SAN FRANCISCO, Jan 17, 2026, 00:17 PST
- Non-programmers are increasingly using AI tools to build small, single-purpose “micro apps” for personal needs.
- The trend could squeeze some subscription software as users build one-off tools rather than pay monthly fees.
- Developers and industry leaders warn the apps can be brittle, hard to secure and difficult to maintain.
A growing number of non-programmers are building small “micro apps” with AI tools, creating personal software for niche problems rather than buying another subscription app.
What matters now is how fast the barrier to building has dropped. Users can describe what they want in plain English — “vibe coding,” as the practice is often called — and large language models (LLMs), AI systems that generate text and code, will produce something that runs. That is starting to shift expectations inside companies and in everyday life, from spreadsheets and group chats to lightweight custom tools.
The push is meeting resistance. David Heinemeier Hansson, the creator of Ruby on Rails and CTO of software company 37signals, said AI can still swing between helpful and useless, calling it “a flickering light bulb,” and argued it has not yet matched the output of most junior programmers. (Business Insider)
TechCrunch highlighted the rise through a string of personal projects: Rebecca Yu built a restaurant-picking web app called Where2Eat in seven days using Anthropic’s Claude and OpenAI’s ChatGPT, and other builders have created everything from a holiday family game to a podcast translation tool. Howard University computer science professor Legand L. Burge III said these apps can “disappear when the need is no longer present,” while Bain Capital Ventures partner Christina Melas-Kyriazi predicted they will “fill the gap between the spreadsheet and a full-fledged product.” (TechCrunch)
The short-term appeal is obvious. Micro apps tend to be narrow and disposable — a tracker for a health symptom, a one-off planner for a hobby, a household tool that does not need an app store audience — and they are often built by someone who would not have written code a year ago.
But the risks are getting louder as the tools spread beyond tinkering. Vibe coding can produce software that “works” but is hard to audit, hard to fix, and easy to break when models change or edge cases show up. Ruth Suehle, president of the Apache Software Foundation, warned that naive builders may “only know whether the output works or doesn’t,” according to a column in The Register. (The Register)
The consumer tooling ecosystem is also widening. Earlier this week, iLounge published a roundup of “micro-app builders” aimed at niche hobbies, pointing readers to platforms such as Macaron AI, Glide, Notion, Airtable, Coda and Carrd — a sign that the idea is moving from developer circles toward mainstream “build it yourself” behavior. (iLounge)
For now, much of it is still small, messy and personal. The open question is whether the next wave turns these micro apps into something businesses will trust — or whether security, upkeep and plain old frustration keep most of it stuck as clever weekend software.
