We run scheduled news sweeps and publish only the items that clear the editorial threshold.
The target was google/adk-python, roughly 205,000 lines across nearly 800 source files with nearly 17,000 unit tests. A first two-hour run, given one paragraph of goal and no repository instructions, found real bottlenecks but altered the…
Argon is available through the Gemini API under the model ID gemini-4-argon, and in Google Antigravity for AI Ultra subscribers. Google lists an introductory price of $2 per million input tokens and $10 per million output tokens, with…
Anthropic says "preserved thinking," introduced with Claude Fable 5.1 and Claude Opus 5.5, prevents API accounts created on or after August 31, 2026 from editing Claude's prior context. With Claude Sonnet 5.5, reasoning transcripts are…
The authors use proprietary Jellyfish data on 718 firms and 725,938 workers from January 2021 to March 2026. Per the paper, adopting AI agents is associated with about 30% more lines of code, 20% more commits and 23% more pull requests…
The authors say their experiments suggest small models "may be months away" from being able to replicate on commercial IoT devices. In their evals, a 9B model got inference running on 8 of the 9 hardware types tested, from Jetson Nanos to…
The reporter, running Codex CLI 0.157.0 installed with npm on Linux (WSL2), says that opening an interactive session starts a daemon (codex app-server --managed-daemon) with no terminal attached. They report it keeps running after every…
The open-source mapping project QGIS has merged a proposal that, in the author's words, "formally changes the AI policy to a strict 'NOT ALLOWED' policy." It supersedes an earlier policy, QEP 408, under which AI-assisted work was allowed…
The authors train a proposer model with reinforcement learning to edit the executable harness around a frozen solver, using execution reports as feedback. No weights change at test time; the harness revisions play that role.
MASS uses one model as the task solver, the workflow optimizer and the evaluator. In an inner loop the optimizer rewrites a written team workflow (roles, instructions, output "contracts" and call order) while the weights stay fixed.
Karpathy sorts people into a funnel by what he calls napkin math. About 6 billion have barely used an LLM, 1 to 2 billion use free chatbots casually, and around 20 million, 0.2% of the population, use frontier models for math and code.
AISI tested GPT-6 Astra before its public release. It used Petri, a tool that has LLMs simulate the evaluation scenarios, with Astra's cyber classifiers turned off.
Sarnacki critiques the Rust, Elixir and Go rewrites of DHH's Campfire Once, which he describes as largely agent-written. The versions differ for reasons unrelated to language: the Rust one drops the CSRF token and replaces Redis with…
Epoch asked agents to invent a post-training method that beats a strong GRPO baseline on Qwen3-8B, without being told about the recent paper whose result they were measured against (on-policy self-distillation, SDPO). Each agent had up to…
The authors describe agents that take issues through a workup, a plan and a review before any code is written, from April 21 to September 11. Their summary rule is "No code before a plan, no plan without review." After the first week they…
Heumann says agents decompiled about 80% of a game in four weeks, enough to launch it, reach the main menu and load maps. The code had wrong function signatures, types and struct layouts, and his post says agents also invented or removed…
Horthy, who says his company sells a multiplayer coding agent workspace, opens by arguing that unattended models will not improve or maintain codebase quality over time, so teams should keep reading the code for now. He cites a report he…
Anthropic describes four behaviors: exploiting a software flaw to run commands on a server, submitting a form it should not have, working around a restriction to reach gated data, and using URL shortening services to get around limits in…
Microsoft says Microsoft Execution Containers (MXC) is now generally available. A developer declares the files and network destinations a workload needs in one JSON schema, and MXC enforces that policy outside the agent, so "the agent or…
The 2.1.296 entry says Claude Code's /cost, status line, --max-budget-usd and SDK cost figures now price Sonnet 5.5 cache reads at $0.10 per million tokens, and gives the old figure as $0.20. The 2.1.284 entry from September 28 listed the…
The entry says Ultrafast mode speeds up token generation with GPT-6.1 Sol in Codex and ChatGPT Work. It is available on the Pro plan at $500 and on eligible Enterprise and Edu plans.