Research & Insights141 articles

What we're thinking about

Technical writing, architecture decisions, and honest assessments of AI tooling from engineers who build with it daily.

Cursor vs Claude Code in April 2026: What Actually Changed

Our original Cursor vs Claude Code comparison got valid pushback — Cursor has shipped agent mode, subagents, a CLI, cloud agents, and terminal access since we wrote it. Here is the corrected picture as of April 2026, with what actually matters for choosing between them now.

Read articleApr 8, 20268 min read

141 articles

Compliance

AI in Legal Practice: Ethics, Regulation, and Reality

Law firms and legal departments are deploying AI at an accelerating pace despite an ethics framework built for a pre-AI world. Here is what is actually being used, what the ethics rules actually say, and where the malpractice exposure lives.

Apr 10, 202614 min read

AI News

Z.ai Just Open-Sourced an AI That Codes for 8 Hours Straight — Here’s What It Actually Means

Z.ai’s GLM-5.1 is a 754B open-source model that runs autonomously for 8 hours, tops SWE-Bench Pro at 58.4, and was trained entirely on Huawei chips. The agentic coding game just changed.

Apr 10, 20264

Engineering

How Database Connection Pooling Actually Works Under the Hood

Every serverless function, every microservice, every ORM opens database connections. Most engineers treat connection pooling as a checkbox — turn it on and move on. But the difference between a pool that quietly handles 10K concurrent users and one that collapses at 500 is entirely in the details nobody reads.

Apr 10, 202615

Opinion

AI Coding Tools Solved the Wrong Problem and the Industry Is About to Find Out

The entire AI coding tools industry optimized for code generation speed. Writing code was never the bottleneck. Understanding, maintaining, and debugging it was. We built a 10x accelerator for the part of software engineering that was already fast enough.

Apr 9, 202612

Engineering

12 Things to Check Before You Trust Your AI Coding Tools With Your Codebase

A fake Gemini npm package just got caught stealing tokens from Claude, Cursor, and other AI tools. Here are 12 things every engineer should check right now to make sure their AI development environment is not leaking credentials, context, or code.

Apr 9, 20267

AI & Engineering

Why RAG Still Outperforms Fine-Tuning for Enterprise Knowledge

Fine-tuning gets the marketing. RAG gets the production deployments. After two years of both approaches running in enterprise environments, the data is clear on when each wins — and when the comparison misses the point entirely.

Apr 9, 202613 min read

Tools

Replit Agent vs Lovable: After Using Both, Here's the Truth

Replit Agent and Lovable both promise to turn prompts into working apps. After building real projects on both platforms, here is what actually ships and what falls apart at the seams.

Apr 9, 20266

Tools

Cursor vs Claude Code in April 2026: What Actually Changed

Apr 8, 20268 min read

Engineering

How We Built CI Quality Gates for AI-Generated Code at Fordel (And What We’d Do Differently)

AI coding tools produce code at 10x speed. Your CI pipeline was built for human-speed PRs. Here is how we added quality gates that catch AI-generated bloat, hallucinated imports, and dead code — without blocking every PR or annoying your team.

Apr 8, 202612

Engineering

Observability-Driven Development

Observability is not monitoring with better branding. Monitoring tells you when something breaks. Observability lets you understand why it broke — and why it is slow, and why that one user's experience differs from everyone else's.

Apr 8, 20268 min read

Showing 10 of 141