Ch 1 Of my New Book on AI
Submitted for your Review
I am collaborating with Chris Kieff, on a book that has the placekeeping title of Thriving in the AI Era. (The subtitle is yet to come). It is targeted at business decision-makers and we hope that when they have completed it they will understand how to cope with the fundamental changes we all face and to not only survive but prevail.
I’m sharing it with you here because we are big believers in crowd wisdom. I have written seven previous books asking my online readers for feedback on early drafts of each chapter, asking for edits, ideas, content, or links that will always help make my books better, one chapter at a time.
So, please tell me what you think of this early version. Please be candid. I would rather get my feelings hurt at this early phase and have a book that people want to read and buy when the book is completed several months from now.
Contact me any way you wish. My old-fashioned email is shel@shelisrael.com
Chapter 1
A Digital Holy Grail
“Some things must be believed before they can be seen.”
– Guy Kawasaki, marketing guru
November 30, 2022, was a slow news day. The New York Times’ top stories included reports of Congress trying to head off another rail strike, more Russian atrocities in Ukraine, and China, once again, violently suppressing protesters.
Not one traditional media organization paid attention to a modestly worded announcement by OpenAI, a-little known San Francisco-based startup who, on that day, released an upgrade of something even less known called ChatGPT.
In the article, OpenAI asserted that ChatGPT, “makes it possible to answer follow-up questions, admit mistakes, challenge incorrect premises, and reject inappropriate requests."
While big media yawned, social media ignited. ARS Technica, a website community for developers, reported that ChatGPT could not only correct coding mistakes, and generate its own code, but it was also reasonably good at writing poetry, expounding on a philosophical debate of whether or not a hot dog in a bun could be classified as a sandwich, and explaining the worst-case time complexity of the bubble sort algorithm in the style of a fast-talking wise guy from a 1940s movie.
The overall impression we got from cumulative online conversations was that some sort of digital Holy Grail had been attained, and in reflection, that is precisely what had happened.
Less than a week later, ChatGPT had been downloaded more than one million times. Sixty days later, 100 million people had downloaded the beta product, marking the shortest period for any digital product to ever reach that milestone. Not only were people visiting the site to check out something being talked about, but they were also using it and seeing value.
In January 2o23, Reuters reported that ChatGPT was receiving 13 million unique visitors daily. In February, Similar Web upped the estimate to 25 million. By June, it was estimated 200 million people had visited it, putting it on track to hit over 9 billion by its first birthday.
(We should note that since then, Threads, a new social media platform from Meta, which attracted 100 million users in just five days, records seem to get shattered faster and faster.)
The Last Laugh
ChatGPT might go down in history as one of the great underdog stories like Jack and the Beanstalk, Rocky, Harry Potter, or Slumdog Millionaire. In 2020, Sam Altman was speaking at a Silicon Valley event, when he asked about his financial strategy.
“I have no idea. We have never made any revenue. We have no plans to make revenue, and we have no idea how we will one day make revenue,” he told attendees who laughed heartily.
He went on to say that when ChatGPT is built, they would ask it how to monetize, which may or may not have been what actually happened.
“We have never made any money so far, and we have no idea how we will ever make any money. Right now, we’re thinking that when it is ready, we will ask ChatGPT, and hopefully, it will have the answer,” he told the audience, which laughed heartily.
Three years later, OpenAI forecast its revenues would top $1 billion in 2024, and NBC speculated that as an investor, Microsoft’s top line would increase by $40 billion.
It looks like Altman will have the last laugh.
Previously, he earned a stellar reputation as a founder of Y-Combinator, a venture group founded in 2005, that focused on Unicorns: start-ups that achieved billion-dollar valuations while still private.
Under Altman’s tenure, Y-Combinator produced more unicorns than any investment group with home-run offerings such as Airbnb, DoorDash, Dropbox, Instacart, and Stripe.
In 2015, Altman left to collaborate with other industry luminaries including Reid Hoffman, Peter Thiel, Greg Brockman, Jessica Livingston, and Elon Musk to found OpenAI with a commitment to freely sharing with other organizations whatever it developed.
As of this writing, co-founder Brockman serves as Head of Nonprofit & Strategic Initiatives at OpenAI and is generally perceived as the guy running the company. He is co-founder of Stripe, a wildly successful online payments company, and former mayor of Mountain View, California in Silicon Valley.
When it began operations, Altman assembled an accomplished tech team, recognized as among the best in AI. In 2021, OpenAI launched its first product, DALL-E, an AI program that can generate digital images from natural language descriptions. While there are several competing services, DALL-E is recognized as among the most successful with over 3 million users as of this writing.
Metaphoric Gun
To our thinking, the launch of ChatGPT is a metaphoric starting gun represents that has triggered the entry of a new AI-driven Era. It is reminiscent of the embattled farmer who stood on a bridge and fired a shot heard around the world starting the American Revolution. This AI revolution fulfills the vision of a small handful of computer scientists who first envisioned what was happening now in the 1950s.
ChatGPT was far from the first AI technology for end users, and the category of AI Search is only a tiny sliver of what has already come, not to mention so much more that will arrive over the next few years.
But when we look back a decade from today, we will see ChatGPT served as the inflection point when businesses, partners, customers, and employees all collectively realized that AI will be an important tool in nearly everything and started to adjust course.
In fact, we have already started adjusting course, because of AI. We use it in driving maps that predict hazards, notices shift in traffic patterns, and learn the routing preferences of each individual driver.
AI is also the secret sauce in our virtual assistants, chatbots, recommendation engines, and social networks.
So why, some may wonder, was this product so wildly popular compared with all those other AI-powered offerings? We asked ChatGPT, and it gave us five good reasons. (We have edited its precise words, for brevity.)
Humanlike text. GPT communicates in a conversational and personal style for each user. Its technology understands context and generates creative and varied responses.
Versatility. ChatGPT performs a wide variety of tasks, ranging from answering questions and creating written content to tutoring on various subjects and even simulating characters for video games.
Reputation. OpenAI is known for its cutting-edge AI research. The release of ChatGPT was more trustingly embraced by so many because of the company’s reputation.
Potential. People could see that ChatGPT held much promise for significant new applications and advancements. We look forward to the day when people will just talk with it, rather than have to use text, which will feel antiquated in future years, like rotary dials on old phones.
Accessibility. Unlike AI used exclusively within enterprises, ChatGPT was broadly available to everyone with a connected computer.
We would also add two additional points:
Timing. After 50 years of refinement, the world seems ready to embrace a technology that developed slowly but now seemed to have come of age. What may have seemed freaky a few years ago has already become part of the newest normal, because earlier, more limited chatbots such as Amazon Alexa, Google Assistant, and Apple Siri started us talking with technology. Each is valued, but many users are accustomed to and sometimes frustrated by their limitations.
Ease of Use. Anyone who has used Google can figure out any GPT product in minutes. The conversational capabilities make it easier than some people experience when interacting with co-workers.
Jargon Translations
If you already are familiar with AI jargon, we suggest you just skip this section and go down to the next bold-faced heading. If you are just getting started, you may have already heard these terms, but need clearer definitions, then we hope this section will help.
The term GPT is short for Generative Pre-training Transformer, which enables people to speak naturally and interactively with a software algorithm, which uses Natural Language Processing (NLP) which allows you to converse as if you were interacting with another person, thus eliminating any need for coding. GPT can create or summarize large quantities of text, generate code, write poetry, or even create impressive art, design, or music.
Two additional terms, you are likely to encounter are Large Language Models or LLMs, trained on massive datasets with billions, perhaps trillions of parameters. The software employs self-supervised learning to understand the relationships between words and phrases as well as to correct corrects grammar. It also can perform many complex tasks, such as generating text, translating languages, and producing creative content.
The products we have discussed so far are all LLMs. While both ChatGPT and Microsoft Bing are built on GPT, Google Bard is built using Google's Pathways Language Model (PaLM), which the company says, is a newer and more powerful language model than GPT.
There are also Small Language Models, (SLMs) that operate using less than 100 million parameters them faster and less expensive to train and deploy than LLMs which usually incorporate billions or even trillions of parameters. They are used in a variety ways with online learning tools such as Blacboard Learn, VanvasLMS, Sakai, and D2L Brightspace were among the most popular when we wrote this Chapter in mid-2023.
LLMs have a great number of advantages over SLMs, except in one critical area. As of this writing, SLMS are more accurate and are used in areas such as medical diagnoses and financial trading, where accuracy is a critical factor.
The Big Three
ChatGPT, Google Bard, and Microsoft Bing are without question the Big Three of Search AI. As we write, ChatGPT is the largest, followed, by Bard, and then Big. By the time you read this, the order may or may not have changed many times over. All three have many similarities and differences and perform tasks that are unprecedented in the history of digital technology.
It could be argued that the battle is really one between a Big Two. Microsoft has invested $10 billion in OpenAI and has built Bing on GPT technology from OpenAI, but currently, they perform differently. While ChatGPT is generally perceived as having superior capabilities to Bing, Microsoft enjoys a user base estimated at 2.5 billion, surpassed in the software industry only by Google with 4 billion. As we see it, it will be a battle between giants that users are bound to win either way.
Over the next 10 years according to questions we asked both Bard and Bing, users can expect both AI Search platforms to provide:
· Greater contextual understanding: Future versions will understand understand and respond to the context of a conversation in more meaningful ways and conversations started in one session can carry over into later indefinitely.
· Closer Personalization. Future versions will get to know each individual as people evolve in personal and business relationships. The AI will be able to ask each of us how your vacation, business presentation, or history test went, and then figure out how they can become more helpful in the future. They will understand your way of talking where you use slang or curse words, what sports teams you follow, and your preferences in art, literature, reading, music, sports teams, exercise, etc., and will grow in their abilities to recommend what you like as it filters out what doesn’t interest you.
· Emotional understanding. Future versions might be more capable of understanding human emotions expressed in text and respond in a way that is perceived as emotionally appropriate and empathetic.
· Active learning. GPT might be able to ask clarifying questions when given ambiguous queries, helping it to provide more accurate and helpful responses.
· Fact-checking. AI might be integrated with real-time databases to check facts or update information when previous responses are no longer current.
· Join the Internet of Things (IoT). The AI will interact with and control smart devices and manage your connected home or workplace.
· Improved Security. Cybersecurity is a perpetually escalating battle. ChatGPT, Bard, and Bing Bard and sophisticated and nuanced ways to detect may be able to prevent harmful instructions, misinformation, inappropriate content, or breeches in confidentiality,
· Personal Assistance. As AI gets smarter, it will know more of your personal needs and serve users increasingly in most aspects of our lives.
· Universal Translators. Real-time translation between two languages is already possible with AI and is likely to improve. You will, be able to communicate in your language to people who receive it in their own language and respond accordingly. You will receive it in whatever it is you speak. This will happen first by typing, but eventually, it will be conducted by voice.
· Other. Somewhere in the world, some right now have a great idea that no one else has had yet. That person will do something with AI that is remarkable and transformational. Perhaps we’ll write a book about it.
Global Micro-Markets
ChatGPT, Bard, and Bing are not just consumer-facing platforms, thousands of businesses are being built upon them, and we would guess, most are tiny and modestly funded, but these AI-powered engines give these micro-startups access to the global micro market in a broad array of categories ranging from jewelry makers, online teaching, software.
Using any of our Big Three AI platforms, or anything new that is yet to come, these micro-companies will have access to global markets and be able to create brand awareness to a level that has previously not been possible. For example, AI will allow their websites to be multilingual. One-person operations might soon provide responsive global customer support and each micro-company will have greatly improved opportunities for growth.
Bringing Up Baby
There are many analogies to what AI is like, and none of them felt quite right until Chris came up with the idea that it is very much like a baby, to which Shel added the analogy that it is like a baby with Einstein Syndrome, a term used to characterize a child who is slow to start taking but is simultaneously gifted in analytical thought.
When any AI platform comes into being, it knows absolutely nothing and has no answer to any query or prompt. But it has an enormous and insatiable ability to collect and retain data.
Babies, learn what they are taught, just like AI. You need to protect what babies learn and teach them ethical behavior, just like AI. Sometimes, like babies, they get answers wrong, but in the case of AI, that is always because they were not properly taught.
Additionally, parents instinctively know that babies need to be protected from all sorts of known and unknown factors that are constantly changing—just like AI.
Babies grow up in most cases to be very much how their parents raised them to be and so does AI.
Our point is that. If you are mindful of what you teach your AI, you, your partners, customers, employees, and prospects will benefit. If you choose another direction, you may suffer unanticipated and unwanted consequences.
What they become depends on what we humans teach them.
In the coming chapters, we will share with you what research we have collected on how AI will improve your efficiency, relationships, brand security, and environment while posing threats to some of those very same assets.
There is no doubt that AI is going to change your business, and this book is intended to equip business decision-makers like you with the information and insights into how this AI era will change how you operate, starting with a need to be increasingly agile
.