Three summary cards for the Jekyll build of this blog: 8.0 seconds as-is, 4.4 seconds without the last-modified hook, 3.3 seconds with the hook batched into a single git log call

Why This Jekyll Blog Took 8 Seconds to Build: 278 Git Subprocesses in One Plugin Hook

Profiling a 139-post Jekyll build: a last-modified plugin forks git twice per post and costs half of an 8 s build. One batched git log cuts it to 3.3 s.

Oct 14, 2026 performance, programming

Bar chart of added latency for a .NET 9 and a Node 22 LLM API gateway at 3,000 requests per second: p50 1.6 vs 2.4 ms, p99 6.8 vs 14.9 ms, p99.9 18.2 vs 46.3 ms

.NET 9 vs Node 22 as an LLM API Gateway: Requests per Second, p99 Latency and Memory, Measured

Same streaming LLM gateway in .NET 9 and Node 22 on 4 vCPU: .NET sustains 11,200 req/s vs 6,400, p99 6.8 vs 14.9 ms, RSS 212 vs 168 MB. Where Node still fits.

Oct 14, 2026 performance, languages, ai

Grouped bar chart comparing TypeScript, Go and Rust implementations of the same MCP server: cold start 410 vs 28 vs 19 ms, idle memory 92 vs 14 vs 6 MB, lines of code 540 vs 820 vs 1,130

TypeScript vs Go vs Rust for LLM Tooling: One MCP Server, Three Languages, Measured

The same MCP server and streaming CLI client in TypeScript 5.9, Go 1.25 and Rust 1.91: cold start, memory, tool-call latency and SDK maturity measured.

Oct 14, 2026 languages, typescript, go, rust

Four-step card for the AI SDLC with Copilot agents: spec issue with acceptance criteria, agent pull request with tests and log, review where the agent self-checks and the human sees three threads, merge once scope diff, tests and scan are green; footer notes 35 agent PRs over 6 weeks, 24 merged, 8 reworked, 3 closed

AI SDLC with Copilot Agents: From Spec Issue to Agent PR to Review, End to End

One .NET issue walked through the agent-driven SDLC: spec issue, Copilot coding-agent pull request, agent self-review and human merge, with screenshots.

Oct 14, 2026 ai

Four-step card: two Azure OpenAI deployments in two regions behind one APIM backend pool; circuit breaker trips on 429 and 5xx so the priority-2 region takes traffic; health probe closes the circuit and traffic returns; game day result of region pulled for 30 minutes with 0.3 percent errors and p95 up 410 ms

Enterprise AI: Multi-Region Failover for Azure OpenAI with APIM Backend Pools and Circuit Breakers

Fail Azure OpenAI traffic over between two regions with an APIM backend pool, circuit breaker and retries: the trace, the policy and game-day numbers.

Oct 13, 2026 ai, enterprise-ai

Header card with three numbered boxes joined by arrows: 1 Token Counter (170 tokens, 0.13% of context), 2 Schema Builder (create_support_ticket), 3 Validator (2 errors caught), under the title Prompt to Schema to Validator

One Workflow, Three Tools: Prompt to Schema to Validator with the Lamplit Labs Tools

The Lamplit Labs Token Counter, Schema Builder and Structured Output Validator run end to end on one support-ticket task, with a screenshot per step.

Oct 13, 2026 ai, lamplit-tools

Header card for AI SDLC practices listing scope contracts on every task, mutation score instead of line coverage, agent PRs earning a human review and a weekly 20-minute agent retro

AI SDLC: Seven Practices That Held Up After a Year of Coding Agents in .NET Teams

Twelve months of coding agents on nine .NET services: the seven AI SDLC practices we kept, the three we dropped, and the before-and-after numbers.

Oct 13, 2026 ai

Header card for the AI SDLC code-review agent workflow listing four stages: triage the diff, blocking versus advisory comments in one review, agent re-checks after the author pushes, human reviewer sees only unresolved threads

AI SDLC: A Code-Review Agent Workflow That Reviewers Actually Trust

The four-stage AI code-review agent workflow we run on .NET repos: triage, one review with blocking vs advisory comments, agent re-check, then human review.

Oct 13, 2026 ai

Lamplit Labs Tiktoken BPE Tokenization Playground header showing the GPT-4 model selector, the Code Snippet example loaded, and stat cards for tokens, characters, average characters per token, input and output cost, context used and tokens remaining

Tiktoken Playground: Seeing Where Your Prompt Actually Splits

How the Lamplit Labs Tiktoken Playground shows approximate BPE token boundaries, per-token byte counts, cost and context-window usage in the browser.

Oct 12, 2026 ai, lamplit-tools

Lamplit Labs Model Context Window Comparison header showing the provider filter, sort-by selector, model search box, provider chips and the top of the comparison table sorted by context window

Model Context Window Chart: How Much Fits Before You Start Trimming

How the Lamplit Labs Model Context Window Chart compares context window, max output and knowledge cutoff for 26 LLMs, and turns tokens into pages.