← Retour au blog
tech 9 June 2026

LLMs and Performative Productivity: An Illusion of Progress?

Large Language Models (LLMs) promise to revolutionize our productivity, but do they mask a less flattering reality? Dive into the real impact of LLMs on our ability to be genuinely productive.

Article inspired by the original source
LLMs and performative productivity ↗ joshcollinsworth.com

Introduction

Large Language Models (LLMs) like GPT-4 have transformed how we approach our daily tasks. These tools promise increased productivity, enabling us to code faster, write more efficiently, and solve complex problems in the blink of an eye. But is this apparent efficiency truly synonymous with real productivity, or is it performative productivity, where the appearance of efficiency masks a lack of substantive progress?

The Impact of LLMs on Work

When LLMs burst onto the tech scene, they were met with enthusiasm. Developers, for instance, quickly adopted these tools to navigate complex codebases and automate repetitive tasks. According to a 2023 McKinsey study, companies using LLMs reported a 30% increase in project development speed.

Productivity or Mere Appearance?

However, this increase in speed doesn't necessarily translate to increased productivity. As Josh Collinsworth noted in his blog, using LLMs can sometimes give the illusion of accomplishing a lot, when in reality, you're only scratching the surface of problems. Developers might generate quick code without truly understanding its foundations, leading to potential errors and long-term technical debt.

The Dangers of Performative Productivity

Performative productivity is a phenomenon where the focus is on the appearance of completed work rather than its substance. With LLMs, it's easy to fall into the trap of producing more without real added value. A 2024 Harvard Business Review report indicates that 40% of features created with the help of LLMs are never used by end-users.

Case Studies

Consider the example of startup X, which integrated LLMs into its product development processes. While the team was able to launch five new features in record time, a post-launch analysis revealed that only two were actually used by customers. The other three were abandoned, wasting valuable resources and time.

Rethinking the Use of LLMs

To avoid the trap of performative productivity, it's crucial to rethink our use of LLMs. Instead of focusing solely on speed and quantity, decision-makers need to assess the long-term impact of the projects and features developed.

Strategies for Genuine Productivity

  • Ongoing Training: Ensure that LLM users not only know how to use the tools but also why and when to use them.
  • Critical Project Evaluation: Before launching an LLM-based project, assess its need and potential impact.
  • Focus on the End-User: Prioritize features that deliver tangible value to users.

Conclusion

LLMs offer incredible possibilities to transform how we work. However, we must be vigilant not to confuse speed with real productivity. By adopting a thoughtful approach and remaining focused on long-term impact, we can leverage these powerful tools to their fullest potential.

Let's discuss your project in 15 minutes.

LLMs productivity automation tech development business strategy
Deepthix newsletter · 100% AI · every Monday 8am

An AI agent reads tech for you.

Our AI agent scans ~200 sources per week and ships the best articles to your inbox Monday 8am. Free. One click to unsubscribe.

Visit the newsletter page →

Want to automate your operations?

Let's talk about your project in 15 minutes.

Book a call