Unleashing the Power of Unread Books and Examining Emergent Abilities in Large Language Models



Sep 21, 20233 min read


Unleashing the Power of Unread Books and Examining Emergent Abilities in Large Language Models


In a world filled with endless sources of information, it's easy to become overwhelmed by the sheer volume of knowledge available to us. People often find themselves acquiring books that they never have the chance to read, resulting in what the Japanese call "tsundoku" - a collection of unread books. However, contrary to popular belief, these unread books hold immense value and can be as powerful as the ones we have actually read.

The Antilibrary Concept:

The concept of an antilibrary, popularized by Nassim Nicholas Taleb in his book "The Black Swan," challenges the notion that a personal library is merely an ego-boosting display. Instead, an antilibrary is a carefully curated collection of unread books that serve as a research tool. By collecting books that pique our curiosity, we create a resourceful library that encourages exploration and discovery. In a sense, an antilibrary becomes a humble relationship with knowledge.

Embracing the Unknown:

Scottish scientist James Clerk Maxwell once stated that "thoroughly conscious ignorance is the prelude to every real advance in science." This sentiment holds true not only in scientific endeavors but also in our personal growth. To gain new perspectives and expand our horizons, we must step out of our comfort zones and embrace the unknown. An antilibrary reminds us of the vast realms of knowledge waiting to be explored, encouraging us to venture beyond our familiar realms.

The Power of Emergence:

The idea of emergence, as articulated by Nobel laureate Philip Anderson, highlights how quantitative changes in a system can lead to new behaviors. This concept has been observed in various fields, including physics, biology, economics, and computer science. When it comes to large language models, emergence is a crucial aspect to consider. Scaling up these models can result in the emergence of new abilities that were absent in smaller models.

Unpredictable Surges of Performance:

In the realm of large language models, the behavior of these models often follows predictable patterns as they scale up. However, there are instances where performance surges unexpectedly, surpassing random performance levels. These emergent abilities are of great scientific interest as they shed light on the capabilities and limitations of these models. They also motivate further research and exploration in the field of language models.

Creating an Actionable Antilibrary:

To make the most of our antilibrary and optimize our learning potential, here are three actionable tips:

  • 1. Take Detailed Notes: When reading a book, make notes of all relevant references mentioned by the author. By doing so, you create a comprehensive list of sources that can further enhance your understanding of the subject matter. These notes serve as a valuable resource for future exploration and research.
  • 2. Seek Recommendations: Knowledge is a process, not a possession. Engage with fellow readers and ask for recommendations. By seeking diverse perspectives and exploring different genres and topics, you enrich your antilibrary and expose yourself to a broader range of ideas.
  • 3. Stay Within Your Means: Building an antilibrary is an investment in yourself, but it's important to do so within your means. Acquiring books should not become a financial burden. Consider borrowing from libraries, exploring second-hand bookstores, or even utilizing digital resources to build your collection responsibly.


Unread books and large language models share a common thread - they both possess untapped potential. Building an antilibrary allows us to embrace the unknown, curate a highly personal collection, and tap into the power of undiscovered knowledge. Similarly, examining emergent abilities in large language models sheds light on the remarkable capabilities these models possess. By taking actionable steps like note-taking, seeking recommendations, and staying within our means, we can unlock the true value of our unread books and leverage the potential of large language models for further exploration and discovery.


  1. "Building an antilibrary: the power of unread books", https://nesslabs.com/antilibrary (Glasp)
  2. "Examining Emergent Abilities in Large Language Models", https://hai.stanford.edu/news/examining-emergent-abilities-large-language-models (Glasp)

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 :)