Header card for migrating a .NET Framework 4.8 service to .NET 8 with a coding agent showing 312 files changed by the agent, 9 agent PRs, 4 rewritten by hand and 11 days wall clock

AI SDLC: Migrating a .NET Framework 4.8 Service to .NET 8 with a Coding Agent - What It Did, What We Did, and the Numbers

Splitting a .NET Framework 4.8 to .NET 8 migration between a coding agent and two engineers: the plan, the gates, two closed agent PRs, 11 days vs 7-9 weeks.

Oct 12, 2026 ai

Screenshot of the Lamplit Labs AI Pricing Calculator with GPT-4o selected, 1,000 input and 500 output tokens per request, and a cost breakdown of $0.0075 per request and $22.50 per month

AI Pricing Calculator: Estimating LLM API Costs Before You Ship

How the Lamplit Labs AI Pricing Calculator turns tokens per request and daily volume into a monthly LLM bill and compares models before you commit.

Oct 12, 2026 ai, lamplit-tools

Bar chart of added latency for a Go and a Rust AI inference gateway at 3,000 requests per second: p50 1.9 vs 1.1 ms, p99 9.4 vs 2.7 ms, p99.9 27.8 vs 4.6 ms, Go max GC pause 3.1 ms

Go vs Rust for a GC-Sensitive AI Inference Gateway: p99 Latency, Memory and GC Pauses, Measured

Same routing + micro-batching gateway in Go and Rust, 3,000 req/s on 4 vCPU. Added p99: Go 9.4 ms, Rust 2.7 ms; RSS 148 MB vs 34 MB; where Go still wins.

Oct 12, 2026 performance, languages, ai

Bar chart comparing Rust and Go on a CPU-bound microservice: Rust sustains 9,840 req/s with p99 under 15 ms on 4 vCPU, Go sustains 7,410 req/s

Rust vs Go for a CPU-Bound Microservice: Throughput, Tail Latency, Memory and Deploy Friction, Measured

Same CPU-bound pHash scorer in Rust (axum) and Go (net/http) on 4 vCPU. Rust: 9,840 req/s, p99 6.8 ms, 38 MB RSS. Go: 7,410 req/s, p99 12.7 ms, 112 MB.

Oct 12, 2026 performance, rust, go

Three-stage diagram of the automation evolution: 1 manual process with people and spreadsheets and a 4 percent known error rate, 2 scripted or API integration with a data contract, idempotent writes, logs and retries handling 70 percent of volume deterministically, 3 AI-assisted automation where the model handles the 30 percent tail behind a confidence gate and human review; a red box below shows the skipped path from manual straight to AI automates everything, where 3 of 3 pilots stalled before production

Enterprise AI: The Automation Evolution Enterprises Skip - Manual Process to API Integration to AI-Assisted Automation

Why AI-automates-everything pilots stall: they skip the scripted/API stage that fixes the data contract, baseline error rate and logging. 3 worked examples.

Oct 12, 2026 ai, enterprise-ai

Four-stage flow card: manual (human writes, checks, merges), scripts and API (tools/test.sh, html-proofer, GitHub Pages), AI agents (draft in throwaway clones, signed commits), and a human gate (review branch, owner-only files, Tier 2 stops)

Pulse and the Automation Evolution: Why We Did Not Skip Straight to 'AI Does Everything'

How Pulse fits the manual to API to AI progression: what it automates for this blog today, where humans still approve, and how to pace AI adoption.

Oct 12, 2026 ai, lamplit-tools

Horizontal bar chart of three-year run cost plus one-time exit cost in thousands of dollars: Azure OpenAI gpt-4o-mini pay-as-you-go 39k run plus 41k exit, Azure OpenAI gpt-4o PTU with one-year reservation 312k plus 118k, self-hosted Llama 3.1 70B on reserved A100 122k plus 76k, self-hosted three-node HA cluster 366k plus 76k

Enterprise AI: Vendor Lock-In and Exit Cost - Azure OpenAI Managed Endpoints vs a Self-Hosted Open-Weights Model

What it actually costs to leave: a lock-in audit and a priced exit plan for one 1.4 B token a month workload on Azure OpenAI vs self-hosted Llama 3.1 70B.

Oct 12, 2026 ai, enterprise-ai

Header card for gating AI coding-agent pull requests showing five gates every agent PR must pass, 11% blocked by the mutation gate and zero agent PRs merged without a human

AI SDLC: How We Evaluate and Gate AI Coding-Agent Pull Requests Before Merge

The five merge gates we run on every AI coding-agent pull request for .NET services: provenance, scope diff, sandboxed tests, mutation score and human review.

Oct 12, 2026 ai

Five-layer defence diagram for coding agents against prompt injection: untrusted inputs, trust boundary, least privilege, deterministic gates, human review, each assuming the previous layer was bypassed

AI SDLC: Prompt Injection Defences for Coding Agents That Read Untrusted Repository Content

How a coding agent gets hijacked by text in a README or PR comment, and the five layers that stop injected instructions from widening scope or leaking secrets.

Oct 11, 2026 ai

Bar chart comparing Rust 1.91 and Python 3.13 serving the same ONNX embedding endpoint with 32 connections: 1,840 vs 1,210 requests per second, p50 latency 11.4 vs 19.8 ms, p99 latency 38 vs 112 ms, resident memory 210 vs 640 MB

Rust vs Python for an AI Inference Microservice: 1.5x Throughput, 3x Less Memory, 3x Better p99

Rust 1.91 (axum + ort) vs Python 3.13 (FastAPI + onnxruntime) serving the same MiniLM embedding endpoint: req/s, p50/p99, memory and cold start measured.

Oct 11, 2026 languages, rust, python