Learning to Live With AI
Listen on Spotify, Apple Podcasts, or Substack. Collect this episode on PODS.
Continuing on Maslow’s Hierarchy Level 2 with Financial Security and Mental Safety from Entries 6-8, this week I am going to touch on artificial intelligence (AI). This topic is massive. It’s hard to emphasize how much detail and research goes into truly understanding AI. I have been researching AI for many years, much more seriously since 2020. I still feel as though I am barely scratching the surface.
But that’s okay. Understanding even some of the basics can provide a solid foundation for understanding the current and future impact of AI. Before we get to the stage of using AI to bolster Financial Security through future-proofing and automation, let’s take a few minutes to address the Mental Safety component ✨
The goal of this entry is not to teach you everything about AI, that would be impossible. The goal is to identify a few concepts that will hopefully teach you about why AI is in the news so much lately and help alleviate the “AI-nxiety” that comes from public mass hysteria over change. Much of this anxiety arises from trying to deal with the uncertainty of AI.
Prior to the hype in the news, AI was an unknown unknown to many people (something that operated in the background that most people were unaware of). With the publicity over the last year, AI has now moved into a known unknown for many people: they recognize that AI exists and is having an impact on the world around them, but they do not understand why or what it is. There is an inability to understand the paradigm shift that is occurring from a lack of foundational knowledge.
Let’s change that 😌
Note: that this entry will touch on a few topics related to AI. As mentioned in Entry #7, there is now a fork in the road of Recalibrating. I will be continuing this discussion behind a subscription paywall, including generating specific guides on how I am building my second brain as a way future-proof myself by collaborating with and protecting myself against AI. This newsletter and other creations have become my primary source of income. If you find value in this newsletter, I would greatly appreciate the support of your subscription so that I can continue to share more about what I’m learning and how it can help you ✨
AI-nxiety
Are you feeling worried or anxious about how AI can affect your career? Your privacy? Are you uncertain about whether what you are seeing online is “real” (human-made) or AI-generated?
You are not alone. You might have “AI-nxiety”.
What is AI-nxiety?
Definition: Unease about the overarching ramifications of AI on human creativity and ingenuity. The sense of foreboding as to whether or not what you’re seeing is being created by man or machine.
The unease and sense of forebording are both related to a single anxiety-inducing issue: uncertainty. Fear of the unknown is one of our most primal human fears, sometimes referred to as the “one fear to rule them all“. No wonder AI has been so stressful, it is full of uncertainty in so many ways:
- Will AI replace my job? (financial insecurity)
- Is democracy still a valid political system? (political insecurity)
- Is what I am looking at real or AI-generated? (a lack of truth)
- Is who I am talking with online a person or AI? (fear of being misled)
- Are there going to be battle robots fighting wars in the future? (a genuine issue)
Much like the limitless uncertainty, there is no end to why people are experiencing stress about AI. Each person is going to experience different stresses related to AI, and that is okay.
The goal here is to identify some overarching ways that people can reduce their AI-nxiety,
How To Reduce Anxiety From AI (AI-nxiety)
Taking a Break
With bottomless rabbit-holes available to deep dive into AI research, sometimes the best way to cope with anxiety is to take a break. Disconnect from the online world a bit. Go touch some grass. The research will be there whenever you are ready.
You do not need to FOMO (fear of missing out) that you are going to fall behind if you take a few days or weeks off. Rest is okay and encouraged.
That said, there is value for people who think for a living (especially for creators and knowledge workers) to learn about advancements in AI and how they may leverage these tools to stay ahead of the curve and increase the security of their jobs. In other words, undergo AI learning to future-proof yourself.
Learning
The best way (in my experience) to deal with Ai-nxiety is to convert the known unknowns into a known known.
Instead of panicking over “AI is going to take my job and I’m going to lose my financial security”, we can take the time to learn what AI is, how it is currently being used, and how best to future-proof ourselves in an ever-changing world.
For example, AI has been used for years as part of the Duolingo app to help people learn languages. Siri, Alexa, and other smart phone voice “assistants” make use of AI to translate voice to text and conduct searches online. Tesla uses AI for vision recognition and vehicle control for autonomous steering. There are many, many ways that AI has been used for years without you thinking about it.
When looked at in this light, the light of our past and common uses of AI, artificial intelligence starts to become less scary and more manageable.
Let’s take a deeper look.
What is Artificial Intelligence (AI)?
Definition
Artificial Intelligence (AI) refers to the simulation of human intelligence in machines that are capable of performing tasks that typically require human intelligence. These tasks include problem-solving, decision-making, language understanding, and more. AI systems use data and algorithms to mimic human cognitive functions, enabling them to learn from experience and improve over time.
Types of AI
There are two main types of AI:
- Narrow AI: This type of AI is designed and trained for a specific task. It excels at performing that particular task, but it lacks the ability to generalize its knowledge to other areas. Examples include virtual personal assistants like Siri and recommendation systems on streaming platforms.
- General AI: This is a more advanced form of AI that can understand, learn, and apply knowledge across a wide range of tasks, similar to human intelligence. However, true General AI is still largely theoretical and hasn’t been achieved yet.
What is ChatGPT?
LLM & Foundation Models
ChatGPT is powered by a cutting-edge AI language model called a “Large Language Model” (LLM). An LLM is a program designed to understand and generate human-like text. It achieves this by leveraging complex algorithms and patterns learned from vast amounts of text data. ChatGPT is built upon foundation models like GPT-3, which stands for “Generative Pre-trained Transformer 3.” These models are initially trained on a diverse range of texts, which enables them to respond coherently and contextually.
Put more simply:
LLMs are in the business of finding the next set of words that is maximally coherent with the words that came before it. Its outputs should make sense to the user. What’s so startling about recent LLMs like GPT-4 is that they can make sense without knowing what they’re saying. An LLM doesn’t understand its outputs. It doesn’t grasp the meanings of words. It’s a surprisingly successful (though far from perfect) next-words-predictor.
https://hbr.org/2023/08/generative-ai-nxiety
NLP (Natural Language Processing) & Prompting
ChatGPT is part of the field of Natural Language Processing (NLP), which focuses on enabling computers to understand, interpret, and generate human language. In the context of ChatGPT, users communicate with the model through prompts.
Prompts are input instructions or questions given to the model. Based on the provided prompt and its learned knowledge, ChatGPT generates text in response. Crafting clear and specific prompts is crucial to obtaining accurate and relevant responses from the model.
For example, this section (What is AI? What is ChatGPT?) was actually written by ChatGPT 👀 (except for the quote I included). To get ChatGPT to write this section for me, I used this prompt:
The rest of this newsletter (and all other newsletters) have been entirely written by me. The point of using ChatGPT to write this section was to illustrate how it may be used to save time (e.g., explaining facts about AI).
Why Should I Care About AI?
Now that you understand more of the basics of AI (at least the publicly hyped portions), let’s consider why you should care to continue learning more about this disruptive technology.
Future-Proofing Yourself
For the first time ever, we are starting to see the replacement of tasks done by “thinkers” rather than “doers”. It is important to recognize that, at this point at least, the AI systems being developed are not at a stage where they can fully replace most workers. Instead, these AI systems are replacing TASKS that the workers do, not their entire jobs.
How long will this distinction remain true? The annoying answer? It depends.
“AI won’t take your job, but someone who knows how to use AI might” is a fairly common saying that has been going around for the last 6 months or so. The reason my answer to the above question is “it depends” is based on this concept that jobs are going to adapt and change to leverage AI.
People who learn to use AI to improve their productivity, efficiency, and quality of work will be able to stay ahead of those who do not.
The type of AI skills needed to “stay ahead” of the curve is going to vary with almost every job type. This variation is part of the reason for AI-nxiety, there is so much uncertainty about who will be affected and how.
The best way to deal with this uncertainty is to make it more certain: learn about what is happening in your industry and how you can leverage AI to improve your own experience. By learning how your specialized skillsets are adapting to account for AI, you can keep up with new skills that will help you stay relevant for many years to come.
Customization vs. Generalization
Consider what we know about ChatGPT now: it is part of a foundation model that has been trained on the ENTIRE Internet (or a good chunk of it). That’s a lot of data, both good and bad. It’s also very general.
It’s actually a good analogy for how knowledge workers are valued. The more specialized your skillset, the more valuable you become as a worker. The more generalized your skillset, the more replaceable you are with someone else who has the same general skills.
ChatGPT is a bit like the generalist, it’s been trained on so much data that the answers it provides are also generic. Of course, better prompting can improve the quality of the output from ChatGPT. That prompting is a result of increased specialized knowledge of learning how to prompt.
For example, instead of asking ChatGPT a general question, I can say “you are a lawyer, a patent agent, an engineer, a photographer, a blockchain expert, etc.” “based on this information, answer my question in a specialized manner”. The output will be much better than the general prompt without limitations.
The more you learn how to prompt with ChatGPT and other NLP-based models, the more effectively you will be able to leverage AI in your life.
OpenAI has recognized this fact and has actually updated ChatGPT to have a custom instruction section that can factor out some of my common prompt-limitations (like my skillsets) so that the AI response is more customized to me as an individual.
These custom prompting instructions can save you time and energy when asking ChatGPT questions.
Geoffrey Hinton’s Commentary in Toronto (2023)
I was going to write up a full section on all of the reasons why you should care about AI, but it began to be too daunting of a task. Instead, I’ll take a clip from one of the world’s leading experts on AI who spoke in Canada recently.
Geoffrey Hinton, the “godfather” of AI, spoke at the Collision 2023 Tech Conference and gave an excellent overview on where we are at with AI. Famously, Hinton resigned from Google to be able to speak more openly about AI and the current concerns.
The talk is about 20 minutes long; I have embedded the video to start at Geoffrey Hinton’s interview so you do not need to watch the entire video (which is 7 hours long).
I suggest watching the video to get a high-level overview on the main concerns with AI. The goal of this interview is not to make people panic, but to increase awareness over what goes into AI and how we can leverage it to build a better future (while avoiding a dystopian one).
The issues Geoffrey Hinton lists in the video are not to be taken lightly. They are also not intended to evoke fear, but rather to inspire people to learn to keep up with the changes that are happening in our lifetime so that we can continue to build a better world together.
If you are interested in learning more about Fake News and its impact on our democracy, I wrote a paper back in law school (2018) on this topic that you might find interesting.
Where is AI Headed? Building for the Future
The more you learn about how to use AI in your life, the more you will be able to adapt with AI as it develops over the coming years.
The Internet is developing to become more semantic (language + logic). This semantic web will make it easier for humans and machines (AI) to communicate, thereby improving how we interface with the Internet. Understanding prompting (using your own language as an input for AI) will be a valuable skill moving forward. More details on the semantic web may be found here.
Knowledge graphs power many semantic search sites and are becoming more and more popular (Google has been using them for years). Knowledge graphs are a network of connected data points that enable deeper understanding of the information provided. More on knowledge graphs may be found here.
You are already familiar with the display of knowledge graphs. When you search for a restaurant, person, book, etc., the display of the location, details, images, videos, and more is being pulled from the knowledge graph for that search result. It’s a way to display many related pieces of information, connected through semantics.
Remember, this semantic web powered by knowledge graphs is something that is happening already, and is not something to fear. Instead, these technologies can encourage you to create your own specialized version of knowledge that will help you communicate with yourself and others, rather than dealing with the generalized versions shared online.
This specialized version of your own database can form intellectual property that can be extremely valuable for creators and knowledge workers. As this IP is captured and organized in a knowledge graph, a second brain (digital collection of knowledge) can be created. For example, here is a timelapse of my knowledge graph from a few weeks ago:
The knowledge graph is extremely versatile and can be customized in many ways. By clicking on a circle (node made of a note) I can follow an edge (connecting line) to related notes that share similar topics. My system involves using colour as way to help identify connections, as shown in the view below:
AI will provide a valuable tool to be able to interface with our second brains, to gather insights about the knowledge we have accumulated that we may have missed on our own.
If you are interested in learning more about the creation of a second brain and how to leverage AI to build your personal brand while improving your productivity, efficiency, and learning, I will be writing about this building process for my paid subscribers.
Researching creation systems and emerging technologies takes a lot of time and is now my full-time job. Paying for a subscription helps me to continue creating valuable content for your education, so I appreciate any support you can give ✨
Next week: Automation and the Creator Economy
I hope that this overview helps to alleviate your concerns and anxiety over AI. While it is a disruptive technology, it is one that anyone can learn to leverage to improve the quality of their lives. In particular, AI is an excellent task-automator, a topic we will get into next week.
The creator economy is built by creators and knowledge workers spending many hours performing tedious and time-consuming tasks. I’m excited to get into how AI can help give us some time back. Stay tuned ✨
P.S. If you would like to see a historical record of my AI knowledge accumulation, you can see my Flipboard Magazine: AI Learning that goes back for several years.
