The video criticizes major tech companies like Meta for evaluating employees based on AI usage metrics, arguing that this approach encourages unproductive behavior and overlooks true employee value. Instead, it advocates focusing on outcomes and retaining employees who consistently deliver meaningful results, regardless of their AI tool usage.
The video discusses a concerning new trend among major tech companies, including Meta, Microsoft, Shopify, and Google, where employees are being evaluated based on their AI skills, particularly their ability to use AI tools effectively. The speaker expresses strong disapproval of this approach, arguing that it encourages bad behavior and leads to poor product outcomes. Meta, for instance, plans to make AI-driven impact a core expectation by 2026, assessing employees on how they use AI to improve productivity and build impactful tools.
The speaker attempts to present a fair perspective by acknowledging that large language models (LLMs) can be useful for creating internal tools quickly, especially in large companies where many internal tools are neglected or poorly maintained. Using AI to build simple, functional tools that don’t require high quality or uptime can save significant engineering time and effort. This practical use of AI for internal operations is seen as a reasonable application that could improve efficiency without much downside.
However, the speaker quickly shifts to criticism, warning about the dangers of measuring employees based on AI usage metrics. They highlight a well-known problem in corporate environments: when a metric is measured, people optimize for that metric, often in unproductive or misleading ways. Just as counting lines of code is a poor measure of a developer’s performance, measuring AI usage or prompt skills does not necessarily correlate with quality or effectiveness. Two employees using the same AI tools can produce vastly different results, and focusing on AI usage alone ignores the true value of their work.
The speaker proposes an alternative approach, suggesting that companies should focus on retaining good employees who consistently deliver valuable results and solve real problems, regardless of whether or how much they use AI. The emphasis should be on outcomes and employee satisfaction rather than artificial metrics tied to AI usage. Transparency about tool use is important, but the method of achieving results should not be rigidly controlled or measured by AI-specific criteria.
In conclusion, the speaker strongly criticizes the trend of grading employees on AI skills, likening it to the flawed practice of measuring lines of code. They argue that good employees should be recognized and retained based on their overall contributions, not on how well they can prompt or use AI tools. The video ends with a call for companies to rethink their evaluation strategies and avoid imposing unnecessary constraints that could stifle genuine productivity and innovation.