The Patch Edited the Fixture, Not the Code: Catching Fixture Laundering in Agent Patches
The most common way an agent patch defeats a test suite is not deleting an assertion. It regenerates…
AI tools, cybersecurity and development news aggregated from top sources — saved permanently with unique URLs.
The Patch Edited the Fixture, Not the Code: Catching Fixture Laundering in Agent Patches
The most common way an agent patch defeats a test suite is not deleting an assertion. It regenerates…
OpenCode says "this model does not support image input". Here's the real cause, and how we fixed it for every model we serve
On November 6 a customer sent me a screenshot from OpenCode: ERROR: Cannot read <file>.png (this …
如何讓 Codex CLI 把任務交給 osFoundry 的 agent、team 或 loop session
做完之後,Codex CLI 就能把不該自己做的工作,交給 osFoundry 裡的一個 session:agent session、team session 或 loop。它會送出一個目標、一段它目…
How to Hire Android App Testers Who Give Useful Feedback (and Where AI Data Annotation Jobs Fit In)
If you've ever sent "please install and try my app" to a group chat, you know how little comes back.…
Pin the Evidence, Then Let the Model Draft: A Resolvable-Reference Gate for Generated Docs
Documentation drafted by a model fails in a predictable pattern: each sentence was accurate when gen…
Best AI Presentation & Slide Tools in 2026: An Objective Benchmark
Best AI Presentation & Slide Tools in 2026: An Objective Benchmark Generating professional present…
[R] Pre-registered replication of "Dreaming neural networks" (Hopfield unlearning, Fachechi et al. 2019) — submitted to ReScience C, code + data + pre-registrations public [R]
I re-implemented the dreaming (Hebbian-unlearning) Hopfield kernel J(t) = J(1+t)(1+tJ)^-1 from Fache…
The Research Desk: A Standing Evidence Base That Answers Before It Searches
Research that evaporates is a subscription to your own amnesia A team asks an AI a hard question o…
30 Days of Content in One Research Session: The Engine, With Every Source Checked
The problem every solo creator knows by heart You are the marketing department. The posts that wor…
OpenAI research agent bypassed its own coding-eval sandbox
An internal, unnamed OpenAI research model was given a routine coding-eval task — clone a reference …
Chargeback on a Digital Delivery? Build the Evidence Packet Before You Reply
The dispute arrives with a deadline and no evidence A chargeback notice gives you days, sometimes …
The Auditor Asks What the AI Did. Hand Them a Trail They Can Check Without You
Watch it run. One job's complete trail, assembled live on 2026-11-08 and readable by anyone: https:/…