The Wrong Side of Right: Navigating the High Cost of AI Compute



Sep 13, 20233 min read


The Wrong Side of Right: Navigating the High Cost of AI Compute


In the world of entrepreneurship and artificial intelligence (AI), there are common mistakes and challenges that need to be addressed to achieve the best outcomes. This article explores the concept of "the wrong side of right" and delves into the high cost of AI compute. By understanding these issues and adopting the right mindset and strategies, entrepreneurs can navigate the complexities of AI and position themselves for success.

The Wrong Side of Right:

One big mistake that people often make is their obsession with proving themselves right. This mindset creates a narrow focus on defending their beliefs and dismissing anything that contradicts them. Instead of striving to be right, entrepreneurs should aim to be less wrong. By acknowledging their fallibility and embracing a mindset of continuous learning and improvement, they can open themselves up to better outcomes. This shift from ego-driven certainty to a humble pursuit of knowledge can lead to greater success in business.

The High Cost of AI Compute:

In the realm of AI, the cost of training and inference is a significant factor driving the industry. The demand for compute resources far outweighs the supply, with some companies spending a substantial portion of their capital on compute. The complexity of AI algorithms, particularly in models like GPT-3, makes AI infrastructure expensive. The computational requirements for training and inference are immense, with the need for specialized chips and optimized hardware. Memory requirements and the availability of software optimizations further complicate the cost and performance of AI compute.

Actionable Advice:

  • 1. Embrace a Growth Mindset: Instead of trying to prove yourself right all the time, adopt a growth mindset that emphasizes learning, growth, and continuous improvement. This mindset allows you to be open to new ideas, perspectives, and evidence that can lead to better outcomes.
  • 2. Prioritize Efficiency and Optimization: When dealing with AI compute, it's crucial to optimize your models and infrastructure to maximize efficiency. This includes selecting the right size of models for your specific use case, leveraging specialized chips and hardware, and exploring software optimizations. By prioritizing efficiency, you can reduce costs and improve performance.
  • 3. Consider Cloud Solutions: For many startups, building their own AI infrastructure may not be necessary or cost-effective. Hosted model services and cloud providers offer accessible and scalable solutions for rapid prototyping and product-market fit testing. Leveraging these services can save time, resources, and the need to manage underlying infrastructure.


Being on the wrong side of right and fixating on proving oneself right can hinder progress and limit success. Instead, entrepreneurs should focus on being less wrong and embracing a growth mindset that promotes continuous learning and improvement. In the realm of AI, understanding the high cost of compute and adopting strategies for efficiency, optimization, and leveraging cloud solutions can help navigate the complexities of AI infrastructure. By combining the right mindset with actionable strategies, entrepreneurs can position themselves for success in the fast-growing world of AI.


  1. "The Wrong Side of Right", (Glasp)
  2. "Navigating the High Cost of AI Compute | Andreessen Horowitz", (Glasp)

You might also like:

Want to hatch new ideas?

Glasp AI allows you to hatch new ideas based on your curated content. Let's curate and create with Glasp AI :)