Why Zoho co-founder Sridhar Vembu is comparing AI tokens with ‘drug addiction’

zoho

• Zoho's Sridhar Vembu likens AI usage limits to drug withdrawal
• Debate grows on workplace reliance on generative AI tools
• Concerns raised about over-automation and skill resilience

Did our AI summary help?

Sridhar Vembu, co-founder and chief scientist at Zoho, has sparked a wider debate in the tech community after drawing a sharp analogy between artificial intelligence usage limits as drug-like dependency. His comment came in response to a real workplace moment — a team member struggling after running out of AI “tokens,” the usage credits that power many generative AI tools.

Vembu shared that one of his colleagues complained about receiving a “tokens exhausted” message and felt the AI was no longer helping as effectively. The frustration was not about software malfunction, but about suddenly losing access to a productivity boost they had grown accustomed to.

For Vembu, the emotional response felt familiar. He likened it to withdrawal — not from chemicals, but from efficiency. The sudden slowdown, he implied, mirrors how people react when something they depend on is abruptly taken away.

Why AI tools are creating strong dependency patterns

Modern AI systems dramatically compress effort. Tasks that once took hours of coding, writing, or analysis can now be completed in minutes. This speed creates rapid feedback loops — quick wins, faster results, and constant progress.

Psychologically, that rhythm can become reinforcing. When users hit token limits, the abrupt loss of momentum feels heavier than a normal slowdown. Several professionals responding to Vembu’s post echoed this sentiment, describing AI as a “productivity multiplier” that workflows are increasingly built around.

Not addiction — but workflow reliance

Some voices in the discussion pushed back on the addiction framing. They argued the issue isn’t craving AI itself, but having work systems designed around constant AI availability. When the tool disappears, productivity drops — similar to losing access to cloud services or internal software platforms.

Others compared it to running out of paid credits in an arcade: you know the limit exists, but the interruption still breaks the flow.

A larger warning about over-automation

Beyond humor, Vembu’s analogy points to a deeper concern. As AI becomes embedded in everyday work, there’s a risk of outsourcing too much thinking, problem-solving, and creativity to machines.

If teams can no longer function smoothly without constant AI assistance, businesses may be trading long-term skill resilience for short-term speed. The conversation now shifting across tech circles is less about token pricing — and more about how dependent modern productivity is becoming on generative AI tools.

In highlighting this moment, Vembu has opened a broader debate: whether AI is simply accelerating work, or quietly reshaping how humans think, build, and operate — and what happens when that acceleration suddenly stops.