跪拜 Guibai
← All articles
Frontend · AI Programming · Claude

Four Structural Reasons Most Developers Won't Cash In on the AI Boom

By 乘风gg ·
Read original on juejin.cn ↗ Google Translate ↗ Alt translation

The piece names four specific structural filters — consumption vs. creation, domain expertise, organizational rank, and risk appetite — that explain why tool access alone doesn't redistribute opportunity. For a developer evaluating whether to stay employed or go independent, these are the exact gates that determine who captures the margin AI creates.

Summary

Using ChatGPT or generating images makes you a token consumer for someone else's platform, not a value creator. The people actually capturing AI upside are those who already understood a business domain and then applied AI to known pain points, not those who started with the technology and went looking for a problem. Inside companies, AI transformation budgets and decision rights sit with department heads and above; individual contributors are treated as replaceable agents, which is why the 'one-person company' model is gaining traction among senior engineers who see the ceiling. A fourth barrier is risk tolerance: chasing AI headlines daily isn't the same as leaving a stable salary to build something, the same way an earlier generation had to take on debt to open a factory.

Takeaways
Calling an API or prompting a model makes you a token consumer on someone else's meter; value capture belongs to the platform, not the user.
Profitable AI applications come from domain experts who retrofit AI onto existing business pain points, not from technologists who start with a model and hunt for a use case.
Inside enterprises, AI initiative budgets and decision rights stop at the team-lead level; individual contributors are seen as interchangeable agents, not stakeholders in the transformation.
Daily AI usage and trend-chasing is not the same as industry participation; capturing a wave requires the willingness to leave a stable salary, comparable to taking on debt to start a factory decades ago.
Conclusions

The 'one-person company' trend is framed here not as a productivity flex but as a rational response to the fact that employers already treat individual contributors as disposable agents — so the IC might as well capture the full margin directly.

The argument inverts the usual AI-upskilling narrative: domain expertise is treated as the scarce input, and AI fluency as the commodity. That ordering has real consequences for which career moves pay off.

The generational bookend — a future child claiming 'joining any AI company led to an IPO' — is a deliberate mirror of the opening complaint about missed real-estate and factory fortunes, suggesting the structural blindness recurs every cycle.

Concepts & terms
Business + AI vs. AI + Business
A prioritization framework: start with a known business problem or domain workflow, then apply AI to it (Business + AI), rather than beginning with a model's capabilities and searching for a commercial use case (AI + Business). The former tends to produce revenue; the latter tends to produce demos.
One-person company
A solo founder business model, increasingly viable with AI tooling, where one person handles product, engineering, marketing, and sales without employees. The term is used here to argue that if employers already view individual contributors as replaceable agents, the IC has little to lose by operating independently.
From the discussion

The conversation pushes back against the article's premise with three structural counterpoints: earnings are capped by personal understanding, survivorship bias inflates the visible success rate, and the cost of failure is prohibitive for those without capital. A parallel thread reframes the developer's role as a bricklayer in someone else's house — coding alone doesn't grant ownership. A more cynical take predicts a shift toward one-person companies fueled by social connections and baijiu diplomacy, where technical skills become irrelevant and EQ dominates.

People cannot earn money beyond their cognitive limits — even when told an opportunity exists, belief, know-how, and execution all remain barriers.
Survivorship bias makes a winning trend look universally achievable, while the many who fail are invisible.
The cost of trial and error is not uniform: a lottery ticket is trivial, but a 20-million-yuan loan is existential for a wage earner and pocket change for the wealthy.
Developers are like bricklayers — they build the house but rarely own it or capture the upside, which goes to those who contract or own the asset.
Working for someone else is framed as a temporary state, but age discrimination eventually closes that door.
The future belongs to one-person operations that win projects through relationships and offload execution to AI, making social skills and EQ far more valuable than technical ability.
Many who think they are leveraging AI are actually the product being sold to.
Featured comments
TokensSurging 6 likes

First, it's hard to earn money beyond your own understanding. For things we don't grasp, even if someone tells us doing this can make money, we might not believe it; even if we believe it, we might not know how; even if we know how, we might not do it well. Take livestream selling — is it that people don't want to? [facepalm] Second, you can't ignore survivorship bias. Even if the general direction is right and the trend is correct, it only means a higher probability of success, not that everyone will succeed. Who speaks up for those who failed 30 years ago? [shush] Third, you can't ignore the cost of trial and error. Everything involves both risk and opportunity. If the cost of trying is low enough, of course people will give it a shot — buying a 2-yuan lottery ticket, if you lose, you lose, anyone can try. But for a wage earner to take out a 20-million loan to start an AI company, how many dare to play that game? Yet if I had a few hundred million, what's wrong with gambling 20 million? Everyone's cost of trial and error is different, and the number of chances to try is different — you can't generalize. [lightbulb moment]

乘风gg  · 1 likes

Though it's hard, you've got to try, otherwise you'll never turn your life around.

永远的十七岁  → 乘风gg

Tried opening a brick-and-mortar shop, lacked understanding, ended up with 300k in loans. Learned my lesson. [tears]

鸡翅喝可乐变成可乐鸡翅 2 likes

It's actually easy to understand. Think of the projects we develop as a house (real estate). You're just writing code (a bricklayer). You could build your own house and become a landlord, or directly contract projects and profit — those are like the people who cash in on the boom. But generally, you're just a worker.

乘风gg  · 1 likes

So working for someone else is only temporary.

鸡翅喝可乐变成可乐鸡翅  → 乘风gg

After all, no company wants older workers.

阳火锅 2 likes

From now on, it's all one-person companies. Everything is about relationships and PR. Rely on connections to land projects, then hand them over to AI once you get them. In the future, China's baijiu culture and foot-massage culture economy will thrive even more. These are all lubricants for maintaining relationships. For the next generation, especially ordinary people, don't study technical stuff when choosing a college major — you can't beat AI. Everyone should study social skills. [look]

每天一道编程题

Indeed, in the AI era when technology is leveled, EQ becomes more important than IQ.

See top comments, translated →
Source: juejin.cn ↗ Google Translate ↗ Backup ↗