An agent takes twelve minutes to finish a task. The tool report shows twenty minutes of activity. One test command took three minutes, although the same command normally finishes in seconds. Before changing the model or the test runner, check…
Document-level accuracy hides the errors that cost money. Score fields, weight them by consequence, and gate every prompt change in CI. A team demos an agent that reads supplier quotes and customer RFQs and turns them into structured line items. The…
DeepSeek V4.1 Flash 1017 Nemotron 3.5 Super VL Hy4 is an open-weight model from Tencent, released November 4, 2026. It ranks #13 of 33 models on the Vals Index with 59.35%. Its best result is #15 of 81 on Tax Agent Bench. Release Date: Nov 4, 2026 59.35 % ± 1.08 Cost / Test (Vals Index) 51 min 31 s…
The inference from these sources is limited but practical: identity permissions, ticket-state consistency, a human owner, and an acknowledged handoff should be testable properties of the workflow. The sources do not prescribe the 16 cases below; those cases are an original…
Includes an EC2 liveness check solution Last month I lost half a day to a workflow that was working perfectly. It just was not running. An order-sync job on my self-hosted n8n had been switched off during an unrelated edit. No red…
I run a site of free calculators, and I wrote a gate that stops me from building a page that competes with one I already have. On its first real day it was wrong twice, once in each direction. Here…
I run a small site of calculators (pay raise, credit card payoff, overtime tax and so on). Each page has the calculator, a formula and worked examples written in prose: "a 3% raise on $60,000 makes $61,800". The usual way those drift apart…
Every AI-powered testing tool I've seen sends screenshots to the AI model. It works, but it's expensive: What if instead of a screenshot, you sent the AI a structured description of every interactive element — its position, label, type, and state?
Those workflows reduced manual review effort by about 60%. Most of our engineering time went into the cases where the model misread a document and what the system should do next. I didn't come to this through machine learning. My first…
I merged two PRs to fix my dependency alerts on GitHub. One of them (#89) had an empty diff. GitHub kept showing 11 open Dependabot alerts (10 high, 1 moderate). It was right, and I was wrong. When I dug into my terminal, I…
Unit, integration, E2E, contract tests, test data, coverage, and performance/chaos testing.
An article joins this feed when it matches these terms. Each one is also a search of its own.