Cognition raises $1 billion and CEO Scott Wu emphasizes that AI agents should not replace programmers
Artificial Intelligence is at the center of one of the hottest conversations in tech right now, and Cognition just threw more fuel on the fire.
The startup, known for creating Devin, one of the most talked-about AI coding agents in recent memory, just raised $1 billion in a round that valued the company at a staggering $26 billion.
A number that big turns heads no matter who you are, right?
But what grabbed even more attention was what CEO Scott Wu said right after all of this went public.
In a landscape where people already have their finger on the trigger waiting for the next mass layoff headline because of AI, Wu went in the opposite direction and made it crystal clear that Cognition never intended to take anyone’s job, least of all programmers.
Sounds contradictory?
It might seem that way, but the story behind it is way more interesting than the billion-dollar headline. 👇
What Cognition is actually building
Before diving into the whole replacing-humans-or-not debate, it’s worth understanding what Cognition is really doing. Devin was introduced to the world as an Artificial Intelligence agent capable of executing complete software development tasks, from writing code to debugging, testing, and even handling deploys. In practice, that means it’s not just an assistant that autocompletes lines of code. It can take a problem, plan a solution, and execute entire stages of a programming project autonomously.
That’s different from anything we’ve seen before in tools like GitHub Copilot or ChatGPT, which function more like a coworker helping you out while you do the heavy lifting.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Devin operates at a level that a lot of people in the industry are still trying to fully wrap their heads around. It uses its own development environment, accesses the internet, reads documentation, writes and runs code, and even learns from its mistakes throughout the process. All of this happens in a chained fashion, like a continuous cycle of reasoning and execution. And it’s exactly this autonomous agent model that made the market’s eyes light up, putting Cognition on the map as one of the most promising AI startups in the world in record time.
According to Scott Wu himself, the goal from the very beginning was to create something that works as a teammate. In his words, when they started building Devin, the idea was simple: this is your buddy that helps you build more. Wu even showed off a little stuffed animal holding a computer that he keeps on his desk, a kind of physical mascot that represents the spirit of the AI agent. For him, Devin is that companion who’s there to add value, not to compete with the programmer sitting next to you.
Scott Wu and the passion for programming that shaped Cognition
To understand where this more human-centered vision of AI agents comes from, it helps to know a bit about Scott Wu’s background. The guy started coding at nine years old and is considered one of the most talented young competitive programmers of all time. While still in second grade, he won a national math competition designed for seventh graders, which kicked off an entire childhood dedicated to math and programming tournaments.
That phase also connected him with other young prodigies who would later go on to found their own tech startups, like Alexandr Wang, founder of Scale AI. In other words, Wu’s journey is deeply rooted in the world of programming, and that explains a lot about why he insists that AI shouldn’t steal the joy of coding from people.
Wu made a point of stressing this in his statements. He highlighted that most software engineers love building things. When you ask them why, the answer is usually something like: I can create something from scratch, turn an idea into a real product, into an experience. And it’s exactly that emotional connection to the act of creating that Wu wants to preserve, even as increasingly powerful autonomous agents enter the picture.
Devin’s actual role inside Cognition itself
This is where things get especially interesting. Cognition revealed that Devin is already responsible for a massive share of the software development work inside the company itself. According to data shared by the startup, 89% of the code committed by the company’s engineers was actually committed by Devin. The rest was handled by local agents from Windsurf, the competing coding tool that Cognition acquired last year.
That number is impressive and might seem contradictory to the whole not-replacing-humans message, but Wu explains the context. Devin’s role within the company is primarily focused on tasks that many programmers don’t enjoy doing in the first place:
- Updating legacy software and bringing old systems up to modern standards
- Migrating applications from one platform to another
- Handling repetitive, long-tail maintenance tasks
- Dealing with incremental adjustments that eat up time but don’t require creative decision-making
In Wu’s view, this frees human programmers from a huge chunk of operational work and allows them to dedicate more time to the creative side of software development. Instead of spending hours updating dependencies or refactoring old code, the engineer can focus on designing new features, thinking about system architecture, and solving problems that genuinely demand human reasoning.
Replacing humans or partnering with them? Scott Wu’s take
Scott Wu was pretty direct when asked if Devin could replace, say, a mid-level programmer. The answer was yes and no at the same time. He acknowledges that the agent can work independently on certain tasks, operating at a level somewhere between a junior engineer and a mid-level one, depending on the complexity of what’s being asked.
But Cognition’s official position is clear: the mission was never to put software engineers out of work. Wu emphasized that everyone at the company is a programmer and that he personally started coding as a kid. The core idea is that Devin can take on repetitive, time-consuming, and lower-value tasks, freeing up human professionals to focus on more complex, creative, and higher-impact problems.
This perspective aligns much more with the reality of tech companies than the robot-stealing-jobs narrative that goes viral on social media. Anyone who’s ever worked on a software development team knows that a huge portion of engineers’ time is consumed by tasks that require neither creativity nor strategic decision-making. If an AI agent can handle that autonomously and reliably, the practical outcome isn’t anyone getting fired. The outcome is that the human team gets to work on things that actually matter for the business and the product’s evolution.
Wu also made an interesting analogy to explain how he sees AI agents in the context of computing history. Just as visual development environments abstracted software creation away from machine instructions, he sees agents as another layer of abstraction between a product vision and its execution. In other words, it’s not the end of programming. It’s programming evolving, just like it has so many times over the decades. 🤖
The 2026 landscape and the wave of layoffs in the name of AI
Wu’s stance carries even more weight when you look at the current landscape. The year 2026 has been marked by a steady wave of tech CEOs announcing staff cuts justified by the adoption of Artificial Intelligence. Every week brings a new headline about companies shrinking their teams with the argument that agents and language models can do the work of dozens of people.
In this environment, Wu’s statement sounds almost like an act of resistance within the AI startup ecosystem itself. While other industry leaders seem to treat replacing workers as inevitable and even desirable, Cognition’s CEO is saying it doesn’t have to be that way and that, in fact, it shouldn’t be that way.
That doesn’t mean he’s ignoring the transformations happening around us. Wu acknowledges that we’re facing a profound shift and went so far as to say we’re on an intense journey when it comes to the concept of self-directed software, where the agent learns and improves on its own. The difference is that he believes this transformation should keep humans in control of the decisions.
AI agents beyond code and the future according to Cognition
Wu also sees a horizon that extends well beyond software development. He believes AI agents will move into other sectors and learn tasks in areas like customer service and even medicine. But his hope is that in those fields too, the goal will be to augment the capabilities of human professionals, not eliminate them.
According to him, code and software were the first to move in this direction, but we’ll see this dynamic repeat across many other industries. And he makes a point of reinforcing a principle he’s considered fundamental from the start: the final decision should always belong to the human. That applies to software engineering, but also to any other profession that ends up being impacted by autonomous agents.
This vision of the future matters because it positions Cognition not just as a product company, but as an important voice in the ethical and practical debate about how AI should be integrated into the workforce. At a time when the concept of RSI (Recursive Self-Improvement) is gaining traction as the sector’s new buzzword, Wu’s stance offers a necessary counterpoint to the unchecked hype.
What this all means for the programming market
When a startup is valued at $26 billion before it even has a widely established product on the market, that says a lot about the expectations investors have for the future of software development with Artificial Intelligence. The big money is betting that AI agents like Devin will completely redefine how tech teams work over the next few years, and that companies that figure out how to integrate this technology smartly will pull ahead of the competition.
For anyone working in programming, this scenario calls for a more nuanced read than what’s circulating on social media. The end of the programmer’s job narrative is catchy, but way too simplistic to capture the complexity of what’s actually happening. What the data and statements from Cognition itself show is that demand for professionals who know how to work with AI, not against it, is set to grow significantly.
That includes people who understand how to guide, review, and validate the work of these agents, who can identify when the tool got something wrong and why, and who can translate business needs into instructions that an autonomous agent can execute with quality. The software engineer’s profile is changing, but it’s not disappearing. It’s becoming more strategic.
Cognition’s acquisition of Windsurf last year also reinforces this trend. By integrating different AI coding tools under the same umbrella, the company is building an ecosystem where autonomous agents and local assistants coexist, each with a defined role within the workflow. This suggests that the future won’t be dominated by a single tool that does everything on its own, but by a set of specialized agents working side by side with human engineers.
What we take away from all of this
The software development market is going to change — that’s pretty much a given. But the direction that change takes depends heavily on how AI companies like Cognition position their products and how professionals in the field adapt. The good news is that, at least for now, the most influential voices in this space are signaling a transition built on collaboration, not replacement.
Scott Wu could have easily ridden the wave of apocalyptic hype and sold Devin as the ultimate programmer replacement. That would have generated even splashier headlines and probably boosted the startup’s valuation narrative even further. But he chose a different path, and that says a lot about how Cognition plans to position itself for the long haul.
At the end of the day, the most important message that came out alongside the billions from this funding round has nothing to do with money. It’s about the idea that Artificial Intelligence can be built to empower people, not to discard them. And at a time when the tech industry seems increasingly willing to swap humans for algorithms without a second thought, that message is more valuable than any funding round. 💡
