AI Mania Is Eviscerating Global Decision-Making
Consultant Nik Suresh exposes the absurdity of AI decision-making in large companies, from executives who have never used AI to engineers performing for AI leaderboards, revealing the twisted logic behind the hype.
- Some executives approve billion-dollar AI strategies without ever using any AI tool.
- An engineer had AI rewrite an entire Go codebase in Zig just to stay on the AI leaderboard and keep his job.
- Vendors dare not correct customers’ unrealistic AI expectations for fear of offending executives and losing contracts.
- AI decision-making is hijacked by social narratives and career risks, rather than driven by technical feasibility.
Origins: A spicy revelation sparks reflection
The AI frenzy has been sweeping the globe for over two years now, with companies rushing to adopt AI projects in fear of being left behind. But a recent article by independent consultant Nik Suresh (recommended by well-known developer Simon Willison) exposed several jaw-dropping cases that force us to pause and rethink: are we collectively sliding into “AI theater”? When rational decision-making is replaced by hype, the consequences can be worse than technical failure.
Dissection: Three typical “AI decision-making maladies”
Executives making AI strategies while being “AI illiterate” Nik recounts an extreme case: a senior executive at a company with over $2 billion in revenue privately admitted to never having used ChatGPT or any AI tool, right after presenting a comprehensive AI strategy. This strategy was based not on hands-on understanding but on sales materials, peer pressure, and fear of being left behind. It’s like letting someone who has never driven design a car manufacturing process—the outcome is predictable.
Employees performing AI to keep their jobs At another company, an internal “AI adoption leaderboard” pressured engineers to compete. One engineer confessed to creating a parallel repo and having AI rewrite an entire Go codebase in Zig while he slacked off, just to boost his stats on the leaderboard and keep his job. Here, AI is not a productivity tool but a new weapon in workplace rat race.
The vendor’s “emperor’s new clothes” The most ironic part is the sales cycle. Customer executives publicly claim “100x productivity gains” from AI, and vendors know it’s unrealistic, but no one dares to speak up. Because pointing out the truth would undermine the customer executive’s credibility, be perceived as an attack or heresy, and likely lead to contract cancellation. The cost of honesty is too high, so everyone maintains the collective illusion.
Trend Insight: Why is AI decision-making hijacked?
These phenomena reveal an underlying logic: in today’s corporate environment, AI is no longer just a technical issue; it has become a complex symbol burdened with social pressure, career risk, and market narrative. Decision-makers care more about appearing “AI-ready” than about the actual value AI can create. This isn’t new in hype cycles—during the dot-com bubble, companies just added “.com” to their names to boost stock prices—but AI is unique because it’s harder to intuitively grasp, leaving more room for “performance.”
Digging deeper, the incentive structures are twisted. Ordinary employees game AI leaderboards because the company measures their “innovativeness” that way; executives boast about AI results because the board and investors want to hear it; vendors echo the lies because it’s the only way to survive. Individual rational actions at each node converge into collective irrationality.
Practical Value: How can we navigate this?
For tech professionals, it’s not hopeless. First, maintain critical thinking: when you hear exaggerated productivity claims, ask “Where’s the data? Has it been validated?” Second, push for genuine AI literacy in your organization—let decision-makers get hands-on experience with basic tools, not just PowerPoints. Third, design healthy incentive metrics, such as measuring actual business outcome improvements rather than AI feature usage counts. Finally, if you’re a vendor or consultant, consider being the child who says the emperor has no clothes; honesty may cost you in the short term but builds trust in the long run.
Counterintuitive insight: The most dangerous factor is often not the technology
We habitually blame AI failures on inaccurate models or insufficient data, but these new examples show that the “human” element may be more damaging. A strategy set by an executive who has never used AI is harder to fix than a software bug. Moreover, job security in the AI era may not depend on how much you can do with AI, but on how much critical thinking and domain insight you have that AI can’t replace. That engineer who used AI to rewrite Go into Zig may appear cutting-edge, but he might actually be training the tool that replaces him.
Nik Suresh’s revelations, though spicy, serve as a much-needed dose of sobriety for this overheated industry. AI’s value is real, but only if we step out of the “AI theater” and return decision-making to rationality.
Analysis by BitByAI · Read original