artificial intelligence
Sergey Tarasov -stock.adobe.com
AI Conundrum, Caleb Briggs, and REX BRIGGS (Mit Press, 2024) are very good books with normal flaws that do not reduce technology. I recently sent a book from a publisher, but the book was so bad that I couldn’t put two or more chapters. I reviewed a problem book, but it was not enough to create an article around. Shortly thereafter, I saw this book in a new section of my local library. I love the library and this book washed away the bad taste from others.
ai undrum
Mit press
This book aims for non -technical management that wants to better understand the risks and rewards of AI in business environments. It does it well by starting one of the best non -technologies explanations of neural networks below the current general solution. Chapter 2 and Chapter 3 are very easy to access, and you need a little high school mathematics.
Mathematics debate brings one of the major issues with large language models (LLM) and suspicious results of mathematics itself. The system is essentially statistical, not understanding higher mathematics, but acquiring information. The questions explained in deep learning and mathematics are that if you have a hammer, everything looks like a nail. Even the simplest program does not do its own mathematics. Call the module. I wrote about what some front -ends and back -end expert systems were, so I tried to clean up the data. The neural network is excellent, but only part of the final system. When a programmer can teach the network when calling a subroutine, it helps.
In the next chapter, we will provide several wonderful case studies, define three foundation stones: accuracy, input control, and theoretical basis (transparency), and shift to an excellent discussion on business risks in Chapter 6. 。
The second section of the book starts with Chapter 7 and several case studies. If it seems to be a repeated theme with people who focus only on AI, the historical growth of the solution skips business intelligence (BI). What is presented as a new AI has been in the BI for decades, including normative maintenance. Yes, AI has some improvement, but business management needs to do ROI analysis to check if the result is needed. Please refer to my mentions of the above “If you only have a hammer” and discussions on the accuracy of the author, not only between AI models, but also in the entire technology options. If the non -AI solution provides enough results at a much lower cost, ignore the trend and expenditure. If the accuracy of AI is needed, it is no longer a trend.
The main problem is repeatedly mentioned in the review. They filled the problem of unemployment in the back 3 of the book and rejected it as a real problem. My favorite line is that they are “assuming that they will tell some of these savings to their customers” … What are the large companies that show their support for the past few decades? There was no. As the article on this site pointed out in 2023, the goal is to keep transferring more money in other classes to the upper class.
As AI becomes wider, there is a big social change. This is not an industrial revolution. The impact will be much larger. Through international agreements from cities, the government has already dealt with its impacts and potentially support great profits.
As mentioned earlier, the negative is not a new one for a business book about AI. As a whole, this is one of the best management books about AI that I have read in a few years. Based on how AI works, the risks and rewards of business -based latest technology, case studies indicating actual implementation, and great frameworks for business evaluation using AI It is strongly recommended for those who want to understand risks and rewards better.
