The next shift in software creation is already underway, and Patrick Ma has been building at the center of it.
As a senior AI engineer and Quora’s first dedicated AI agent engineer on Poe, Ma helped architect the first agent on Quora’s AI product and has worked on tools that let users create AI-powered apps through natural language. His work sits inside one of the most important changes in technology today: software is becoming something more people create without first becoming professional programmers.
“For a long time, software creation belonged mostly to people who could code,” Ma says. “AI changes that. The question becomes how to give more people access to the power of software without losing quality.” That question has shaped Ma’s recent work on Poe, Quora’s AI platform that gives users access to leading AI models and supports AI-powered app creation. Ma served as tech lead on several Poe launches covered by TechCrunch, including Previews, App Creator, and group chats across AI models. These products are part of a larger movement toward making software creation more interactive and accessible.
The significance is the change in who gets to begin. A person who has an idea for a tool, workflow, interactive page, or small application no longer has to start with a blank code editor. The first step can be a plain-language description of what they want the software to do. “Natural language is becoming a creation interface,” Ma says. “That opens the door for people who have useful ideas but never had the traditional path into software.”
Ma’s authority on that subject comes from working on AI products before the current agent wave became a mainstream industry focus. He has worked on Poe for more than three years and has watched the field move from chat-based AI toward tools that can act, build, evaluate, and assist users across more complex tasks.
At Quora, he architected Poe’s first agent before the industry had fully settled on what the word “agent” should mean. That early position required a close understanding of how users would actually try to create with AI. “Building early in a category forces you to stay close to the user,” Ma says. “The terminology may change, but the product still has to help someone get from an idea to something useful.”
His work has had unusually broad reach. Quora has more than 400 million unique monthly visitors, and Poe serves millions of users. Poe has also been consistently named among the top 100 consumer generative AI apps by Andreessen Horowitz. Because Ma’s work focuses on core product experiences across Quora and Poe, the impact reaches users at scale.
That scale matters because AI-powered creation is not meant for a small circle of technical users. At its best, it reaches people with very different needs: someone pressure-testing a new product, an educator building something for the classroom, a small team trying to simplify operations, or a professional who sees a repetitive task and wants software to handle it.
“People should not need to know every technical detail before they can make software useful in their lives,” Ma says. “AI can give more people a way to start.” That belief also shows up outside Ma’s formal product work. One of his side projects saved a doctor more than 1,400 hours by applying software to a real professional workflow. The project reflects a practical version of the same principle behind his AI-agent work: technology is most powerful when it helps people solve problems they have lived with for too long.
The doctor did not need a flashy demo. The need was specific, time-consuming, and tied to decades of medical notes. Ma built software to reduce the burden and give that time back.
“That project mattered to me because it was so concrete,” Ma says. “It was not about proving that AI was exciting. It was about helping someone with a real problem.” Ma is interested in tools that expand what people can do, not only tools that impress other engineers. His focus is on giving more people access to software-building power while keeping the product useful enough to hold up in real life.
Poe’s App Creator points in that direction. The product gives users a way to create AI-powered apps and experiences, shifting some of the creation process into language. For non-engineers, that can make the difference between having an idea and actually producing something they can use or share. “Many people know the problem better than the software industry does,” Ma says. “They know the workflow. AI can help turn that knowledge into something functional.”
That does not mean every user becomes an engineer overnight. It means the distance between idea and prototype gets shorter. That shift opens the door to broader experimentation. People can test ideas sooner to find out whether a tool is actually useful. Then turn practical knowledge into software, even when they never had an easy path to build it before.
Ma’s earlier work at Quora also reflects his ability to build for visible, high-impact product needs. On the ads team, his work helped grow company revenue by more than $3 million per year through improvements to the ads product. During his first half year, he led a project that reduced fraudulent advertisers by 98 percent. Later, when Quora began its AI initiative with Poe, he became one of the first engineers working on the iOS app and contributed to major feature launches.
Throughout, Ma has taken on product problems with high user or business impact. “My career has been shaped by working on the most important problem in front of the company,” Ma says. “That is where the work becomes meaningful.”
His technical judgment has also been beyond Quora. Ma serves as a venture scout for UpHonest Capital, where he sourced a deal that resulted in a $1 million investment. He has joined the CRV venture scout program and has judged AI and technology hackathons, including the UC Berkeley AI Hackathon, Modal Autoresearch Hackathon, and WeaveHacks.
In those roles, Ma evaluates builders at a different stage of the creation process. The same question appears in a new form: who is turning a strong idea into something people can actually use? “Good builders pay attention to the gap between the idea and the person who needs it,” Ma says. “That is where the real product work happens.”
His next chapter will take him to Cognition, maker of Devin, the first AI software engineer. The move brings Ma closer to the frontier he has already been working toward: AI systems that help people build better software, whether they are engineers trying to move faster or non-technical users trying to turn an idea into something functional.
For Ma, the future of software creation reaches beyond code automation. It changes who gets to build. A founder can test an idea sooner, a professional can shape a tool around the way work actually happens, and a user with a specific problem no longer has to wait for a software team to understand it.
“AI can help more people become builders,” Ma says. “That is the exciting part. My work is about making that power available to more people.”
