As "AI programming" becomes increasingly easy, a new wave is emerging: "micro-applications" that anyone can create and "super programmers" who can outperform entire teams.

Silicon Valley is frantically vying for a new breed known as “Cracked Engineers.”

Under the AI gold rush in the Bay Area, traditional hiring logic is failing. J.X. Mo, co-founder of the robotics startup Gradient, directly cancelled the company’s internship recruitment program. The reason is simple and brutal—no one is “Cracked” enough.

In his view, in an era of AI-assisted programming, if a newcomer is not “Cracked”—cannot perform with explosive skill and technique like a top gamer and is tirelessly productive—then they do not deserve an employee badge.

Today, as AI dramatically lowers the threshold of programming, the global software industry is experiencing unprecedented polarization.

On one hand, the bar for programming is dropping infinitely. Thanks to tools such as Claude and ChatGPT, ordinary people who don’t know how to code are beginning to build highly personalized “micro apps” using natural language; software is shifting from something you purchase as a product to something you can DIY as a tool.

On the other hand, competition among professional engineers is intensifying. In Silicon Valley, startups are fighting over the aforementioned “Cracked Engineers” (hard-core engineers). These young individuals leverage AI to magnify their personal output to the extreme, trying to replace entire traditional development teams with their own abilities.

Micro Apps for Everyone: From “SaaS Subscriptions” to “DIY Tools”

For a long time, if you needed software to solve a specific problem, you would usually go to the App Store to download or purchase SaaS services. But now, a new consumption habit is emerging: make it yourself, use it, and move on.

This type of software, built by non-professionals, is called “Micro Apps” or “Fleeting Apps.” Their features are distinctive: the scenarios they address are highly vertical, solve instant pain-points, and often lack any commercial promotion intent.

Howard University Computer Science Professor Legand L. Burge III compares them to the flash trends in social media, except this time the protagonist is software itself—“When the need goes away, the software disappears too.”

Rebecca Yu’s story is highly representative.

To solve her friends’ “decision paralysis” at group dinners, she—without any technical background—used Claude and ChatGPT and created a web app called Where2Eat in just seven days. The app can recommend restaurants based on her and her friends’ common interests.

In the past, this would typically require hiring a professional full-stack engineer or using complex low-code platforms.

Rebecca Yu says:

“Once I learned how to efficiently prompt the AI and troubleshoot, the building process became much easier.”

(1) Filling the Gap Between Excel and SaaS

Christina Melas-Kyriazi, Partner at Bain Capital Ventures, sharply notes that micro apps are filling the huge market vacuum between ‘Excel spreadsheets’ and ‘full-featured SaaS products.’ As Shopify made opening online stores easy, AI is making software development as casual as making a spreadsheet.

People have begun customizing software for ultra-vertical—even “trivial”—needs, and similar cases are erupting in Silicon Valley:

• Medical Records: Software engineer James Waugh developed a simple log app for a friend with heart palpitations, specifically for displaying heart data to a doctor.
• Daily Trivialities: Media strategist Hollie Krause, who disliked the doctor-recommended apps, made her own allergy tracker. She described the speed: “By the time my husband returned with dinner, I’d finished writing it.”
• One-time Entertainment: Founder Jordi Amat created a web game for a family holiday gathering, and once the holiday ended, the app’s mission was done and it was shut down.
• Bad Habit Tracker: An artist developed a “bad habit tracker” to record how many hookahs were smoked or drinks were consumed over the weekend.

(2) The Business Paradox of “Use and Go”

Darrell Etherington, VP of SBS Comms, predicts that people will stop subscribing to monthly-fee utility apps, and instead, based on specific needs, “self-supply” using Claude Code, Replit, or Bolt.

However, the trend faces resistance. Although the threshold for web app development has dropped to zero, mobile app development still faces the “Apple tax”—the annual $99 developer account fee, which is too costly for a “one-off app.”

Nonetheless, venture capital has sensed opportunity. Startups like Anything ($11 million raised) and VibeCode ($9.4 million seed round) are dedicated to solving the last-mile problem of “Vibe Coding” on mobile.

Of course, these personal apps have inherent flaws in quality, security, and maintenance, and are destined not to scale. But for creators, they don’t need to serve the masses, only themselves—this in itself disrupts the original supply-demand relationship of the software industry.

“Super Coders” Who Replace Entire Teams: The “Cracked Engineers” of the AI Era

If micro apps are a product of lowered coding barriers, then “Cracked Engineers” are a reflection of intense competition among professionals.

“Cracked” comes from gaming slang, referring to those with superb gameplay and “explosive” states.

In today’s Silicon Valley, it is used to define the ideal software engineer in the AI era: young (usually in their 20s), intensely ambitious, extremely sharp, and able to use AI tools for astonishing output.

J.X. Mo, co-founder of robotics startup Gradient, recently made a ruthless decision: cancel all intern recruitment. After interviewing applicants, he found it “not worth the time”—no one was “Cracked” enough.

Under the AI gold rush, startups are fanatical about peak productivity. Turing CEO Jonathan Siddharth believes that, with AI, a small, elite team can realistically reach $100 million in revenue in a year. Founders no longer need mediocre code workers who go by the book—they need special forces.

“Cracked Engineers” fundamentally differ from two familiar groups:

1. They are not “Vibe Coders”: Vibe Coders lack deep technical fundamentals and are just AI “prompt operators” (“Cursor Jockeys”). Cracked Engineers, however, have solid tech expertise, greatly enhance their output via AI, while also being capable of reviewing and correcting AI mistakes. They ride AI—they’re not mere passengers.
2. They are not traditional “10x Engineers”: The last generation’s “10x” engineers were typically over 30, working in big companies like Google, following procedures, and sometimes skeptical of AI coding tools. Cracked Engineers are younger, more rebellious, uninterested in corporate politics, and live by “work is everything.”

“One person replacing a whole team” is becoming reality.

Ron Arel, CEO of Intology, points out that a few people highly focused and adept at using Claude Code now output more than a 15-person team could previously do without AI.

Adam Gleave, co-founder of Far.AI, reveals that one employee—using AI assistance—built a prototype large model fine-tuning software in weeks, work that would have taken an open-source community a year.

This high output often comes with extreme work intensity. These engineers generally accept, even embrace, the “996” work schedule (9am-9pm, 6 days a week), knowing that in the AI race, falling behind is being eliminated.

James Hawkins, co-founder of PostHog, describes them:

“They don’t care about office politics, clothing style, or grooming. The results speak for themselves.”

However, the craze for “super programmers” is not without risk.

Deedy Das, partner at Menlo Ventures, observes that some young engineers, eager to appear “Cracked,” deliberately act socially distant, use obscure language, or abandon all hobbies for work. He warns:

“The most effective technical leaders are usually great communicators—this isn’t a solo game.”

Recruitment expert Kelsey Bishop is more blunt: Many founders try to cover up business model flaws by hiring a “Cracked Engineer.” “They treat it as a bandage, but that doesn’t solve the real problem.”

Conclusion

As AI programming becomes much easier, the middle ground is vanishing.

For ordinary users, the right to build software is being handed down—everyone can be their own “product manager”; in professional spheres, the bar is rising infinitely, and only those who can truly master AI and fuse body and algorithm can survive the fierce gold rush.

This is an age of gold, but it is also an age of brutality.

This article is from WeChat Public Account “Hard AI”. For more cutting edge AI news, visit here.

Risk Warning and DisclaimerThe market has risks; investing requires caution. This article does not constitute personal investment advice and does not account for individual users’ specific investment goals, financial circumstances, or needs. Users should consider whether any views, opinions, or conclusions in this article suit their particular circumstances. Invest accordingly, at your own risk.