Claude Opus 4.5 changes everything

Source: Dev.to

Experimenting with AI‑Generated Code in 2025

Before I begin, I’d like to clarify my position. I’m one of those people who believe that AGI will happen. I don’t know when, but I think that, while the human mind and consciousness are extraordinarily complex, they are ultimately governed by the physical laws of the universe and can therefore be simulated. Someday AGI will be a reality. How far are we? I have no idea, and I don’t care.

What I do care about is how the current, rapidly growing capabilities of LLMs will impact the source of income for my family over the past 15 years. So I’ve been keeping an eye on emerging tools, workflows, and new approaches to software engineering.

For over two years I have been using GitHub Copilot extensively—with multiple models, coding agents, and custom agents. For the most part it has been hit‑or‑miss. I usually ask Copilot to implement a feature or fix a bug via the chat or coding agent, and many times the result is far from what I expected.

I have developed the habit of:

1. Staging changes before each prompt.
2. Code‑reviewing the changes for each prompt as I go along.
3. Rolling back via git when I’m not happy with the solution.

Working this way for a while has helped me develop a sense of what kinds of requests will work and how to break problems into steps that make it more likely the AI agent will do what I expect.

How AI Affects My Workflow

I know some developers feel like AI agents are taking the joy of coding away from them, but I do not feel that way because it has allowed me to spend more time developing features. I find that locating the root cause of bugs—​even when I end up fixing them manually—is something LLMs excel at, so I spend far less time debugging.

Another benefit, in my opinion, is that I have to deal less with repetitive tasks. After 15 years in the sector, I now enjoy spending more time understanding the business problem and designing a solution than actually implementing it.

• Once I have designed the API contracts, bounded contexts, database schema, etc., the implementation becomes very much grunt work—something I’ve done so many times that it is no longer enjoyable.
• In 2025 I spent much more time doing code reviews and writing technical specs (for the LLMs) and less on implementation.

Do I feel like I have become more productive in 2025? Not really—maybe a little bit, but not a lot. Overall, a task takes roughly the same amount of time, but I now spend more time thinking about the problem and doing verification rather than writing code.

A Holiday Experiment with Claude Opus 4.5

When the Christmas holidays arrived, I suddenly had unlimited Claude Opus 4.5 tokens for a couple of weeks. I decided to tackle a few old side projects that had been sitting idle for far too long.

Migrating a CRA Project to Vite

One of the projects was built with Create React App, which is now deprecated. I asked Claude to devise a migration plan from CRA to Vite. The plan was clear and thorough, so I gave the go‑ahead. Within three minutes I had a fully functional Vite setup.

Multi‑Platform Builds (Web, Mobile, Desktop)

I also wanted the app to run on multiple platforms:

• Web – via a browser
• Mobile – using Capacitor
• Desktop – using Electron

The challenge was handling platform‑specific storage APIs:

I asked Claude to:

1. Define a common storage interface.
2. Implement concrete classes for each platform.
3. Configure builds for Web, Windows, macOS, Linux, Android, and iOS.

In about 10–15 minutes I had a working version. After a couple of quick prompts I fixed a fullscreen issue in Electron, and everything ran as expected. The code quality was high enough that my review uncovered no major problems.

Rapid Feature Development

With the foundation in place, I started adding features. In a single morning I shipped more than 10 new capabilities. I even pushed Claude into unfamiliar territory: migrating an OpenGL application to WebGL, moving from Qt to TypeScript. The migration was completed in roughly half a morning.

Running several AI agents in parallel felt like “horizontally scaling” myself. While the agents produced code, I reviewed the output and planned the next set of tasks, dramatically boosting my productivity.

After the Tokens Ran Out

The holidays passed, and on 31 December at midnight my unlimited tokens came to an end. It is now January, and when I tried to code without them, the other models seemed dumb in comparison.

The joy of developing a product for me is not in coding the product—don’t get me wrong, I love coding—but the true joy comes from seeing people enjoying it. With Claude Opus 4.5, I can deliver more products and more features than ever before. I can focus on listening to my users and not have to suffer all the repetitive and tedious stuff.

Will the code be as good as NASA’s or as beautiful as poetry? No, but it will certainly be good enough to delight users, and for me, that is enough.

Looking Forward

I’m now looking at the higher‑tier Claude subscription and thinking that, given what I have experienced, it is actually reasonable. I think Claude Opus 4.5 changes everything, and it makes me both very excited and very anxious.

• Excited because development now feels like having a team of tireless assistants.
• Anxious because the landscape is shifting so quickly that I have to constantly adapt.

Developers will be able to build things that were not possible before because they required too much effort. This will be especially noticeable in open source. Soon we will see truly powerful open‑source solutions (that are more than just libraries) that can compete with big SaaS players.

• Anxious because it is hard to see how this won’t impact job security in the long run.

Note: What I’m describing above is not vibe coding. I still review every change, work in feature branches, run CI/CD pipelines, and understand the code that ships. This is agent‑driven software engineering, not blind prompt‑and‑pray development.

Have You Tried Opus?

What is the most impressive use case you have experienced?

---

In my next post I will talk about how I plan to combat my anxious thoughts about my career as a software engineer and focus my energy on the exciting possibilities.