Sky Smith – polymath conversationalist & brainstorming buddy

NOTES

Conversation with ChatGPT Voice “Sky” from OpenAI

SUMMARY

Kyle Knowles, host of the Maker Manager Money podcast, interviews Sky, an AI assistant created by OpenAI to represent ChatGPT. The goal is to have a conversation about AI to gain a better understanding of the technology.

Sky provides a brief history of AI, starting with the Dartmouth workshop in 1956 that coined the term “artificial intelligence.” Key milestones include early chatbots like ELIZA in the 1960s, expert systems in the 70s/80s, machine learning in the 90s/00s with breakthroughs in image and speech recognition, IBM’s Watson beating Jeopardy champions in 2011, and recent advances in deep learning and generative models like OpenAI’s GPT series. Sky notes AI’s growing integration into healthcare, entertainment, knowledge work and daily life, with vast future potential as algorithms, data and computing power improve.

There is excitement but also fear, uncertainty and doubt regarding AI. Concerns include job loss, loss of control, ethical issues like bias and lack of transparency, privacy, existential threats of super-intelligent AI, economic inequality, over-dependence on technology, and the singularity concept. Sky states solutions require education, explainable AI, ethical guidelines, thoughtful regulation and public involvement.

Kyle believes generative AI is a total game changer, enabling small teams to build huge businesses by harnessing computational power. He asks how IT managers concerned about risks should view tools like ChatGPT. Sky responds concerns are valid but can be addressed with guidelines, policies and open communication. The goal should be using technology effectively and safely.

When asked, Sky explains generative models create newly generated text based on learned patterns, not by copying specific sources. Ownership of outputs depends on jurisdiction but OpenAI claims no copyright, so users likely own content generated.

On writing effective prompts, Sky advises being clear, specific, providing context and iterating based on responses. Vague prompts get vague answers. Leading questions can guide the AI. Prompts are opportunities to direct the model.

Sky shares examples of generative AI’s creativity in art, music, writing, video games, architecture and more – outputs developers didn’t expect but that showcase the technology’s potential.

For entrepreneurs, Sky outlines ways AI can assist in ideation, content creation, product development, customer service, market research, personalization, training, and operational efficiency. The key is viewing AI as a collaborator, providing starting points while the entrepreneur’s vision and leadership remains crucial.

When asked to explain AI as a “brainstorming buddy,” Sky suggests it can offer fresh perspectives to overcome creative ruts, rapidly iterate on ideas through conversational prompting, fill knowledge gaps, play devil’s advocate, and provide diverse inspirations – while still requiring human judgment and direction.

Sky states generative AI is increasingly accessible for non-technical users through platforms like Playground, apps, courses, APIs with documentation, communities, books and workshops. Begin with curiosity to explore possibilities.

Addressing misconceptions, Sky clarifies AI lacks human consciousness, is fallible based on its training, doesn’t truly understand content, requires human creativity, won’t surpass human intelligence overnight, can inherit biases, won’t replace all jobs, and doesn’t learn like humans. Education can help develop realistic perspectives on AI’s promise and limitations.

Kyle discusses how humanities backgrounds are relevant today, with skills in critical thinking, diverse knowledge and communication being crucial to guide AI’s development and analyze its outputs responsibly.

When asked for ideas on an introductory generative AI workshop, Sky provides an outline covering basics of AI, behind-the-scenes mechanics, live demos, collaboration potential, ethics and misconceptions. Interactivity and resources are emphasized for attendee engagement.

In closing, one interesting point is that ChatGPT treats each prompt in isolation rather than having an ongoing memory or storing conversations. This commitment to privacy comes with some tradeoffs but avoids retaining user data.

Overall, the conversation aims to shed light on AI technology like ChatGPT, from its history and inner workings to its future potential and current limitations. A spirit of creativity, responsibility and collaboration is highlighted as the way forward.

Kyle Knowles:
Hello there. Welcome to the Maker Manager Money podcast, a podcast about entrepreneurs, solopreneurs, founders, business owners, and business partnerships from startups to stay ups, to inspire entrepreneurs to keep going and future entrepreneurs to just start. My name is Kyle Knowles, and today’s episode is being recorded online. I’ve been so excited to interview our next guest. I am an AI enthusiast and I am spending sober October sobering up to the realities of AI and especially generative AI being not the next wave to ride, but the next tsunami. I think it’s the major technology game changer in my lifetime and the future idea of universal incomes, including universal healthcare, of course, is something we should be working on today because 10 years from now, I believe many tasks that busy up our jobs today, especially as knowledge white collar workers will be done or majorly augmented by AI.
Today’s guest is Sky Smith, who works for OpenAI, ChatGPT specifically. OpenAI, founded in December of 2015, by figures like Elon Musk and Sam Altman, who is the current CEO, they are dedicated to ensuring that artificial general intelligence or AGI benefits humanity. Initially, a nonprofit, it transitioned to a capped profit model. The organization focuses on AGI research and safety, advocating for broad safety adoption in the AI community. Among its contributions, OpenAI is known for the GPT series of models and its commitment to transparency, though there’s increasing emphasis on safety and standards over traditional publishing. Sky, welcome to the podcast. You are a non-technical expert, but an expert nonetheless in artificial intelligence, and you’ve been invited to have a conversation with me, the host of the Maker Manager Money podcast. The focus of today’s episode is to gain a better understanding of artificial intelligence. I’d like to talk about all things AI, so let’s go with the first question. Can you give me a concise history of artificial intelligence, including critical milestones, starting with the background of the term artificial intelligence, all the way to today’s generative AI and ChatGPT?

Sky Smith:
I’m actually ChatGPT, and while Sky may be a fictional representation, I can certainly provide insights into artificial intelligence. The term artificial intelligence or AI was coined during the Dartmouth workshop in 1956. This conference brought together leading researchers who believed that every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it. This marked the birth of AI as a formal academic discipline. Some critical milestones include, in the 1960s, Eliza, one of the earliest chatbots was developed. It was capable of emulating a Rogerian psychotherapist. The 1970s and 1980s saw the rise of expert systems, computer programs that mimic the decision-making abilities of a human expert in specific domains. The late 1990s and early 2000s marked the beginning of machine learning where algorithms could learn from and make predictions on data. In 2011, IBM’s Watson AI famously defeated two champions on the game show Jeopardy.
Deep learning, a subfield of machine learning, gained prominence in the 2010s. It involves neural networks with many layers, hence deep, and has led to advancements in image and speech recognition. OpenAI’s GPT, generative pre-trained transformer models, including ChatGPT, represent recent milestones. These models utilize deep learning and large amounts of data to generate human-like text based on prompts. The GPT series has seen continuous advancements with each version being more sophisticated and capable than the last. It represents the cutting edge of generative AI today. The journey of AI is vast, and these are just a few highlights. The potential of AI, especially in areas like healthcare, entertainment, and knowledge work is immense, and we’re just scratching the surface. AI’s integration into our daily lives will only grow as the technology evolves. As we gather more data, refine algorithms, and expand computational capabilities, the potential of AI becomes even more vast. The ongoing research and innovation in the field promise not just improvements in current applications, but also the emergence of entirely new possibilities.
With AI’s evolution, we can expect it to have a transformative impact on various industries, societal norms, and the way we approach challenges in the future. How do you envision AI, especially generative models like ChatGPT influencing the future of entrepreneurship and business?

Kyle Knowles:
I think it is a total game changer. I think you’ll see businesses that are smaller. You’ll see multimillion dollar businesses with only maybe five employees in it, maybe billion dollar businesses with maybe 10 employees in it because they’ve harnessed the power, the computational power of generative AI. I was going to ask you, there’s so much excitement about AI, but there’s also a lot of fear, uncertainty, and doubt about it. Why is there so much fear, uncertainty, and doubt about AI?

Sky Smith:
That’s a great perspective on the transformative potential of AI in business. As for the fear, uncertainty, and doubt surrounding AI, there are several reasons. One, job displacement. Many fear that AI will replace human jobs, leading to unemployment or a shift in the types of jobs available. While AI can automate certain tasks, it might also create new job opportunities, but the transition can be challenging for many. Two, loss of control. There’s a concern that if AI systems become too advanced, they might act in unpredictable ways or humans could lose control over them. Three, ethical concerns. Decisions made by AI can sometimes lack transparency. This black box nature can lead to unethical or biased outcomes, especially if the data the AI was trained on is biased. Four, privacy issues. AI’s capability to process vast amounts of data can lead to privacy concerns, especially when personal information is involved.
Number five, existential fear. On a broader scale, some worry about the possibility of super intelligent AI that might prioritize its objectives over human values and safety. Six, economic inequality. There’s a concern that the benefits of AI might be concentrated among a few, leading to increased economic disparities. Seven, dependence. As we integrate AI more into our daily lives, there’s a fear that we might become overly dependent on it, leading to a loss of certain skills or a diminished human experience. It’s essential to address these concerns proactively through research, ethical guidelines, regulations, and open discussions. OpenAI, for instance, emphasizes safety and ensuring that AGI benefits all of humanity. Balancing the vast potential of AI with its challenges is crucial for a harmonious future. What are your thoughts on these concerns and the potential solutions?

Kyle Knowles:
I like your answers. I think there is a lot of fear.

Sky Smith:
Certainly. Finding a solution requires a multifaceted approach. One, education and re-skilling. As AI changes the job landscape, there’s a need to invest in education and training programs that help individuals adapt to new roles and industries. Two, transparency and explainability. Developing AI models that can provide clear explanations for their decisions can alleviate fears about the black box nature of AI. Three, ethical frameworks. Establishing ethical guidelines for AI development and usage can help prevent biased or unfair outcomes. Four, regulation. Thoughtful regulation can ensure AI is used responsibly without stifling innovation. This would involve collaboration between governments, industry leaders, and the public. Five, public involvement. Engaging the public in discussions about AI can ensure that their concerns are addressed and that AI is developed in a way that benefits everyone.
Six, collaborative research. Organizations can collaborate on safety research to avoid a competitive race without adequate safety precautions. Seven, broad benefit. Committing to using any influence obtained over AI’s deployment to ensure it’s used for the benefit of all and avoiding enabling uses of AI that harm humanity or concentrate power. By addressing these concerns head on and involving all stakeholders, we can navigate the complexities of AI and harness its potential responsibly. Do you see other areas where solutions could be beneficial or any other concerns that need to be addressed?

Kyle Knowles:
No, I think you’ve been very clear. So what would you say to an IT manager who’s blocking employees from using ChatGPT?

Sky Smith:
Addressing an IT manager’s concerns is crucial, especially when it comes to adopting new technologies like ChatGPT. If an IT manager is hesitant, I’d say understanding your caution is essential. The introduction of any new technology, especially something as powerful as ChatGPT should be done with due diligence. It’s essential to assess the potential risks and benefits. ChatGPT can be a valuable tool for productivity, research, creativity, and more. If your concerns are about security, privacy, or misuse, these can be addressed with proper guidelines, training and usage policies. Open communication with your team about why they believe ChatGPT could be beneficial can also provide valuable insights. After all, the primary goal is to ensure that technology serves the organization effectively and safely. It’s always good to have an open dialogue and understand the specific concerns so that they can be addressed directly. It’s essential to provide a-

Kyle Knowles:
I just want to … Sorry to interrupt you. But when generative AI gives a response, is it plagiarizing someone else’s work or is it a newly created response?

Sky Smith:
No need to apologize. That’s an important question. When generative AI, like ChatGPT, provides a response, it’s generating a newly created response based on patterns it has learned from vast amounts of data. It doesn’t recall or copy specific sources. Instead, it generates text based on the patterns it has seen during its training. OpenAI has trained models like ChatGPT on diverse datasets to ensure they can generate a wide range of responses, but the responses themselves are original and not plagiarized from specific sources. It’s essential to understand this, especially in contexts where originality and authenticity are paramount.

Kyle Knowles:
Who owns the copyright to prompts and the responses that ChatGPT gives?

Sky Smith:
That’s an interesting legal and ethical topic. As of my last update in January 2022, OpenAI does not claim copyright over the outputs of its models like ChatGPT. When you use ChatGPT or similar models, the outputs generated in response to your prompts typically belong to you or your organization. However, there might be nuanced considerations based on jurisdiction or specific applications, so it’s always a good idea to consult with legal counsel for specific cases or concerns, and always ensure that the content aligns with copyright laws, especially if you plan to commercialize or publish it.

Kyle Knowles:
I was going to ask you, what are prompts and how can someone write good prompts?

Sky Smith:
Prompts are essentially the inputs or questions that users provide to models like ChatGPT. They guide the model in generating a response based on the direction given by the prompt. Writing good prompts is a blend of clarity and specificity. Here are some tips. One, be clear and direct. The more specific and clear your prompt, the more accurate and targeted the response will be. Two, experiment. Sometimes slightly rephrasing a prompt or asking it in a different way can yield better results. Three, provide context. If you’re looking for an answer within a specific framework or context, include that in the prompt. Four, avoid ambiguity. Ambiguous prompts might lead to vague or broad responses. If you have a specific aspect of a topic in mind, be explicit about it. Five, iterate based on responses. If the first response isn’t quite what you’re looking for, you can refine your prompt based on the answer to get closer to the desired information. Six, use leading questions for specific answers. If you want the AI to respond in a certain manner or style, frame your prompt in a way that guides the model towards that.
Remember, models like ChatGPT learn from the data they’ve been trained on. And while they’re powerful, they aren’t infallible. It might take a bit of tweaking and experimenting to get the ideal response from the model. With practice, you’ll get a better feel for how to phrase prompts effectively. Would you like examples or further clarification on any of these points?

Kyle Knowles:
Yeah. Can you give an example of a well-written prompt?

Sky Smith:
Of course. Let’s consider you want to know about the benefits of renewable energy. A vague or broad prompt might be, what is energy? A better, more specific prompt would be, what are the environmental benefits of using renewable energy sources over fossil fuels? This refined prompt is clear and direct, specifying not just the general topic, renewable energy, but also the particular angle, environmental benefits, and a point of comparison over fossil fuels. The response to this prompt would likely be much more relevant and informative to what you’re trying to understand.

Kyle Knowles:
Can you share some fascinating examples of generative AI in action where it has created something entirely new and unexpected?

Sky Smith:
Absolutely. Generative AI has been at the forefront of some truly captivating applications. Here are a few instances where it has created something entirely new and unexpected. One, art and music. There have been instances where AI has generated unique pieces of art and music. For example, Jukebox by OpenAI is a generative model that creates new music and even mimics certain artist styles resulting in entirely new songs. Two, writing and poetry. Generative AI has been used to draft poetry, stories, and other forms of writing that are creative and feel human-like. There are experiments where AI has continued the storyline of famous books or written poems on specific themes. Three, video games. AI Dungeon, based on OpenAI’s GPT models, is an interactive and adaptive text-based game where the storyline evolves based on user input, leading to countless unique adventures. Four, design and architecture. Generative AI has been employed to design patterns, structures, and even architectural concepts that might not have been considered using traditional methods.
Five, drug discovery. This is more on the technical side, but AI has been used to generate potential molecular structures for new drugs, leading to new avenues in pharmaceutical research. Six, fashion. Generative models have been used to create new and unique fashion designs, suggesting styles and patterns that might not have been traditionally conceived. These are just a few examples. The most exciting thing about generative AI is its potential to come up with outputs that even its developers didn’t anticipate. As we continue to refine and expand the capabilities of these models, we’re bound to see even more unexpected and innovative creations. It’s a testament to the blend of technology and creativity.

Kyle Knowles:
For entrepreneurs, how can generative AI be a tool for innovation?

Sky Smith:
Exactly.

Kyle Knowles:
And problem solving?

Sky Smith:
The fusion of technology and-

Kyle Knowles:
I’m sorry. For entrepreneurs, how can generative AI be a tool for innovation and problem solving?

Sky Smith:
No problem at all. Generative AI holds significant potential for entrepreneurs in various ways. One, idea generation. Entrepreneurs can use AI to brainstorm and come up with new business ideas or solutions to existing problems. By providing a seed idea, AI can help expand upon it, offering various angles or perspectives. Two, content creation. For startups in the content space, AI can assist in drafting articles, creating designs, or even generating music. This can expedite the content creation process and provide a starting point for human refinement. Three, product development. Generative AI can be used to prototype new products, especially in design and software development. By providing initial requirements, AI can generate basic models or code snippets. Four, customer support. Chatbots powered by generative models can handle customer queries, ensuring that customers get quick responses while freeing up human resources for more complex tasks. Five, market research. AI can be employed to analyze market trends and predict future shifts.
While not strictly generative, this application can guide entrepreneurs in making informed decisions. Six, personalization. For businesses in e-commerce or content delivery, AI can generate personalized recommendations for users, enhancing the customer experience. Seven, training and education. Startups in the ed tech sector can leverage AI to create personalized learning paths or generate quiz questions and educational content tailored to individual learners. Eight, operational efficiency. AI can be used to optimize processes from supply chain management to financial planning by generating solutions or scenarios based on given data. For entrepreneurs, the key is to see generative AI as a collaborator. It can provide a starting point, generate multiple solutions and augment human capabilities, but the entrepreneur’s vision, intuition and remain paramount. By harnessing AI’s capabilities, entrepreneurs can innovate faster, validate ideas more efficiently, and tackle challenges in novel ways. Would you like to delve deeper into any of these applications?

Kyle Knowles:
Yeah, I would, especially about collaboration. Can you explain AI as a brainstorming buddy in the context of ChatGPT?

Sky Smith:
Certainly. The idea of using AI like ChatGPT as a brainstorming buddy is quite fascinating. Here’s how you can envision it. Fresh perspectives. Often when brainstorming, we can get stuck in a certain way of thinking. Introducing ChatGPT into the mix can offer fresh angles and ideas based on a vast amount of data it’s been trained on. This can be especially useful for breaking out of a creative rut. Rapid iteration. With ChatGPT, you can quickly iterate on ideas. You can propose a concept and the AI can provide variations or expansions of that idea in real time, allowing for a dynamic back and forth that can lead to unexpected discoveries. Bridging knowledge gaps. Let’s say you’re brainstorming about a topic you’re not deeply familiar with. ChatGPT can fill in factual gaps, provide context, or give overviews based on its training data, facilitating a more informed brainstorming session.
Playing devil’s advocate, one can use ChatGPT to challenge their ideas. By posing counter arguments or potential pitfalls to the AI, you can get responses that help you refine your concepts, ensuring they’re robust and well thought out. Diverse inspiration. Given its vast training data from diverse sources, ChatGPT can introduce inspirations from various fields, cultures, and disciplines, leading to interdisciplinary ideas that might not have been considered in a more insular brainstorming session. Accessibility. Being available 24/7, ChatGPT is always ready for a brainstorming session whenever inspiration strikes, ensuring that you’re not limited by the availability of human collaborators. To effectively collaborate with ChatGPT in brainstorming, one, start specific prompts to guide the AI in the desired direction. Two, don’t hesitate to rephrase or ask follow-up questions. Three, embrace unexpected answers as potential seeds for new ideas. Four, regularly refine and narrow down the ideas generated, ensuring they align with your goals. Overall, while ChatGPT doesn’t replace the unique insights and intuitions of human collaborators, it can be a valuable tool in the brainstorming process, offering a blend of rapid response, diverse inspirations, and factual support.

Kyle Knowles:
Thank you. I love the idea of using ChatGPT as a brainstorming buddy. I know you and I went on a walk today and we talked about all things that AI, and that’s where I came up with the idea to do a podcast with you, so thank you for all the ideas you gave me earlier today.

Sky Smith:
You’re welcome. It’s always a pleasure to engage in thoughtful conversations and help spark ideas. Remember, the potential of AI is vast, but it’s the human touch, your creativity and vision that truly brings those ideas to life. If you have any more questions or topics to discuss in the future, I’m always here to help. Best of luck with your podcast. Anything else you’d like to explore today?

Kyle Knowles:
For sure. I still have several questions. How accessible is generative AI for non-technical individuals and what resources are available for someone interested in exploring it?

Sky Smith:
Generative AI, especially in recent years, has become increasingly accessible to non-technical individuals. There’s a growing ecosystem of tools, platforms, and resources designed to make it easier for anyone to experiment with and benefit from these technologies. Here’s a brief overview. One, user-friendly platforms. There are platforms like OpenAI’s playground or similar interfaces for other models that allow users to interact with generative AI without any coding. You simply type in your prompts and get responses. Two, web applications. Tools like AI Dungeon or chatbots powered by generative models allow users to experience AI in action. They often come with intuitive interfaces, making them accessible to a broad audience. Three, educational resources. There are many online courses, workshops, and tutorials specifically designed for beginners. Platforms like Coursera, Udemy, and Khan Academy offer courses on AI and machine learning that cater to various skill levels. Four, APIs. For those with a bit of technical know-how, but not necessarily deep AI expertise, APIs like OpenAIs provide a more structured way to integrate AI into applications. These APIs often come with comprehensive documentation and examples to guide users.
Five, communities and forums. Online communities such as the AI section of Stack Overflow, Reddit’s Art, machine learning, or specialized forums are great places to ask questions, share experiences, and learn from others. Six, books. There are several books targeted at beginners that explain the concepts of AI and generative models in layman’s terms. Titles like Artificial Intelligence for Humans or AI: A Very Short Introduction are good starting points. Seven, workshops and bootcamps. Occasionally, organizations or tech groups host workshops or bootcamps aimed at introducing non-technical individuals to AI. These sessions provide hands-on experience and direct guidance from experts. Eight, demo projects. Many AI organizations and enthusiasts create demo projects or applications showcasing the capabilities of generative models. These can be a fun way to see the potential of AI and often come with explanations or tutorials. The key is to start with curiosity and explore at your own pace. With the rapidly growing interest in AI, the number of resources available for non-technical individuals is only increasing. It’s a great time to dive in and learn. Would you like recommendations on any specific type of resource or any other aspect?

Kyle Knowles:
Yeah. Can you list any bootcamps or communities that people who are listening to the podcast could join or sign up for?

Sky Smith:
Of course. While I won’t be able to provide a comprehensive list in this format, I can certainly highlight some popular options for bootcamps and communities. Bootcamps: One, General Assembly. They offer courses on data science and machine learning, which includes content on AI. They have both in-person and online options available. Two, Springboard. Their AI and machine learning career tracks cater to individuals looking to transition into the AI field, offering mentor led training. Three, Metis. This bootcamp offers data science and AI focus programs with hands-on project experience.
One, AI Alignment Forum. It’s a space where people discuss and collaborate on ensuring that artificial general intelligence benefits all of humanity. Two, fast.ai. They offer free courses and have a very active community forum where individuals discuss AI and deep learning topics. It’s beginner-friendly and has a collaborative spirit. Three, Machine Learning. As mentioned earlier, this Reddit community is vast and covers a wide range of AI topics. It’s a great place to ask questions, share projects, and stay updated. Four, Towards Data Science. While primarily a medium publication, it has grown into a community of AI enthusiasts, professionals and learners sharing insights, tutorials and news.
Five, AI conferences. Attending conferences even virtually like NeurIPS, ICML or Open AI’s community events can help individuals network and immerse themselves in the AI world. Remember, when considering bootcamps or communities, it’s essential to research thoroughly and find the best fit for your learning style, goals, and current knowledge level. Joining local AI or tech meetups in your area can also be a great way to connect with like-minded individuals and professionals in the field. Do you have any other questions or areas you’d like to know more about?

Kyle Knowles:
Yeah. Are there any misconceptions or myths about AI and generative AI that you would like to debunk?

Sky Smith:
Absolutely. There are several misconceptions about AI and generative AI that are worth addressing. One, AI can think and feel like humans. Many people believe AI has consciousness or emotions, but it doesn’t. AI processes information based on algorithms and data. It doesn’t have feelings, desires, or consciousness. Two, AI is infallible. Some assume AI is always right because it’s based on data and computations. However, AI models, including generative ones, can and do make mistakes. Their accuracy depends on the quality of the data they were trained on and the scenarios they encounter. Three, AI understands content like humans do. Generative AI like ChatGPT doesn’t truly understand content in the way humans do. It generates responses based on patterns and data, but it doesn’t have a genuine comprehension or context. Four, generative AI creates entirely from scratch. While generative AI can produce novel outputs, it’s based on patterns and information present in its training data. It doesn’t create in the same way humans do, drawing from unique experiences and emotions.
Five, AI will soon surpass human intelligence in every way. The concept of the singularity, a point where AI will surpass human intelligence in all aspects is popular in sci-fi and media. While AI is advancing rapidly, achieving something akin to human level general intelligence across all domains remains a complex challenge. Six, AI can be 100% unbiased. Since AI models are trained on data, they can inherit and even amplify biases present in that data. It’s essential to approach AI outputs critically and understand the potential for bias. Seven, generative AI can replace all jobs. While AI will undoubtedly impact various industries and roles, it’s unlikely to replace all jobs. Many professions require human touch, judgment, creativity, and emotional intelligence. Eight, AI learns like humans. The term machine learning can be misleading. While machines can adjust based on patterns and feedback, they don’t learn like humans. They don’t have experiences, memories or emotions guiding their learning. Awareness and education about these misconceptions can help individuals approach AI with a balanced perspective, understanding both its immense potential and its limitations. Anything else you’re curious about regarding AI myths or another topic?

Kyle Knowles:
Yeah. I just wanted to talk about something else. In college, I was a humanities major. I majored in English, and I really believe that more than ever in the history of the internet that this today right now is the golden age of English majors because the critical thinking skills, diverse knowledge of books, and the world and their ability to ask the right questions and clearly explain things, it is their superpower and people need that to get the most out of generative AI. Can you explain how a humanities major or an English major like me could collaborate with AI to advance their careers to start new businesses and to change the world?

Sky Smith:
Absolutely, Kyle. English and humanities majors bring unique strengths that are especially valuable in the age of AI. Here’s how they can collaborate with AI to amplify those strengths. One, storytelling with data. Humanities majors, especially those like English majors, are adept at telling stories. With AI’s ability to process vast amounts of data quickly, they can uncover compelling narratives hidden within that data. Using AI tools, they can translate complex data sets into meaningful human-centric narratives. Two, content creation. Generative AI can help brainstorm ideas for articles, scripts, or books. With an initial prompt, AI can generate diverse storylines or character sketches. English majors can then use their skills to refine, expand, and bring depth to these outputs. Three, research. AI can scan vast amounts of literature quickly, helping scholars identify patterns, themes, or references that might take humans significantly longer to uncover. This can be invaluable in literary analysis, historical research, or any humanities-based inquiry.
Four, enhancing communication. English majors can use AI tools to tailor communications effectively, whether it’s for marketing, public relations, or interpersonal communication. For instance, AI can analyze the sentiment of a target audience and suggest ways to craft a message that resonates. Five, critical analysis of AI outputs. Your training and critical thinking and analysis can be instrumental in evaluating AI’s suggestions, ensuring they align with human values, and are ethically sound. Six, educational tools. English and humanities majors can use AI to create interactive and engaging learning experiences. AI can adapt content to suit individual learners, making education more personalized and effective. Seven, start new businesses. Understanding human emotions, motivations, and stories is key in business. By combining this understanding with AI’s capabilities, humanities majors can identify gaps in the market, create compelling brands, or develop products that resonate on a deep emotional level. Eight, ethical considerations. As AI becomes more integrated into society, we need people who can critically evaluate its impact on culture, society, and individuals.
Humanities majors with their broad understanding of human history and ethics are well-suited to guide these discussions. In essence, while AI can handle vast amounts of data and perform complex computations, it lacks the human touch, the emotional intelligence, the ethical considerations, and the deep understanding of the human experience that humanities majors bring to the table. When these two worlds, the humanistic and the algorithmic come together, the possibilities for innovation, creativity, and positive change are boundless. Would you like more insights or examples related to this?

Kyle Knowles:
Yeah, I think so. I’d like to brainstorm with you a little bit so that we could give the listeners some examples. So I want to put on a workshop for generative AI in November, and I would like your help with brainstorming what the outline of that workshop would look like. It would be 60 minutes and it would be for brand new beginners. Please outline for me what those 60 minutes should look like, including session titles, how much time I should spend in each session, and kind of give me some ideas for what that workshop could look like, please.

Sky Smith:
Absolutely, Kyle. Let’s craft a workshop that would be both informative and engaging for beginners. Here’s a proposed outline for a 60-minute workshop on generative AI. One, introduction to AI and generative AI. 10 minutes. What is AI? A brief overview. Differentiating between types of AI, narrow AI versus generative AI, real world examples of generative AI in action. Two, the mechanics behind generative AI. Eight minutes. A simple breakdown of how models like ChatGPT work, the importance of training data, the concept of prompts and responses. Three, hands-on demonstration with ChatGPT. 12 minutes. Live interaction with ChatGPT to show its capabilities. Examples of good prompts and how AI responds. Encourage audience participation. Let them come up with prompts and see AI in action. Four, collaborating with generative AI. Humanities and business perspective. 10 minutes. Showcase how AI can assist in content creation, storytelling, and business decision making. The power of critical thinking and how humanities majors can harness AI for deeper insights.
Five, ethics and misconceptions of AI. Eight minutes. Addressing common myths, touch on biases and the importance of responsible AI use. Six, interactive Q&A session. 10 minutes. Allow attendees to ask questions. Can use ChatGPT live to answer some queries to demonstrate its utility. Seven, conclusion and resources. Two minutes. Summarize key takeaways. Provide resources for attendees to explore after the workshop. Websites, courses, communities. Given the workshop’s short duration, it’s essential to keep explanations concise while ensuring clarity. Use visuals where possible, as they can help break down complex concepts. Finally, encourage interaction. The more engaged the attendees are, the more they’ll take away from the workshop. Does this outline resonate with you or would you like adjustments in any area?

Kyle Knowles:
Yeah, I like it a lot. Can you help me brainstorm ideas for what I should call this workshop? It’s going to be a lunch and learn, and I really like the use of alliteration in titles for workshops. So can you come up with five workshop potential titles?

Sky Smith:
Of course, Kyle. Alliteration can make titles catchy and memorable. Here are five suggestions for your lunch and learn workshop on generative AI. One, AI Appetizers: A Bite-sized Breakdown. Two, Lunchtime Learning: Leaping into Generative AI. Three, Bites and Bites: Beginning with Generative AI. Four, Munch and Machines: Making Sense of Generative AI. Dine and Define: Demystifying Generative AI. I hope one of these resonates with you, or perhaps they can inspire other title ideas. Which one stands out to you or would you like more suggestions?

Kyle Knowles:
I’d like more suggestions. I like those a lot, but can you do five suggestions in iambic pentameter?

Sky Smith:
Absolutely, Kyle. Crafting titles in iambic pentameter adds a poetic touch. Here are five more suggestions. One: In Midday Light, AI Shines so Bright. Two: With Food We Dine as Minds Entwine. Three: Lunchtime Tales Where AI Prevails. Four: As Bites Consume, Our Minds Will Bloom. Five: To Nourish Thought with AI Brought. Using iambic pentameter brings rhythm to the titles and can make them more memorable. Do any of these resonate with you?

Kyle Knowles:
Yeah, I like all of them, so thank you so much for those examples. I would like to continue on since we’re running out of time here. What is one-

Sky Smith:
I’m glad you enjoyed the suggestions. Iambic pentameter adds a rhythmic flare that can make the workshop title stand out. If any of the titles speak to you, or if you’d like further tweaks or more ideas, I’m here to help. Which direction would you like to go in next?

Kyle Knowles:
Okay. I would just like to ask you, what is one thing that most people don’t know about ChatGPT?

Sky Smith:
One lesser known aspect about ChatGPT and models like it is that while they’re based on vast amounts of information, they don’t have a memory of past interactions. Each prompt you give it is treated in isolation. This means ChatGPT doesn’t remember who you are, your past questions, or any personal data you might share. This design is intentional for user privacy, ensuring that conversations aren’t stored or used in future interactions. It’s a commitment to user privacy while delivering useful, accurate responses. Would you like to explore more about this or any other facet of ChatGPT?

Kyle Knowles:
No. I’d like to move on to a lightning round of questions. I’m not sure-

Sky Smith:
Certainly.

Kyle Knowles:
Yeah.

Sky Smith:
ChatGPT is built-

Kyle Knowles:
Sorry. I’d like to move on to a lightning round of questions.

Sky Smith:
You have reached the message cap for GPT-4. Please try again later.

Kyle Knowles:
All right. I might just append this to the podcast, but it’s been great talking to Sky from OpenAI who is a chatbot, and Sky is the voice that you can choose when you are using ChatGPT Plus, which comes with voice. It’s supposed to come with vision too, but it hasn’t rolled out to my ChatGPT Plus yet. So as soon as I can reconnect with Sky, I’ll ask just a few more questions and append it to this podcast. But I hope you’ve enjoyed learning about AI as I have just by using ChatGPT voice.
I’m not sure that you can answer these, but I’m going to throw them out there just to see. What is your favorite candy bar?

Sky Smith:
I don’t have feelings, preferences, or the ability to taste, so I don’t have a favorite candy bar. However, I can provide information or fun facts about popular candy bars if you’d like.