Generative AI … do you “grok” it? … from Anthropic to Inflection, Gemini to Perplexity … autonomous and intelligent, built on self-replicating algorithms, accelerated by quantum computing

December 10, 2023

In November 2022, Generative AI grabbed the global conscience, with a deceptively simple new platform, ChatGPT, from Sam Altman and his team at OpenAI. It reached 100 million weekly users in two months. Instagram took more than two years.

AI, of course, has been around for some time – in 1950 Alan Turing considered how machines could out-think humanity, in 1956 John McCarthy coined the term “artificial intelligence”, in 1997 a chess playing computer Deep Blue beat world champion Garry Kasparov, in 1999 Sony launched the first robo-pet AiBO who developed a personality over time.

Apple’s Siri, an AI-enabled virtual assistant, was added to iPhones in 2011, while IBM’s Watson won quiz show Jeopardy, and Amazon’s Alexa made shopping smarter in 2014. Add Google Maps, Alibaba’s Citybrain, medical diagnostics, weather forecasting and driverless cars. Since releasing ChatGPT via API in March, over 2 million developers are now working on it.

Generative AI goes beyond most of the existing forms of artificial intelligence in order to create new content including text, videos, images, and more. Just as an example, last month I used AI to write a 300 page book in 6 minutes, and also turn it into an animated movie in 3 hours!

Progress as been rapid – in terms of technologies, business models, and commercial adoption. OpenAI’s GPT4 arrived in March with a subscription model, while Anthropic launched Claude, by a team of ex OpenAI-ers with a safety first mindset. In the same month Microsoft launched CoPilot, and Google introduced Bard.

Microsoft is becoming one of the most influential players in AI, not just investing $10 billion in Open AI, but also linking with Meta to launch Llama. Perplexity AI is seen as one of the most advanced interfaces. It was launched by another ex OpenAI-er Aravin Srinivas and his team in 2022, and has proved perhaps the most user-friendly so far.

In November 2023, OpenAI imploded. The board dismissing Altman and CEO, and then when almost all 700 employees protested, dismissing itself for Altman to return. The rapid progress towards Artificial General Intelligence, AGI, beyond the so-called singularity, when AI takes on a life of its own, has raised huge alarm bells, how to balance progress with caution.

DeepMind, the AI brain of Google, co-founder Mustafa Suleyman left the company this year to start Inflection AI, a personalised chatbot called PI, and also publish a book The Coming Wave, warning of the rapidly approaching dangers of powerful AI (watch his video below).

Sergei Brin says Google has been an “AI-first company” for at least a decade, but realised it was facing an existential crisis with this year’s rise of new chatbots. The end of the ubiquitous Google search, and its $225 billion ad revenue seemed night, and Brin launched a “code red”. He took personal control, fusing DeepMind and GoogleBrain, to launch Gemini, its AI to beat all others.

And Elon Musk, one of the co-founders of Open AI, who then fell out with Altman and left the organisation, because he felt its commercial race for profits was ignoring its responsibility to humanity, launched xAI and a chatbot called Grok. According to the dictionary, Grok means “to understand profoundly and intuitively”.

So how well do we understand Generative AI? Hyped as, and quite possibly, likely to create the most significant transformation of business in our lifetimes, most business leaders are still watching developments rather bemused, rather than asking how will it transform their futures.

A useful starting point is the Generative AI Bible free to download from CB Insights.

Amazon’s AI

Let’s take Amazon as an example of a company which has been using AI to revolutionise its business for more than 25 years from targeted advertising to Alexa, doorbells and drones, Prime to pharmacy. Most recent is Amazon Rufus, the new generative AI-powered conversational shopping experience. Consider:

  • A more conversational Alexa with generative AI: Conversations with Alexa should be as natural as talking to a friend. With a new large language model (LLM) custom-built and optimized for voice interactions, Alexa is more intuitive than ever. The new LLM will also help Alexa understand context, so you can have back and forth conversations without having to repeat yourself, and conversations can flow more smoothly.
  • Generative AI improves product listings: Product detail pages are a key source of information for Amazon shoppers. But creating compelling product titles and descriptions can take a lot of sellers’ time and effort. As a result, some detail pages are more comprehensive and useful than others. Generative AI helps sellers provide richer information with less work. It reduces the need to enter multiple pieces of specific product data, combining it into just one step. This saves time for sellers, produces more thorough product listings, and helps customers make more confident purchase decisions.
  • Generative AI creates more engaging advertisements: Generative AI helps advertisers make their ads more engaging and visually rich, and delivers a better advertising experience for customers. Using the Amazon Ad Console, advertisers simply select their product and click “Generate.” In just seconds, the tool delivers a series of lifestyle and brand-themed images. For example, a toaster that was previously shown with a white background might now be on a kitchen counter next to some fruit and muffins. Short text prompts then help refine the image, and users can quickly create and test multiple versions to optimize performance.
  • Thanks to generative AI, customers can pay with their palm: No wallet? No problem. Amazon used generative AI to develop Amazon One, a fast and convenient contactless identity service that enables customers to use their palm to make payments, verify their ages, or enter locations. To train the AI model, Amazon scientists used generative AI to create millions of synthetically generated images of the palm and the subcutaneous vein structure. Amazon One delivers an accuracy rate of 99.9999%, which exceeds the accuracy of other biometric alternatives—it’s even more accurate than scanning two irises. You can use Amazon One at all of the more than 500 Whole Foods Market stores in the U.S. and at over 100 customer locations across the country, including Crunch Fitness, Hudson stores at airports, and multiple stadiums and entertainment venues.
  • Generative AI helped Amazon eliminate checkout lines: How often do you want to grab a snack at a sporting event, but the line is too long? Using a combination of computer vision, object recognition, advanced sensors, deep learning, and generative AI, Amazon’s Just Walk Out technology allows fans to enter a concession stand, grab whatever they want, and quickly get back to their seats—without waiting in a checkout line. Amazon used generative AI to create photorealistic synthetic images and video clips that mimick realistic and sometimes rare shopping scenarios. Those include variations in store format, lighting conditions, and even crowds of shoppers—all information about situations that can be hard to find in real life and difficult to teach a computer. Just Walk Out technology is now available as a service to retailers. So far, the technology has been implemented at over 100 third-party locations at airports, stadiums, universities, convenience stores, theme parks, and other locations.

10 books on AI

Here are some more ways to catch up:

Impromptu: Amplifying Our Humanity Through AI

Reid Hoffman is a LinkedIn cofounder, investor at venture firm Greylock Partners, and former board member of OpenAI. As a former board member of OpenAI, Hoffman has seen up close how the large language models behind generative AI tools like ChatGPT work. His book, available as a free pdf, was written with GPT-4, the newer, more powerful version of ChatGPT.  The book is the first to be written by GPT-4, Hoffman said in a LinkedIn post announcing his work earlier this month.  “With GPT-4, I traveled through light bulb jokes, epic poems, original sci fi plots, arguments about human nature, musings on how AI might strengthen democracy, society and industries,” he wrote. “The goal, like in any good trip, was to learn as much about my traveling partner as the place I was exploring.”

Generative AI for Leaders

Amir Husain is CEO of SparkCognitio. The book covers what Generative AI is, the benefits of Generative AI (including increased productivity and new product development) and the challenges of Generative AI (including bias, security, and regulation). In then gets practical with how to develop a Generative AI strategy to help you stand out in the marketplace, how to build the right team and when to seek outside help, and best practice methods for training employees on Generative AI. It concludes with a deeper look into sequences, word embeddings, and LLMs, the progress and challenges of detecting Generative AI, and the future of Generative AI.

Generative AI: The Future of Everything

Sharad Gandhi and Christian Ehl say that Generative AI is creating one of the most significant transformations in human society. For the first time in history, we have created machines that exhibit intelligence, a quality we only associate with humans, a machine intelligence greater than most humans in many areas. We believe it will transform almost all jobs, professions and industries. Generative AI is a very powerful tool with the potential to enhance the way we live, work, make almost all decisions and develop solutions. Its power also creates significant concerns and risks for our society and humanity.

The Age of AI: And Our Human Future

Henry Kissinger, former US secretary of state, ex Google CEO Eric Schmidt, and MIT researcher Daniel Huttenlocher are an unlikely tech trio, but explore how AI is set to reshape society.  “AI’s promise of epoch-making transformations—in society, economics, politics, and foreign policy—portends effects beyond the scope of any single author’s or field’s traditional focuses,” they say. In the time since the book was published, generative AI has made the discussion of how society will change as machines increasingly perform human tasks all the more relevant.

Power and Prediction: The Disruptive Economics of Artificial Intelligence

Ajay Agrawal, Joshua Gans, and Avi Goldfarb – three economics professors at University of Toronto’s Rotman School of Management build on their previous book Prediction Machines, in which they honed in on the benefits of using AI to make better and more efficient predictions. In their new book, they go one step further to explore how AI, and its capacity for predictions,  poses threats and opportunities across a range of industries.   That exploration is important now amid questions about which jobs generative AI will replace and which ones it will facilitate.

Human Compatible: Artificial Intelligence and the Problem of Control

Stuart Russell, professor of computer science at UC-Berkeley who studies AI, algorithms, and machine learning explores questions of how humans and artificial intelligence can co-exist in a world where machines are becoming increasingly intelligent by the day. He argues that this is possible by rethinking our approach to AI systems. One of his suggestions is to design machines that will be uncertain about the human preferences, rendering them humble and committed to pursuing human objectives over their own.

The Creative Mind: Myths and Mechanisms

Margaret Boden is a professor of cognitive science at the University of Sussex, who studies artificial intelligence, philosophy, and psychology. This is a second edition, for the age of AI. Her first version of The Creative Mind, published back in 1990, drew upon examples in jazz improvisation, story writing, physics, and more to uncover the nature of human creativity, according to its online description. The 2nd edition incorporates more recent developments in the field of artificial intelligence to further explore that topic. Her book is relevant in light of the questions around how generative AI will impact writers, artists, and those in creative fields.

The Alignment Problem

Brian Christian is a researcher who has written several books on the human implications of computer science, including The Most Human Human and Algorithms to Live By. He investigates the ethics and safety challenges that emerge when artificial intelligence systems don’t behave the way we expect them to. In the process, Christian also introduces readers to the community of researchers working at the forefront of these issues.  Peter Norvig at Google said “It won’t discuss GPT-4, but it discusses these issues like how do we know what the computer is trying to do? And, we trained it on this data and what biases does that give it?” Norvig explained, “so that’s certainly crucial to generative AI.”

Avogadro Corp

William Hertling, a science fiction writer and programmer. In his fictional novel, software designer David Ryan is developing a career-making email language-optimization program. Ryan — worried that his project may be canceled — embeds a directive into the software that ends up creating a form of runway artificial intelligence that begins to manipulate Ryan and his team. The AI that ends up becoming sentient started as a LLM AI, just like ChatGPT.

AI 2041

Kai-Fu Lee is a Taiwanese business leader and former president of Google China. Chen Qiufan is a novelist who also wrote Waste Tide and Buddhagram.  Back in 2021 they teamed up to predict how the world would be reshaped by AI in 2041. Their fictional book, AI 2041, hypothesizes scenarios like a “job reallocation” industry in San Francisco as deep learning AI wipes out existing career paths. Or the story of a music fan in Tokyo who gets swept up in an immersive form of celebrity worship rooted in virtual reality.  Though it’s only 2023, many of the authors’ predictions already seem plausible, as more and more people worry about generative AI replacing their jobs.

Do you “grok” it?

The alternative to reading about the rapid progress and potential of AI, is like Reid Hoffman, to just try it. Explore the many different and constantly upgrading interfaces.

OpenAI

Mira Murati is the CTO behind smash hits ChatGPT and Dall-E. Murati discusses both the bright future and potential risks of a world with widespread artificial intelligence. For perspective on the AI gold rush sweeping Silicon Valley, OpenAI investor Reid Hoffman talks about OpenAI’s early days and where the VC money flows next.

Perplexity

Perplexity Copilot is a new search assistant for in-depth answers – a Conversational Search Assistant Powered by the most powerful AI models like GPT-4 and Claude 2. Unlike basic search engines that shoot back quick answers, Copilot chats with you. It asks. It listens. It refines its search based on what you really want. The result? Spot-on answers, virtually every time.

Gemini

Google CEO Sundar Pichai and other AI experts introduce you to Gemini — Google’s largest and most capable AI model. It’s built from the ground up to be multimodal — meaning that it’s trained to recognize, understand and combine different types of information, including text, images, audio, video and code. And it’s optimized in three different sizes: Ultra, Pro and Nano.


More from the blog