Better Tools

For years, I’ve been comfortable with the traditional developer grind: Googling errors, digging through Stack Overflow, reading docs, trying, failing, and trying again. That process shaped how many of us learned to build.

When AI tools and “vibe coding” started trending, I remained skeptical. There was a trend of people using AI to blindly generate features, often leading to buggy, insecure products. I do use it to some extent, as an autocomplete on steroids or a quicker search engine, but I never trusted it to build entire features or products.

I thought I had the edge: a deeper understanding of my codebases, cleaner architecture, better control. But not quite. Technical depth is very important, but I underestimated the value of speed and productivity.

Compounding progress

Over time, the ecosystem has simply gotten better. Tools like MCPs and better AI agents are enabling actual engineers to build fast with enough supervision to ensure quality. These aren’t just prompt monkeys shipping broken products. These are thoughtful builders using better tools to augment their speed and scale.

What felt like hype can now translate into practical leverage: faster iteration with guardrails, less glue work, and a higher baseline of quality. We can still take the traditional route to build things — but others now deliver the same results much faster, with quality that’s often on par.

Better tools

This isn’t just about AI. We’ve seen similar resistance with developers rejecting tools like modern frameworks in favor of vanilla JS, ORMs in favor of raw SQL, or infrastructure tools like Terraform and Docker in favor of manually managing deployments via SSH and handwritten scripts. Not because they’ve weighed the tradeoffs, but because they believe the hard or traditional way is the right way. But convenience isn’t the enemy; complexity for its own sake is. At some point, the goal should also be leverage, not just mastery.

Better tools can lead to better products. They raise the floor on quality, cut cycle time, and reduce errors. You still need judgment, but you spend less time on glue work.

Used thoughtfully, good tooling doesn’t replace good engineering — it amplifies it. But without care and skill, even the best tools can’t save a sloppy product.

We still need to stay critical of hype and limitations, but that shouldn’t stop us from recognizing innovation when it’s real.

We need better tools. I’m learning to embrace them.