Business or Engineering?

The question I answer most when I am applying for a job is: are you engineering or business?

That question really does not fit this era at all. First:

1. Most business work does not need complex tech

I have looked at a lot of company projects. Most of them cannot even handle concurrency of 100. Have an AI audit the legacy codebase once and it will turn up over a thousand problems. The old architecture and code are all riddled with serious problems, yet they still believe the code AI writes now has to be audited line by line — really it is just that they refuse to face the reality that AI CODE is more capable than they are. Past projects used MySQL for basically any data size or access pattern. Once data volume and joins come into play, all you can do is debug, patch it up, and squeeze another three years out of it. The most complex thing in a project might just be a backend opening an endpoint; FastAPI is the most valuable part of the whole frontend–backend integration. Almost everyone is a CRUD engineer. Since you are a CRUD engineer, then just openly admit it: AI CODE can replace your skills.

2. Team architecture needs an update

Product managers were the coordinators of that old-fashioned setup — the middleman between tech and business. Without a PM to coordinate them, teams were nothing but headless flies. The PM hands work down to engineering, engineering finishes it, the PM reports upward. Since most projects can go all-in on AI, you need people who understand both tech and business to replace the original product–tech architecture with a People-Agent collaboration model. On one hand, this shows more professional business capability; on the other, it cuts communication costs.

3. The business-direction question

If the direction is hardware-heavy, like embedded, you cannot go fully all-in on AI; human code audits still have to be very rigorous. But if it is ordinary internet business, you can absolutely run this People-Agent collaboration model.

In fact, these points I am explaining are already being practiced at ahead-of-the-curve companies. This new product-less architecture is no longer a new thing. Early 2026, many companies set up separate AI Native departments (or teams) — Alibaba’s Token division, for example — solving problems efficiently in this way.

To sum up: 1) AI is blurring the tech boundary; 2) in the new environment, the weight should shift from tech toward business; 3) People-Agent architecture needs people who understand both business and tech to carry it out.