← Retour au blog
tech 16 July 2026

The LLM Critics Are Right. I Use LLMs Anyway.

Language models (LLMs) are divisive. Yet, despite justified criticism, many professionals still use them. Why this dissonance?

Article inspired by the original source
The LLM Critics Are Right. I Use LLMs Anyway ↗ www.theocharis.dev

Introduction: The LLM Dissonance

Language models (LLMs) have become a staple in the tech world. Yet, their use attracts sharp and often justified criticism. So why do we continue to use them? This question came to mind during the Local-First conference in Berlin, where the paradox was palpable. Armin Ronacher, a recognized software engineer, shared his experience with Pi.dev, an open-source coding agent harness. Despite the criticisms, he continues to explore the potential of LLMs.

Legitimate Criticisms of LLMs

LLMs are accused of several shortcomings:

  1. Copyrighted Content: LLMs are often trained on copyrighted data. This raises major legal issues.
  2. Environmental Impact: Training these models requires massive computing power, increasing their carbon footprint.
  3. Bias and Misinformation: Models can replicate and amplify biases present in training data.

These concerns are not unfounded. For example, a 2023 study revealed that the carbon footprint of LLMs could surpass that of many countries if their usage continues at this pace.

Why Use LLMs Anyway?

Despite the criticisms, LLMs offer undeniable advantages for developers and tech companies:

  1. Efficient Automation: LLMs enable the automation of complex tasks, freeing up time for higher-value activities.
  2. Accelerated Development: They facilitate rapid prototyping of applications and solutions, reducing time-to-market.
  3. Accessibility and Personalization: Thanks to LLMs, even small businesses can access advanced technologies and customize their solutions.

A concrete example is the adoption of LLMs in customer support tools, enabling automated responses while reducing operational costs.

Striking a Balance

The key is to find a balance between using LLMs and managing their drawbacks. Here are some suggestions:

  • Transparency and Ethics: Adopt transparent and ethical practices to minimize biases and respect copyright.
  • Energy Optimization: Invest in greener infrastructures and optimize model training to reduce environmental impact.
  • Continuous Education: Raise awareness among developers and decision-makers about risks and best practices.

Conclusion

The dissonance around LLMs is real, but it doesn't mean their use is impossible. By adopting a thoughtful and responsible approach, it is possible to leverage their potential while mitigating their negative impacts.

Let's discuss your project in 15 minutes.

LLM IA automatisation développement éthique
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