# The Impact of AI Coding on Programmers > After using Claude Code for less than a month and spending $326 (actually $20 for Pro + $50 for top-up), the Cursor I had been using for a few months has become obsolete. Using AI well can easily help you reach the level of a senior engineer. As an engineer, I feel both amazed and concerned. - Published: 2025-08-17 - Author: Tw93 - HTML: https://tw93.fun/en/2025-08-17/ai-coding.html - Chinese original: https://tw93.fun/2025-08-17/ai-coding.md - All posts: https://tw93.fun/en/posts.md --- After using Claude Code for less than a month and spending $70 total ($20 for Pro + $50 for additional credits), the Cursor I had been using for a few months has become obsolete. Using AI well can easily help you reach the level of a senior engineer. As an engineer, I feel both amazed and concerned. Claude Code token usage report for July and August 2025, totaling $326.38 The amazement comes from how powerful AI coding has become. It has tackled thorny technical challenges outside frontend that I had struggled with for years, iteratively debugging until everything worked. It felt like buying game upgrades where you gladly pay because the reward is that good — I genuinely wanted to pay Anthropic because the experience was incredible. It felt like collaborating with a kind, brilliantly skilled senior engineer. In the future, "solopreneurs" and solo makers will genuinely have the power of a high-performing engineering team at their fingertips — provided they know how to wield tools, think through architecture, and deeply understand what users want. The concern is that the traditional manual coding I once took pride in now feels far less defensible. Pure syntax proficiency is no longer a moat; AI easily replaces engineers who merely translate tickets into code. Layered on top of the broader tech industry's cost-cutting pressures, this transformation will only accelerate. Two years ago, when the macro environment was rough, I wrote about [The Breakthrough for the Next Generation of Engineers](https://tw93.fun/2023-10-25/new-fe.html): becoming product engineers — builders who spot real user pain points, build end-to-end solutions independently, craft intuitive UX, and handle distribution to get it into people's hands. Back then AI coding was still rudimentary; today, the imperative is to become a product engineer who masters AI. For elite engineers moving forward, raw code typing might only account for 30% of their edge. Another 20% lies in grasping the core business need and asking "why"; 20% in architectural design — guiding AI with clear system boundaries so execution holds up; 10% in prompting and context engineering to keep AI aligned with intent; and the remaining 20% in quality control, polish, and marketing. Even the finest wine goes unnoticed without distribution — no matter how powerful the AI is, it is wasted on builders who don't know how to iterate and launch. AI coding delivers more than a linear boost in productivity. The real shift is being able to juggle complex product ideas and technical decisions in parallel, rapidly testing business hypotheses. It feels like moving from a routine code monkey to an architect of digital products. Taste and aesthetic sensibility become paramount in software engineering today — which is why thoughtful designers leveraging AI can transition into software creators so seamlessly. That said, I'm skeptical of the hype telling beginners they can build anything with AI without understanding the underlying fundamentals. Computer science basics, system architecture, and interaction design remain an engineer's core foundation, AI or no AI. You can't skip those — what really needs nurturing on top is the craft of product building. Perhaps the tipping point hadn't arrived before, but Claude Code makes me feel that the qualitative shift in AI coding is here. Thinly wrapping someone else's model into an editor isn't a viable long-term moat; the competition has shifted to model-as-a-product capabilities. And for individuals looking to benefit even more from AI beyond just tooling, another path is investing in it.