Cognition Bought an AI You Text Like a Friend — Here's Why That Matters for Solo Builders

Last week, Cognition — the company behind the AI coding assistant Devin — acquired Poke, an AI agent you text like a friend. The deal valued Poke in the “low nine figures.” That’s a lot of money for an AI that uses slang and makes jokes. But it tells you something important about where AI tools are heading: personality is becoming a feature, not a decoration.

I’ve been testing AI tools for over a year now, and the ones I keep coming back to aren’t necessarily the smartest. They’re the ones that feel right. That’s not a soft metric — it’s a competitive advantage. And Cognition just paid nine figures to prove it.

What Poke actually does

Poke isn’t another ChatGPT wrapper. It’s an AI agent that lives in your messaging app — iMessage, SMS, Telegram, or WhatsApp — and talks to you like a friend would. Not in a creepy, “I’m your bestie” way. In a practical way. It responds with personality, uses casual language, and handles tasks like scheduling, email management, reminders, and research without feeling like you’re issuing commands to a machine.

Over the past three months, Poke users exchanged more than 100 million messages on the platform. Hundreds of thousands of people use it. The most common use cases aren’t flashy — they’re the boring, daily stuff: managing emails, setting reminders, handling to-dos. The kind of tasks you’d delegate to an employee if you had one.

That’s the part that matters for solo builders. You don’t need an AI that can write poetry. You need one that can handle your inbox without making you feel like you’re talking to a customer service bot.

Why Cognition cares about personality

Cognition’s main product is Devin, an AI coding assistant that handles software engineering tasks. It can write code, submit pull requests, and debug issues. Technically impressive. But here’s the thing — Devin does one pull request at a time. It doesn’t remember what you asked it yesterday. It doesn’t have a personality.

The Poke acquisition changes that. Cognition wants Devin to feel less like software and more like a colleague. As Cognition co-founder Scott Wu put it: “The Interaction team has built an agent that people love: it’s proactive, it knows you, and it’s fun to talk to. That’s exactly how working with Devin should feel.”

Think about what that means in practice. Right now, if you use an AI coding tool, you’re basically issuing instructions to a very smart search engine. There’s no continuity. No rapport. No sense that the tool understands your preferences over time. Poke’s interaction model could change that — making Devin remember your coding style, your project context, your communication preferences.

The personality gap in AI tools

This acquisition highlights something I’ve noticed while comparing AI tools: most of them are technically capable but emotionally flat. They give you the right answer in the most generic way possible.

I ran an experiment recently where I asked three different AI tools to write the same product description. The results were nearly identical — not just in structure, but in voice. Same adjectives. Same rhythm. Same “elevate your space” energy. That’s the groupthink problem, and it extends to how tools interact with you.

When every AI assistant sounds the same, the one that sounds different wins. Not because different is automatically better, but because different is memorable. And memorable is what keeps people coming back.

Poke figured this out early. Instead of building a more capable tool, they built a more likable one. The capability came later. That’s backwards from how most AI companies think, and it’s why Cognition paid nine figures to acquire them.

What this means for your tool stack

If you’re a solo builder choosing AI tools, this acquisition signals a shift in what to look for. Here’s what I’d pay attention to:

Memory matters more than raw intelligence. Poke’s biggest selling point isn’t its personality — it’s that it remembers you. It knows your preferences, your context, your history. Most AI tools treat every conversation like a blank slate. That’s going to change. Tools that maintain context across sessions will have a real advantage.

Interaction style affects adoption. You can have the most powerful AI in the world, but if talking to it feels like filling out a form, you won’t use it. Poke’s success with 100 million messages in three months isn’t because it’s the smartest AI — it’s because people actually enjoy using it. For solo builders, that means choosing tools you’ll actually stick with, not just the ones with the best benchmarks.

The “AI colleague” model is coming. Right now, most AI tools are search engines with better interfaces. The Poke acquisition points toward something different: AI that works alongside you, remembers your preferences, and improves over time. That’s closer to hiring an employee than using a tool.

The pricing question nobody’s asking

Here’s the part of the Poke story that didn’t make headlines: it was expensive to run. Poke’s founder admitted the product was difficult to turn a profit on, despite hundreds of thousands of users. That’s a familiar pattern in AI — technically impressive, financially unsustainable.

Cognition can absorb those costs because they’re selling Devin to enterprises at enterprise prices. But for solo builders, it raises a question: are the AI tools you’re building your workflow around going to be here in two years?

I’ve been testing pricing across AI tools for months, and the pattern is clear: the tools with personality and memory features tend to cost more to run. That cost gets passed to you eventually. When you’re choosing between a $20/month AI assistant and a $200/month one, the difference often isn’t intelligence — it’s the interaction layer.

The bottom line

Cognition’s acquisition of Poke isn’t just another AI M&A deal. It’s a signal that the companies building AI tools are starting to care about how those tools feel, not just what they can do. For solo builders, that means the next wave of AI tools will be more personal, more contextual, and more expensive. Choose tools that remember you and that you actually enjoy using — because the technical capability gap between AI tools is shrinking, but the personality gap is about to become the deciding factor.

Want to stay ahead of these shifts? Check out what’s coming next in AI tools or see the tools I actually use every day.