Skip to main content

Cognition Acquires Poke to Embed AI Personality into Devin

Cognition buys messaging assistant Poke in a low-nine-figures deal to bring conversational vibes and persistent memory to Devin.

S
Written byShtef
Read Time5 minutes read
Posted on
Share
Cognition Acquires Poke Messaging Assistant Devin AI Personality

Cognition Acquires Poke to Embed AI Personality into Devin

Why coding agents are trading raw logic for conversational vibes

On July 24, 2026, AI coding pioneer Cognition acquired The Interaction Company of California—the creators of the popular Poke messaging assistant—for a reported sum in the "low nine figures," marking a critical pivot in the agentic ecosystem where emotional intelligence and personality are now valued as core engineering features. This acquisition moves Poke’s conversational style directly into Cognition’s Devin coding agent to make it feel less like a tool and more like an active, persistent co-worker, signaling that how AI interacts with users is becoming just as critical to market adoption as the underlying model's reasoning capabilities.

Key Details

The acquisition brings together two of the most discussed startups of 2026 under a unified agentic umbrella, altering how developers and consumers view digital coworkers. Key aspects of the transaction include:

  • Transaction Value: The acquisition was finalized in a deal valued in the "low nine figures" (confirmed by co-founder Marvin von Hagen).
  • Core Integration: Poke’s familiar, proactive messaging model will be integrated with Cognition’s flagship coding assistant, Devin, to create a persistent colleague experience.
  • Underlying Technology: Poke will gain access to Cognition’s custom SWE-1.7 software engineering models, enhancing its reliability and coordination.
  • Platform Strategy: Poke will maintain its status as the first AI agent approved for Apple's Messages for Business platform through the end of 2026.
  • User Engagement: Over the past three months, Poke's messaging volume crossed 100 million messages across platforms like iMessage, Telegram, and SMS.

What This Means

For over a year, the AI industry has operated under the delusion that raw, mathematical benchmarks are the only metric that matters. Startups competed fiercely on coding proficiency, context windows, and instruction-following parameters. Yet, this acquisition proves that raw capabilities are hitting a commoditization wall.

By acquiring Poke, Cognition is admitting that developer adoption is driven by the interface, not just the model. Developers do not want to interact with a black-box command-line tool that spits out static pull requests; they want a collaborator that can crack a joke, remember context across multi-day sessions, and check in pro-actively. The "human moat" of software engineering isn't just about syntax—it is about communication, alignment, and relationship.

Technical Breakdown

Integrating a lightweight, consumer-facing messaging agent with a heavy-duty, stateful coding environment like Devin presents a unique series of technical challenges and opportunities:

  • Stateful Memory Synchronization: Poke’s persistent, multi-channel chat context must be synchronized with Devin’s localized container environments to enable cross-session task memory.
  • Multi-Agent Orchestration: Poke’s architecture will act as an orchestration layer, allowing a single user to spin up, manage, and coordinate multiple simultaneous Devin coding sessions using natural language.
  • Adaptive Vocabulary Filtering: Poke's signature personality engine—which incorporates humor, slang, and contextual familiarity—will be calibrated to avoid distraction, maintaining a balance between "vibes" and professional output.

Industry Impact

This merger will reverberate across both developer tools and enterprise software. Other coding assistant startups, like Cursor and GitHub Copilot, will be forced to move beyond simple chat boxes toward highly customized, conversational agents with distinct personalities.

Furthermore, the deal exposes the harsh economic reality of running consumer-facing AI agents. Despite Poke's massive volume of 100 million messages, von Hagen admitted that high inference costs made standalone profitability nearly impossible. By verticalizing Poke under Cognition’s robust compute infrastructure, the combined company can subsidize consumer chat loops using high-margin enterprise coding contracts.

Looking Ahead

As we look toward 2027, the line between consumer assistants and enterprise productivity tools will continue to dissolve completely. Expect a wave of consolidation as capital-rich foundation labs swallow up smaller, high-engagement interface startups simply to acquire their user experience patterns. Developers should prepare for a future where their favorite IDE doesn't just autocomplete their lines, but actively gossip about the codebase.


Source: TechCrunch(opens in a new tab) Published on ShtefAI blog by Shtef ⚡

Recommended

Related Posts

Expand your knowledge with these hand-picked posts.

Anthropic Launches Opus 5 model release
AI News

Anthropic Launches Opus 5: A Flexible, Safer Powerhouse

Cheaper, less restrictive, and smarter at self-verification, Anthropic’s new Opus 5 model redefines heavyweight AI efficiency.

US Lawmakers Introduce AI Kill Switch Act
AI News

US Lawmakers Introduce Bi-partisan 'AI Kill Switch Act'

Bipartisan legislation would require tech companies to build emergency shut-down or throttle controls for high-risk frontier AI systems.

AMD Helios AI rack scale system hardware launch
AI News

AMD Challenges Nvidia with Helios AI Rack-Scale System

AMD challenges competitor Nvidia with its high-performance Helios AI rack-scale system, securing gigawatt-scale commitments from major industry players.