What If AI Worked Alone for Months?

"Plan and execute a digital marketing campaign for my restaurant." What if you could say this once and have AI work on it for months without another word from you?

Perplexity's new "Computer" tool promises exactly that. It's not just another AI assistant—it's a system that breaks down complex goals into subtasks and assigns them to different AI agents, potentially running for hours or even months autonomously.

How It Actually Works

Computer operates on a "workflow orchestration" principle. Tell it to build an Android app, and it'll divide the project into planning, design, development, and testing phases. Each phase gets assigned to the AI model best suited for that specific task—GPT-4 for creative work, Claude for analysis, others for coding.

Currently available only to Perplexity Max subscribers, the company describes it as a system that "creates and executes entire workflows" without constant human intervention.

The Developer Divide

Silicon Valley's reaction has been polarized. Startup founders are excited about potentially 90% faster MVP development, while experienced developers remain skeptical.

"Can AI really manage complex projects for months without going off track?" questioned one developer on Hacker News. "Without checkpoints, you could end up with months of work in the wrong direction."

The concern isn't just technical—it's philosophical. Are we ready to surrender that much control to automated systems?

Enterprise Implications

For businesses, Computer represents a fundamental shift. Instead of managing AI tools, companies might soon be managing AI workflows. This could dramatically reduce project timelines but also raises questions about oversight and accountability.

Some enterprises are already experimenting. A mid-sized marketing agency reported testing similar multi-agent systems for campaign management, seeing 60% reduction in manual coordination time.

The Control Paradox

Here's the interesting tension: Computer promises more freedom by handling long-term tasks, but it also requires unprecedented trust in AI decision-making. Users must define outcomes clearly upfront, then step back for potentially months.

This challenges traditional project management approaches. Most businesses are used to iterative feedback loops, not "set it and forget it" automation.

What This Means for Workers

If Computer succeeds, certain roles will evolve rapidly. Project coordinators might become AI workflow architects. Middle managers could focus more on strategy than execution oversight.

But there's a flip side: as AI handles more autonomous work, human skills in long-term planning and quality assurance become more valuable, not less.