alt Hacker NewsWithout experience, programming with AI (vibe coding, I guess) can be compared to being a rat in maze... You work your way through a project, but the dead-ends exact a high cost in terms time, attention, and ultimately cost.
With experience, you see these dead ends before they have a chance to take hold and you know when and how to adjust course. It's literally like one poster said: coding with some buddies without ego and without the need to constantly talk people out of using the latest and greatest shiny objects/tools/frameworks.
I've really enjoyed going back a revisiting old ideas and projects with the help of AI. As the OP stated -- it has restored my energy and drive.
Replies
I have always had ADHD and as a consequence have a decades long backlog of things that I want to do “some day”, and Claude just removes all the friction from going from idea to execution. I am also a software engineer, so basically for me it is like having a team of developers available 24 hours a day to build anything I want to design.
I have built and thrown away a half dozen projects ideas and gotten one into production at work in just the last few months.
I can build a POC for something in the time it would take me to explain to my coworkers what I even want. An MVP takes as long as what a POC used to take.
The thing that really unlocks stuff for me is how fast it is to make a cli/tui/web ui for things.
Claude Code has killed my ADHD and turned me into an always-on hyper-focused machine.
I am getting 20x done. This is a literal superpower.
I am not using it in agentic mode yet. I am telling it everything I want it to do. I will tell it where I want the files, what I want structs to be named, how I want the SQL queries to join, etc. I then review every line and make edits (typically with Claude first).
I haven't tried the agentic stuff yet, but I probably will at some point soon. I'm anxious about losing control over the architecture and data model, which is something I feel gives me my speed with Claude Code and that I know is important for my engineering work and quality.
I won't be writing code by hand ever again. This is the future. We'll look back at the old way as horse carriages.
Claude is also really freaking good at Rust, and the fact that it emits proper Rust with tests makes me even more confident of my changes.
We are literally living in the future now. Twenty years of SaaS and smartphone incrementalism and now we have jet packs.
Instead of engineers inventing 50 different frameworks and conventions for any given language or platform, maybe that energy will be directed to creating better AI tools.
Edit: I'll also reiterate what others are saying in that I think this is a tool best leveraged by engineers who know what they're doing and that care about code quality. The results you get back will also depend on your repo/project's code quality. If your project is poorly structured or has a lot of cruft, Claude will see that and spit it right back out. Keeping your code clean and low on tech debt is going to matter tremendously.
Fully agree: I believe my decades of software engineering experience definitely help me fly LLM tools better than less experienced folks.
But the much more interesting question to me: as LLM coding becomes the norm, does it drive the cost of self or small-company generated software to 0?
Like many SW architects/engineers my not-so-developed work-in-retirement plan is to assemble a small team of people I’ve loved working with over the years, start an LLC, and try to make a reasonable (not posh) living doing what we love: making software to solve problems.
On the one hand, it’s clear LLM coding can accelerate and amplify our efforts, but alternately there’s many people claiming there’s no possibility of a moat, your solution/innovation can be cloned in a matter of days … ie. the value of your software is exactly 0.
Not sure which future will be closer to reality. A backup plan that seems reasonable in the 0-value case is to focus our effort on creating actual physical gadgets and systems in the embedded realm, which conceivably can be designed and prototyped by a small team… It seems like these would still be valuable.