Why ‘Multiplayer AI’ is Lagging, According to John Cutler

In a recent LinkedIn post, John Cutler explores the significant barriers preventing the widespread adoption of “multiplayer AI,” a concept where teams collaborate using AI tools rather than solely individuals. Cutler, utilizing a behavior diagnosis skill he developed based on the COM-B model and the Behavior Change Wheel, identifies that the current infrastructure and incentives are heavily optimized for solo use, leaving substantial collaborative value untapped.

Cutler highlights that the fundamental issue isn’t a lack of desire for collaboration, but rather an ecosystem built for individual productivity. He states:

“The short version: Everyone’s deep in single-player mode — massive personal .md repos, custom prompts, individual workflows. It’s working great… for each person, in isolation. But we’re leaving enormous value on the table by not thinking together.”

The Infrastructure Gap for Collaborative AI

One of the primary obstacles identified by Cutler is the sheer absence of multiplayer infrastructure. Current AI tools are predominantly designed for a singular user interacting with the machine. The transition from a personal solution to a shared organizational asset requires considerable effort, a step for which current systems are not equipped.

As John Cutler points out:

“There’s nowhere to go multiplayer. The infrastructure literally doesn’t exist. Every AI tool is built for ‘you and the machine.’ Going from ‘I solved this’ to ‘here, use this’ takes a ton of extra effort that nobody’s set up for.”

Lack of Concrete Practices and Misaligned Incentives

Beyond the technical infrastructure, Cutler argues that a lack of clear, actionable practices hinders the move towards collaborative AI. While terms like “shared context” and “distributed cognition” are discussed, there’s a deficit in defining what these look like in daily work. Without concrete examples, adoption remains elusive.

Furthermore, Cutler emphasizes that the incentive structures are misaligned. He notes that solving personal problems with AI offers immediate satisfaction, whereas contributing to a shared organizational resource often involves extra work with uncertain payoffs and little recognition. This dynamic strongly favors individual, or “single-player,” AI usage.

“The rewards are all wrong. Solve your own problem? Instant satisfaction. Share something with the org? Extra work, uncertain payoff, and nobody notices. Every incentive points toward single-player.”

Shifting Towards Collective Intelligence

Despite these challenges, Cutler offers a hopeful outlook, suggesting that the “seeds” for skill-sharing already exist in various pockets within organizations. He proposes several actionable steps to foster a more collaborative AI environment:

• Start with existing pockets of sharing: Identify and understand current collaborative practices to extract successful patterns.
• Create and seed shared spaces: Establish lightweight repositories with templates and pre-populated wins to provide immediate value.
• Make the invisible visible: Conduct audits to highlight redundant problem-solving across teams and ensure contributors and beneficiaries are recognized.
• Ruthlessly reduce friction: Streamline the process between creating a solution and making it available in a shared space.
• Recognize contributors: Increase visibility for those who share and benefit from shared resources.
• Leadership buy-in: Leaders must visibly use and contribute to shared resources to signal its importance.
• Reframe identity: Shift the focus from individual AI prowess to the collective intelligence that makes the entire organization smarter.

In conclusion, John Cutler’s analysis on LinkedIn presents a compelling case for the need to build the right infrastructure, practices, and incentives to unlock the full potential of collaborative AI, moving beyond the limitations of individual workflows.