The software industry is changing fast. By 2025, Meta plans to use AI as mid-level engineers to write code. This doesn’t just change what software gets built, but also how it’s made. Satya Nadella, CEO of Microsoft recently said that traditional apps might be replaced by AI-driven agents. Mark Zuckerberg, CEO of Meta’s comments show this is already becoming a reality. Software development is shifting from static tools to AI that responds dynamically to tasks.
Meta has the resources to move quickly—money, GPUs, and infrastructure. If an AI can write solid code now, what will it be able to do in 2026, 2030, or 2035? By then, most software might be written by AI on the fly, as it’s needed.
This breaks with decades of how code has been written. Instead of pre-made apps, AI agents could create code in real-time—whether for a user interface, a database connection, or running a process. Applications could become more fluid, designed as they’re used.
The impact on jobs is huge. AI won’t just write more code—it will change how we think about programming. Humans will still oversee AI at first, but as tools improve, the bottleneck might shift to us. Will AI also review code? Will the code even look like what we’re used to? These questions are coming soon.
Zuckerberg sees this as a way to boost human creativity, not replace it. AI can handle repetitive tasks, freeing people to do more interesting work. This has happened before. Automation moved people from farming to services and innovation. The same could happen now. People might work fewer hours, but those who stay engaged will have powerful tools to achieve more.
Not everyone is so hopeful. Salesforce CEO Marc Benioff recently froze engineer hiring, saying AI increased productivity by 30%. Companies like Salesforce are already scaling back teams because AI is so efficient. This shows the transition will be bumpy, with job losses in some areas.
But history suggests demand will grow. When computing got cheaper, use exploded. As AI becomes more accessible, its uses will multiply. Companies will solve more problems with fewer people and lower costs.
This shift will be messy but unavoidable. Entrepreneurs may benefit most. AI could help small teams build billion-dollar companies. Tools already exist, like code generators and advanced AI platforms, letting founders focus on solving problems instead of infrastructure.
The future of software isn’t just writing code faster. It’s rethinking what software is. As AI becomes part of workflows, the line between applications, tasks, and creativity will blur. Those who adapt now will lead the next wave of change.
At Groundsight, we use AI daily to help generate code, leveraging the current state of the art. It works well, though it’s not perfect and sometimes requires guidance. Even so, it offers a huge boost in productivity. Like many in the industry, we’ve seen how reliance on AI can sometimes overlook important nuances in solving complex problems. Over time, we believe AI will improve at understanding these subtleties, making it an even more valuable tool.