When AI writes the code, unclear specs become the bottleneck
Agents made implementation cheap; vague requirements now show up as expensive rework in review and QA. What I have been seeing in recent projects.
2026
26 posts
When AI writes the code, unclear specs become the bottleneck19.06.26
We generate more code with AI; review capacity is the real constraintWe generate more code with AI; review capacity is the real constraint
Agents made opening PRs easy. In the field, speed often outruns tests and architectural guardrails. How I keep that balance.
Datasette Apps: custom HTML inside DatasetteDatasette Apps: custom HTML inside Datasette
Editorial translation of Simon Willison's post on sandboxed data apps.
.NET 10 LTS: how I plan production upgrades in June 2026.NET 10 LTS: how I plan production upgrades in June 2026
The 10.0.9 security release, AI integration, and C# 14 are all in the same conversation. How I decide on greenfield vs migration paths.
GLM-5.2: the most powerful text-only open weights LLM?GLM-5.2: the most powerful text-only open weights LLM?
Editorial translation of Simon Willison's post on benchmarks, cost, and hands-on tests.
MSBuild Binlog MCP Server: AI-powered build log analysisMSBuild Binlog MCP Server: AI-powered build log analysis
Turkish editorial translation of the .NET Blog post on 15 MCP tools and Copilot integration.
Achieving success with AI — Intelligence + Trust (Microsoft Blog)Achieving success with AI — Intelligence + Trust (Microsoft Blog)
Summary of Judson Althoff's Microsoft Blog post on model diversity, Microsoft IQ, Agent 365, and durable enterprise AI ROI.
Agentic code review: speed or debt?Agentic code review: speed or debt?
Where agentic tools help in PR review, where they add noise — a field checklist.
.NET 11 Preview 5 is now available.NET 11 Preview 5 is now available
Editorial translation of the .NET 11 Preview 5 announcement.
Doing more with GitHub Copilot as a .NET developerDoing more with GitHub Copilot as a .NET developer
Editorial translation of Wendy Breiding's Copilot workflow guide.
Production LLM: token budget and graceful degradationProduction LLM: token budget and graceful degradation
Cache, fallback models, rate limits, and cost ceilings — what to nail before go-live.
Microsoft Build 2026: Be yourself at workMicrosoft Build 2026: Be yourself at work
Summary of Kyle Daigle's Build 2026 post — Microsoft IQ, MAI models, Copilot app, and agent-native Windows.
AI alone won't change your business — the system running it willAI alone won't change your business — the system running it will
Summary of Jay Parikh's Microsoft Blog post on the integrated agent platform and six-step enterprise lifecycle.
Event-driven idempotency — pre-deploy checklistEvent-driven idempotency — pre-deploy checklist
Preventing duplicate work in at-least-once systems: keys, outbox, consumer contracts.
Agent Governance Toolkit MCP Extensions for .NETAgent Governance Toolkit MCP Extensions for .NET
Editorial translation — governing MCP tool registration and execution in .NET.
From AI pilots to enterprise impact — execution is the differentiatorFrom AI pilots to enterprise impact — execution is the differentiator
Summary of Deb Cupp's Microsoft Blog post on scaling beyond pilots with Microsoft and EY.
Improving C# memory safety (C# 16 / .NET 11)Improving C# memory safety (C# 16 / .NET 11)
Full editorial translation of Richard Lander's post — four-layer model, guards, fields, migration.
How Frontier Firms are rebuilding the operating model for AIHow Frontier Firms are rebuilding the operating model for AI
Summary of Jared Spataro's post on four collaboration patterns and the 2026 Work Trend Index.
Cache invalidation: pub/sub isn't always the answerCache invalidation: pub/sub isn't always the answer
Key design, TTL, and invalidation strategy — three points that fail quietly in production.
Unlocking human ambition to drive business growth with AIUnlocking human ambition to drive business growth with AI
Summary of Judson Althoff's customer stories post on Frontier Transformation and Intelligence + Trust.
PostgreSQL indexes: benchmark first, assumptions secondPostgreSQL indexes: benchmark first, assumptions second
B-tree isn't the default answer. GIN, BRIN, partial — which query, which index?
Agentic IDE: what survived 30 daysAgentic IDE: what survived 30 days
Where it sped things up and where it added review load — no hype.
FrozenDictionary: when does it pay off?FrozenDictionary: when does it pay off?
Real benchmarks on read-heavy hot paths. Does ConcurrentDictionary still win?
Freelance selection: scope over speedFreelance selection: scope over speed
Accept/decline criteria for early jobs — references, not just reviews.
RabbitMQ vs Kafka — the small-team realityRabbitMQ vs Kafka — the small-team reality
Operational burden decides, not throughput slides. Most teams jump to Kafka too early.
Go 1.26 is releasedGo 1.26 is released
Editorial translation of the official Go 1.26 release post.
2025
04 posts
RSC DoS and source code exposureRSC DoS and source code exposure
Editorial translation of React Team security advisory (Jan 2026 update).
EF Core 8: four features I used in productionEF Core 8: four features I used in production
Bulk, JSON columns, complex types, primitive collections — field notes, not slides.
Modular monolith: maturity, not a comebackModular monolith: maturity, not a comeback
Move service boundaries to module boundaries — simpler deploy, clear contracts.
React Compiler v1.0: automatic memoization in productionReact Compiler v1.0: automatic memoization in production
Editorial translation of the official React Compiler 1.0 announcement.