what an agent can do roughly fixed and varies only how those capabilities are organized and exposed. Six tool setups, three actor models (Qwen3Coder-30B, Kimi K2.5 and Claude Sonnet 4.5), 65 issue-resolution instances…
A first-party account of what a Claude Code turn is billed for and what breaks the prompt cache. The figures are Anthropic's own: output tokens are priced "at roughly 5x input", reading from the cache costs 0.1x the input price, and writing…
Two releases landed after this feed last covered the changelog at 2.1.224. In 2.1.232, "subagent forking is now on by default", and a forked subagent "inherits the full conversation and prompt cache".
This is the follow-up to Harness engineering for coding agent users, the article that supplies the guides-and-sensors vocabulary, and it reports what happened when Böckeler removed one half of that pair. She rebuilt an internal analytics…
DeepSeek promoted V4-Pro to general availability and, in the same post, announced that API billing moves to peak and off-peak rates at 16:00 UTC on August 16. Off-peak is exactly half of peak.
A first-party account of when the Pi coding agent compacts and what it sends when it does. Pi checks for auto-compaction after a turn ends, against a configurable token budget whose current default is 20 thousand tokens, "roughly 5 to 20…
Cursor now prepares "ready-to-use copies of your development environment" in the background, by default once an hour, so a cloud agent forks a warm machine instead of booting, cloning repositories and running the install script at session…
A short post, roughly 450 words, holding that Opus 5 is the more capable model on benchmarks and the worse one to work with, because Opus 4.7, Opus 4.8 and Fable "stop and ask questions if my intent was unclear" and "don't reinterpret or…
DeepSeek released dsh, its own coding agent harness, as a developer preview with the source under an MIT license. The design claim is that every capability is a swappable plugin: "models, tools, skills, sessions, sandboxes, storage,…
Delta is a separate application from Zed, in private beta as of yesterday, built on DeltaDB, the version control layer Zed has been developing publicly. The argument is about where the record of agent work lives.
Not new, published April 2, 2026, and still the clearest map of the vocabulary the day's release is speaking. It splits an agent harness into guides that steer before the agent acts and sensors that observe after, each either…
The Economist piece is paywalled and was not read for this edition, so nothing is reported here about what it argues. The discussion around it is the notable part: roughly 150 points and 188 comments by early evening Eastern on August 13,…
Datadog scaled Cursor and Claude Code to its roughly 4,000-engineer org in under a year, moving from a 100-to-200-person pilot to more than a thousand daily users within a month of launch. Rather than write more rules, Boudrias's team…
Martin's talk lays out the architecture behind Anthropic's newer managed-agents surface for long-running work. The harness running the agent loop is decoupled from the containers executing its work and connected through an append-only…
Osmani's argument is that code review does not scale to agent-generated volume, so quality has to move into the constraints placed around the agent: tests, mutation testing, complexity and line-length limits, and whatever else gates a…
New Claude Code sessions on Pro, Max, and Team plans switch to auto mode by default on August 14, replacing per-command permission prompts with a classifier that blocks actions judged irreversible, destructive, or aimed outside the user's…
Herrengt argues that AI didn't make bad engineering decisions worse, it just removed the limit on how fast a team can make them: a codebase can now absorb more change over one weekend than it used to accumulate while its senior engineer was…