A Better Workflow for AI-Assisted Coding: Lessons from Matt Pocock
I recently watched Matt Pocock's workshop, Full Walkthrough: Workflow for AI Coding . The main takeaway: Good AI-assisted coding is not about writin…
Testing & QA
⚑ Report a ProblemLatest Testing & QA news from Tech News
All topics
agents
ai
api
architecture
automation
backend
beginners
career
claude
cybersecurity
devchallenge
devops
discuss
javascript
llm
machinelearning
mcp
node
opensource
performance
productivity
programming
python
security
showdev
softwareengineering
testing
tutorial
typescript
webdev
Keep Prices Out of the Model: A Rate-Freshness Guard for Jewellery WhatsApp Assistants
A WhatsApp assistant for an Indian jewellery store will be asked about prices constantly. Today's rate, making charges, the cost of a design in a gi…
95% recall, 80% answers: what LongMemEval retrieval scores don't tell you
Memory layers for AI agents love to quote LongMemEval. You will see 94%, 95%, 96%. Those numbers are real, but they usually measure retrieval: was o…
You Cannot Unit Test a Summary. You Can Still Test It.
Every developer I know has the same reflex when a feature is done: write a test that feeds it an input and asserts the output. That reflex is correc…
Nobody will vouch for the answer
Microsoft says so in the overview page for its Agent Framework, the product meant to help you build on its models: "You are responsible for carefull…
Validate bounding boxes from a general vision model
A general model can take an image plus a question and return segmentation, bounding boxes, and object detection, work that used to require its own t…
I built the world's fastest AI gateway in Rust
TL;DR I built Antirion , an open-source (Apache-2.0) AI gateway written in Rust: one binary between your apps and LLM providers, with routing, fallb…
Ouroboros, очередной инструмент для вашего harness
Harness - построение инфры вокруг LLM таким образом, чтобы man in the loop (человек в процессе поддержки LLM) отсутствовал. Я решил поэкспериментиро…
Reflection's Beam is open: 23B active, but you still have to fit 501B
Reflection released the weights for Beam on October 16, 2026. It is the lab's first open-weight model: a sparse mixture-of-experts (MoE) model with…
Pick the LLM eval metric before you build the dashboard
You can ship a polished eval dashboard and still learn nothing about whether the product works. A common way to get there is picking a number becaus…
Claude Sonnet 5.5 and Haiku 5.5 in Python: Two SDKs, Prompt Caching and Haiku's 100K Pricing Cliff
Short answer: call claude-haiku-5-5 for classification, extraction, routing and summaries, and claude-sonnet-5-5 for multi-step coding and long tool…
Your retry logic has never seen a real 429. Here is how to test it with a local fault-injecting mock
Most LLM client code has a retry branch that has never actually run. The provider returns 200 in every test, so the 429 , 500 and timeout paths sit…
When to pay for a reasoning model on a request
A reasoning model spends extra tokens before it answers. Send every request there and you pay latency plus a larger bill for lookups a standard mode…
Why We Kept Ternary Bonsai 2 27B After Testing the Faster 9B
October 16, 2026 Our production classifier runs on a single RTX 2080 Ti with 11 GB of VRAM. It is often busy enough that requests spill over to a ho…
Почему ИИ ошибается и признает ошибку, только когда вы ее найдете? История одного экзамена
Не так давно я поступал в аспирантуру. Готовился по старинке: конспекты, книжки, пробные тесты. Но уже в самой аудитории заметил, как многие полезли…
The Parser Found a Verdict in the Model's Explanation
The Parser Found a Verdict in the Model's Explanation My review gate approved a diff the model had rejected. The reject verdict was sitting right th…
Can decision models like Jev and Laya route a coding agent?
We plugged Laya, a 421M-parameter decision model, into our coding agents as a model router. The plumbing worked in about 112 ms. Deciding when a big…
Reflection AI Beam: Why 501B Parameters That Cost Like 23B Matter More Than the Leaderboard
Reflection AI's Beam, announced October 5, 2026, is a 501-billion-parameter open-weight model that activates only 23 billion parameters per task. Th…
We ran the same LLM evaluation twice. Seven of eleven scores changed.
If you ship any automated score — evals, rankings, quality gates — this is the number you should know about your own pipeline. We measured ours and…
Jev 1.14: ищем объекты, проверяем документы и находим кошку среди 96 фотографий
В предыдущей статье я проверял Jev на текстовых задачах: классификации, уверенности ответов, длинном контексте и грязном входе. 14 октября TypeSafe…
Semantic Caching Explained: When It Saves Money and When It Fails
TL;DR Semantic caching evaluates the mathematical similarity of prompt embeddings against stored responses to bypass redundant upstream model calls.…
The Hidden Backdoor Risk in "Uncensored" AI Models
Overview Plenty of people spin up public models from Hugging Face on a GPU cloud or their own GPU server without a second thought. "Uncensored" mode…
I Burned 95% of My AI Coding Quota in 2 Days — Here's What the Logs Revealed
I use Claude Code every day to build a SaaS product. This week, the usage meter hit 95% before Wednesday. I assumed I had somehow written a mountain…
Audit an AI receptionist in 30 minutes: the 6-part checklist I run on every deployment
Somebody in your company set up an AI receptionist. It's answering calls, it sounds confident, and nobody has checked it since day one. That's the m…
Can AI Write Verilog for FPGAs
Can AI Actually Write Verilog for FPGAs? My Real-World Test As someone who does both full-stack web development and FPGA/Verilog programming, I was…
Coinbase's 9B Fraud Model Beat Opus 4.5: Copy the Pattern With a Verifiable Reward, Tier-Only Output, and Block-Only Harness
Attributed compile + one small offline run on synthetic data Primary sources (Coinbase Engineering, "Owning Intelligence" series): Part 1/3 — Build…
What does an agent actually need from your app to do one useful thing
We started with an assumption about what agents need from an app. Then we started asking developers. The answers keep surprising us. So here is the…
Первый RAG комом: как мы учились анализировать ошибки автотестов с помощью LLM
Привет! Меня зовут Алексей, я автотестер в hh.ru . У нас около 3000 запускаемых классов с e2e (UI) тестами на Java. Они прогоняются по несколько раз…
Why a '35GB' Quantized LLM Still OOMs an 80GB GPU: KV Cache, GQA vs. DeepSeek MLA, and the Real VRAM Math
Every engineer deploying open-weights models hits the same rite of passage: You take a 70B parameter model, quantize it to 4-bit ( 70B * 0.5 bytes ≈…
Three models vote on what happened
I ran the same job past three different AI providers and had them vote on the answer, and I did it for a hobby project about a Sunday night D&D…