As those who have read this column over time understand, I have a soapbox that involves authors, whether academics or consultants, pandering to management rather than teaching them. Sadly, Age of Invisible Machines.

---

The second, and larger issue was mentioned up top. Inventors have a habit, from long before Alfred Nobel, of ignoring the consequences of their inventions. The excuse is the same as scientists often give, that it’s not up to them to decide on the used and societal impact, they’re just discovering and inventing things. While that is true for theoretical science, it’s far past time for technologists focused on applications that directly impact society to give up that attempt to absolve themselves from societal impact.

The ethical AI movement is only an extension of regular movements in society, movements that try to understand how change impacts those societies and to do it from the beginning. Any good programmer looks at system issues from the design phase. Waiting until debugging is too late to create an effective system. Artificial intelligence will clearly impact society in major ways. It will redefine who can work and how society must address a change in the definition of work. It ties into the overvaluing of stocks because of the promise of solutions, in the lack of understanding of most people in what those solutions mean, and a real understanding, among a very few, of what that means.

Society is in a new Gilded Age. The first one led to regulations and protections that have been gutted over the last forty years. This new one has even more challenges and dangers than those that were created in the industrial revolutions. Nobody should be talking about systems with such an enormous potential impact on society without talking about those impacts. This book takes an outdated approach of ignoring society while pushing for major societal disruptions. That means I cannot recommend this book.

The emergence in the last week of a particularly effective voice synthesis machine learning model called VALL-E has prompted a new wave of concern over the possibility of deepfake voices made quick and easy — quickfakes, if you will. But VALL-E is more iterative than breakthrough, and the capabilities aren’t so new as you might think. Whether that means you should be more or less worried is up to you.

Voice replication has been a subject of intense research for years, and the results have been good enough to power plenty of startups, like WellSaid, Papercup and Respeecher. The latter is even being used to create authorized voice reproductions of actors like James Earl Jones. Yes: from now on Darth Vader will be AI generated.

VALL-E, posted on GitHub by its creators at Microsoft last week, is a “neural codec language model” that uses a different approach to rendering voices than many before it. Its larger training corpus and some new methods allow it to create “high-quality personalized speech” using just three seconds of audio from a target speaker.

This blog post was co-authored with Guy Eyal, an NLP team leader at Gong.

TL;DR: In 2022, large models achieved state-of-the-art results in various tasks and domains. A significant breakthrough in natural language processing (NLP) was achieved when models were trained to align with user intent and human preferences, leading to improved generation quality. Looking ahead to 2023, we can expect to see new methods to improve the alignment process (such as reinforcement learning with AI feedback), the development of automatic metrics for understanding alignment effectiveness, and the emergence of personalized aligned models, even in an online manner. There may also be a focus on addressing factuality issues as well as developing open-source tools and specialized compute resources to allow the industrial scale of aligned models. In addition to NLP, there will likely be progress in other modalities such as audio processing, computer vision, and robotics, and the development of multimodal models.

2022 was an excellent year for machine learning, with numerous large language models (LLMs) published and achieving state-of-the-art results across various benchmarks. These LLMs demonstrated their superior performance through few-shot learning, surpassing smaller models that had been fine-tuned on the same tasks [1–3]. This has the potential to reduce the need for specialized, in-domain datasets. Techniques like Chain of Thoughts [4] and Self Consistency [5] also helped to improve the reasoning capabilities of LLMs, leading to significant gains on reasoning benchmarks.

Generative AI and creative innovation go hand in hand that capture people’s attention in 2023. Here in this article, we are going to it explain how.