5,99 €
Mehr erfahren.
DISCLAIMER
This book does not in any capacity mean to replace the original book but to serve as a vast summary of the original book.
Summary of Co-Intelligence by Ethan Mollick: Living and Working with AI
IN THIS SUMMARIZED BOOK, YOU WILL GET:
- Chapter provides an astute outline of the main contents.
- Fast & simple understanding of the content analysis.
- Exceptionally summarized content that you may skip in the original book
Das E-Book können Sie in Legimi-Apps oder einer beliebigen App lesen, die das folgende Format unterstützen:
Ähnliche
Summary of
Co-Intelligence
A
Summary of Ethan Mollick’s book
Living and Working with AI
GP SUMMARY
Summary of Co-Intelligence by Ethan Mollick: Living and Working with AI
By GP SUMMARY© 2024, GP SUMMARY.
All rights reserved.
Author: GP SUMMARY
Contact: [email protected]
Cover, illustration: GP SUMMARY
Editing, proofreading: GP SUMMARY
Other collaborators: GP SUMMARY
NOTE TO READERS
This is an unofficial summary & analysis of Ethan Mollick’s “Co-Intelligence: Living and Working with AI” designed to enrich your reading experience.
DISCLAIMER
The contents of the summary are not intended to replace the original book. It is meant as a supplement to enhance the reader's understanding. The contents within can neither be stored electronically, transferred, nor kept in a database. Neither part nor full can the document be copied, scanned, faxed, or retained without the approval from the publisher or creator.
Limit of Liability
This eBook is licensed for your personal enjoyment only. This eBook may not be resold or given away to other people. If you are reading this book and did not purchase it, or it was not purchased for your use only, then please purchase your own copy. You agree to accept all risks of using the information presented inside this book.
Copyright 2024. All rights reserved.
INTRODUCTION
THREE SLEEPLESS NIGHTS
The cost of getting to know AI is at least three sleepless nights. After a few hours of using generative AI systems, it becomes clear that Large Language Models (LLMs) act more like a person than expected. This realization leads to questions about the future of work and education.
An academic studying innovation, an academic studying AI, has been involved in work on the applications of AI, especially for learning. AI research has always seemed to be on the edge of a massive breakthrough, but most practical uses have advanced slowly. During this time, the author kept experimenting with AI tools, including OpenAI's GPT models, figuring out ways to incorporate them into their work and assigning students to use AI in class.
Just after the release of ChatGPT in November 2022, the author demonstrated the new tool to his undergraduate entrepreneurship class, demonstrating how AI can help generate ideas, write business plans, turn those plans into poems, and generally fill the role of company cofounder. By the end of the class, one of the students had created a working demo for his entrepreneurship project, which led to venture capital scouts reaching out to him by the end of the next day.
The author recognized that he wasn't immune to these changes and decided to get hands-on and put the AI through its paces. He built simulations that teach business skills, such as negotiation, using thousands of lines of code.
Every week, AI serves up new miracles or concerning developments, such as creating beautiful images, coding in Python, and discovering something remarkably close to an alien co-intelligence that can interact well with humans without being human or sentient.
General Purpose Technology (GPT) refers to once-in-a-generation technologies that impact every industry and aspect of life. AI, or generative AI, is a significant advancement that has taken many decades to develop and become useful. For example, the internet took nearly three decades to achieve general use in the 1990s, while computers took decades to become useful. Large Language Models (LLMs) have become increasingly capable and have been adopted by consumers quickly.
AI's impact on work and education may be less than previous GPTs, as it works as a co-intelligence, augmenting or potentially replacing human thinking. Early studies have found that AI can lead to a 20 to 80 percent improvement in productivity across various job types. However, economists have struggled to show a real long-term productivity impact of computers and the internet over the past twenty years.
GPTs also impact every aspect of our lives, such as teaching, entertainment, and self-expression. Schools are grappling with the future of writing, AI tutors may change education, and AI-driven entertainment is allowing personalized stories. Misinformation is already flowing through social networks, making things increasingly strange.
AI has been a topic of discussion for many, with many believing it to be the spark of a new form of intelligence. AI has surpassed the Turing and Lovelace Tests, acing our hardest exams, and maximizing human creativity and sentience. However, the exact reasons behind these abilities remain unclear. As a professor at Wharton, the author has been an influential voice on the implications of AI, publishing some of the first research on AI in education and business. They regularly speak with organizations, companies, and government agencies to understand the world we are creating and keep up with the rapidly evolving field of AI research. The author aims to take readers on a tour of AI as a new co-intelligence, with all the ambiguity that the term implies. To understand what AI means, we need to start with the technology of Large Language Models, discuss how AI can change our lives by acting as a coworker, teacher, expert, and companion, and consider the implications of thinking together with an alien mind.
PART I
CREATING ALIEN MINDS
AI has been a subject of fascination for centuries, with its origins dating back to the invention of the first mechanical chess computer in 1770. The first machine, known as the Mechanical Turk, was a chessboard that beat Ben Franklin and Napoleon in chess matches. However, the idea of machines thinking led to the development of artificial intelligence (AI) in the 1950s through the creation of the jury-rigged mechanical mouse called Theseus and the imitation game by Alan Turing.
The latest AI boom began in the 2010s with the promise of using machine learning techniques for data analysis and prediction. Many applications used supervised learning, which required labeled data to learn from. This phase of AI was the domain of larger organizations that had vast amounts of data, using these tools as powerful prediction systems.
Predictive AI technologies, such as those found at retail giant Amazon, revolutionized many back-office functions, such as managing customer service and running supply chains. Amazon integrated AI into forecasting demand, optimizing warehouse layouts, and delivering goods. It also intelligently organized and rearranged shelves based on real-time demand data, making popular products easily accessible for quick shipping.
However, these AI systems had limitations, such as struggling with predicting "unknown unknowns," difficulty with data not yet encountered through supervised learning, and limited ability to understand and generate text coherently and context-awarely. While these uses of AI are still important today, they were not something most people directly saw or noticed in their daily lives.
The 2017 paper "Attention Is All You Need" by Google researchers introduced a significant shift in AI, particularly in how computers understand and process human language. The Transformer architecture, which uses an attention mechanism, allows AI to concentrate on the most relevant parts of a text, making it easier for the AI to understand and work with language in a more human-like way. This has led to the development of Large Language Models (LLMs) that can generate contextually rich content, showcasing the evolution of machine comprehension and expression.
LLMs are trained on a massive amount of text from various sources, such as websites, books, and other digital documents. This is called pretraining, and unlike earlier forms of AI, it is unsupervised, meaning the AI doesn't need carefully labeled data. Instead, AI learns to recognize patterns, structures, and context in human language through analyzing examples. LLMs can create a model that emulates how humans communicate through written text using a vast number of adjustable parameters (weights).
The original ChatGPT had 175 billion weights, encoding the connection between words and parts of words. Over time, the weights become more organized and accurate, and the apprentice chef transforms into a master chef. When given a prompt, the master chef artfully selects the right ingredients from their vast repertoire and consults their refined spice rack to ensure the perfect balance of flavors. In this analogous way, AI creates humanlike written text that is engaging, informative, and relevant to the topic at hand.
