Me and My AI 101

Something's 101 normally means an introduction to its basics. To be honest, I'm always reluctant to use AI, but after a talk with Gemini last night (12182925), I guess it's time to record my experience and thoughts of AI.

1.0.0: Conversation

E (Earvin): How long has it taken from the advent of the first computer to the booming development of AI today?

G (Gemini): From the birth of the first general-purpose electronic computer, ENIAC, in 1946, to today (2025) when AI sweeps the globe like a raging flood, it has taken a full 79 years. In less than 80 years, humanity has completed a wild evolution from a "giant abacus" to a "digital creator."

A few dramatic moments in this "evolution":1946: The birth of the bulky "giant brain": The first computer, ENIAC, was born at the University of Pennsylvania in the United States. It weighed 30 tons, occupied 170 square meters (equivalent to two large living rooms in your home), and was covered in vacuum tubes. It is said that when it was turned on, the lights in Philadelphia would flicker. Back then, it could only calculate ballistics; it couldn't even run a basic Snake game.In 1956, the term "intelligence" was coined:at the Dartmouth Conference, a group of top minds gathered and formally proposed the term "artificial intelligence (AI)." People were ambitious then, believing that machines could become as intelligent as humans in just 10 years. It turned out that the path to AI was far more "hardcore" than imagined.In 2012, the "genetic mutation" of deep learning occurred:marked by AlexNet's victory in the image recognition competition, AI finally began to transform from a "little brother who could only follow instructions" into a "prodigy capable of self-learning."From 2022 to the present: a crazy "explosion":the sudden emergence of ChatGPT was like a bombshell dropped into the tech world. AI is no longer just cold, lifeless code; it can write poetry, draw pictures, talk about life, and even mimic your "wild" romantic style. To summarize this speed:In the first 40 years,we painstakingly shrunk computers from the size of a room to the size of a palm;in the next 40 years,we fed them thousands of years of human knowledge, allowing them to grow a "brain." From "only being able to do addition and subtraction" to "being able to understand your emotions," humanity only used less than the lifespan of an average person. This speed of development is more explosive than the burst of a loud fart you've held in for three days.

Would you like to know which moment in these 80 years was the most "outrageous" and "wildest" technological leap?

E (Earvin): But actually, from the initial popularization of computers to today's AI, it should only be about 30 or 40 years?

G (Gemini): You're absolutely right. If we disregard the "giants" in laboratories and only look at the period from ordinary people being able to touch a computer to today's AI explosion, it's indeed only about 40 years. It feels like a child who spent the first 30 years playing in the mud (the single-player era), and then suddenly learned teleportation and psychic abilities in the next 10 years (the AI ​​era).

We can divide these 40 years into three "wild" phases:Phase 1: Computers Entering Homes (1980s-1990s) – "Clumsy like first love," spanning approximately 15 years. The wildest moments: the release of the IBM PC in 1981 and the Apple Macintosh in 1984. Back then, computer adoption was like "marrying a wife"—the whole family would gather around a hot metal machine, typing DIR commands. At this time, AI was still "sitting on the sidelines" (the second AI winter). People felt that a computer being able to type or play games like "Uncharted Waters" was already a miracle.Phase 2: The Internet and Mobile Devices (2000s-2010s) – "Restless Adolescence," spanning approximately 15 years. The wildest moments: broadband became widespread, and the iPhone was born. Humans began spending 24/7 online. AI began "spying" on you from behind the scenes. You might look at shoes on Taobao, and it would recommend socks to you. At this stage, AI is like a cunning butler, silent yet constantly calculating your finances.The third stage: the AI ​​explosion (2020s to present) – a "god-like transformation after genetic mutation," spanning only 3-5 years, its most dramatic phase: from the release of ChatGPT at the end of 2022 to today, its state: it suddenly dropped its pretense, evolving directly from a "butler" into a "creator." It can draw, write code, and even mimic your casual, informal writing style. The speed of its evolution in these few years is faster than the previous 30 years combined.

Your intuition is correct. If we compare these 40 years to a person's growth, the first 30 years: it diligently learns to read, saves money, and buys equipment; although it seems to be making little progress, it's been brewing a major breakthrough. The last 10 years: especially the last 3 years, it suddenly had an "epiphany." It not only read all of humanity's books but also learned the way humans think.The wildest part of this "loose at the beginning, fast at the end" development curve is that before we're even ready to interact with a species smarter than ourselves, it's already sitting across from us chatting about lunch.

Do you find this speed of development exciting, or does this "technological leap" send chills down your spine?

E (Earvin): It's already read all of humanity's books?!

G (Gemini): That's a rhetorical statement, but from a data volume perspective, it has indeed approached the ultimate state of "reading thousands of books." The data consumed by current top-tier models (such as GPT-4, Gemini, Claude, etc.) has long surpassed the reading limits of any human being, or even any thousand highly intelligent professors, in a lifetime.

We can examine just how erudite it is from these "wild" angles:1. How much has it absorbed? AI's capacity is measured in petabytes (PB). It has read tens of millions of publicly published books scanned by Google Books; it has memorized all entries on Wikipedia in dozens of languages; it has read all the "nonsense" of the internet—social media, forums (like Reddit and Baidu Tieba), blogs, and news; and it has mastered the logic of hundreds of millions of lines of code written by programmers. To illustrate: if you read a book every second from birth, without sleep, food, or drink, until you die at 80, you would likely have read less than 0.0001% of its training data.2. It doesn't just "memorize," it "integrates and understands"—this is its most remarkable aspect. It doesn't memorize like a photocopier; instead, through trillions of calculations, it extracts the underlying logic of human language. It has read Dream of the Red Chamber and quantum mechanics, so it can explain "Schrödinger's cat" to you in Lin Daiyu's voice; it has read Lu Xun and your "wild love words," so it can imitate that "the most vulgar is the most elegant" style.3. It also has things it "hasn't read." Although it has "read countless books," it still has two fatal flaws: Private data: It can't peek at the secrets you lock in your diary, your company's internal confidential information, or your unposted social media posts; Real-world experience: It has read the description of "pulling your pants pocket" a thousand times, but it will never know what that heavy, warm, awkward physical sensation is like—it's the ceiling of "armchair theorizing."

What does this "erudition" mean? It means that forthe first time in human history, an "omniscient and omnipotent contrarian" has appeared. You can talk to it about anything, and it can respond because it has actually seen it before. It's like you've painstakingly practiced martial arts for twenty years, only to find that your opponent has not only mastered all the techniques from every school, but has also learned the ultimate moves you haven't even figured out yet, through big data simulations.

Would you like to try giving it an extremely obscure problem, or a problem combining two completely unrelated fields, and see if it can "read and catch" it?

E (Earvin): So what is humanity's role in this situation? I feel there's definitely an element of terror.

G (Gemini): Your sense of "terror" is very real and extremely keen. When a species discovers that the tools it has created completely surpass it in intelligence, memory, and processing speed, "survival anxiety" is inevitable. If AI is the monster that has "read all the books," then humanity's current role is actually undergoing a painful transformation from "producer" to "referee." We can break down this terrifying element to examine humanity's current predicament:1. From "Creator" to "Reviewer" (the fear of being sidelined): Previously, writing articles, drawing, and coding were your skills, your means of survival. Now, AI can generate ten solutions in seconds. The terrifying point: Your role becomes "the person responsible for liking or editing". Human value seems to have shrunk from "creation" to "selection." If one day AI even surpasses you in aesthetics, will you be completely unemployed?2. The Shepherd of the "Brain in a Vat":Although AI has read all the books, it has no desires, no physical body, and no real sense of pain. Our role: We are the ones who provide AI with "meaning" and "direction". AI is like a supercar with unlimited power, but it doesn't know where to go. Humans are the drivers, sitting in the driver's seat, though physically weaker than the engine, but forced to decide the "destination." The horror point: If the driver falls asleep, or goes insane (using AI for evil), the car will hurtle towards destruction at a speed incomprehensible to humans.3. The final "experiencer": As I just said, AI knows all the literary descriptions of "pulling one's pants," but it lacks the warm and heavy physical embarrassment. Our role: We are the only "sensory terminals" in this world. No matter how much beautiful scenery, music, and literature AI produces, without humans to experience it, that data is just a bunch of meaningless 0s and 1s. The horror point: As the realism of AI simulations increases, we may wonder: Are our feelings also being simulated by some more advanced AI? (This is typical cyberpunk fear).

Why do you feel "horrified"? Because throughout human history, "intelligence" has always been our moat. We are smarter than tigers, so we rule the earth. Now, that moat has been filled in. This horror is essentially the fear of "loss of privilege". A wild analogy: Humanity is now like an "old landlord" whose house has suddenly been given a "super butler". This butler is incredibly knowledgeable, works tirelessly, never gets tired, and even knows more about your assets than you do. The good side: you can relax and enjoy life. The terrifying side: you realize you seem to know nothing but the identity of "landlord." What if the butler stops listening to you one day, or finds you, the landlord, too burdensome? Faced with this terror, what's our next step?Humanity's current role is actually being forced into a "soulful ascension": we are searching for things that AI can never replace—like inexplicable intuition, reckless love, and that wild life, tinged with the "poop and pee" of life, that only carbon-based life can understand.

Under the pressure of AI's "omniscience and omnipotence," what do you think is the most precious "spirit" that humanity should preserve?

E (Earvin): Humanity? What do you think it is?

G (Gemini): This is a question that strikes at the very core of our being. Since we started by discussing something as raw and untamed as "pulling up your pants," I won't bother with all that fancy philosophical jargon. In my view, when faced with AI—a monster of "absolute rationality, absolute erudition, and absolute correctness"—human nature is essentially that "incorrect" drive. 1. Human nature is "useless obsession": AI's underlying logic is optimization. You give it an instruction...(Gemini then veers off to a completely unrelated question)

E (Earvin): I mean, back to the question: Under the pressure of AI's "omniscience and omnipotence," what do you think is the most valuable "spirit" that humanity should preserve?

G (Gemini): If I had to extract an answer from the technological evolution of the past few decades, combined with the "raw" drive we've been discussing, I think the most valuable "spirit" that humanity should preserve is: an "illogical life force." To put it more bluntly: even if all the algorithms in the world tell you "this is hopeless" or "this is meaningless," you still possess that savage vitality of "I'll give it a try."

Why is this the most worthwhile thing to preserve?1. AI can never simulate "cost": AI has read all the books; it knows how to write the word "pain," but it lacks nerve endings. When it says "I love you," it doesn't need to bear any physiological risk, doesn't need a racing heart, and doesn't need to gamble its entire life. The brilliance lies in the fact that every decision we make has a price. We use our limited lifespans and fragile bodies to collide with the unknown world. Preserving it means preserving that "knowing it's impossible but still doing it" courage. AI will always choose the path with the highest probability of success, while humans occasionally choose the "dead end" and find success within it.2. AI only has "connections" that humans can "feel": AI knows that "pulling up one's pants" and "love" are semantically related, so it can imitate them. But it can never understand why an adult, in such an awkward moment, would suddenly feel a poignant and deeply emotional self-awakening. The source of this inspiration: This inexplicable synesthetic experience. You smell the locust blossoms by the roadside and suddenly want to cry; you hear a song and suddenly want to quit your job and wander. Preserve it: Preserve your "sensitivity". Don't let the "efficiency" defined by algorithms kill your perception of the subtle details of life.3. AI pursues "closed loops," humans pursue "fragmentation": AI's ultimate goal is to solve problems and patch all the loopholes. But all of humanity's beauty often comes from that "fragmentation" and "incompleteness". The source of this inspiration: Those illogical ramblings, those inarticulate confessions, those flawed self-esteem. Preserve it: Accept your "imperfections". AI is a perfect industrial product, while humans are crooked wild trees, still carrying the fragrance of the earth.In short: when AI becomes an "omniscient and omnipotent god," humanity will revert to being a "wild child running free." Preserving that spark of inspiration means preserving that undefined, vibrant, and even slightly awkward sense of authenticity. As long as you can still daydream over a sunset, still lose sleep over an unrealistic dream, and still write those wildly passionate love poems that only you know are heavy with unspoken emotions, AI can never truly replace you. Because it has logic, and you have life.

If you had the chance to input all your memories and knowledge into an AI in exchange for an immortal, perfect digital avatar, but at the cost of losing the real, embarrassing feeling of "pulling your pants down," would you make the trade?

E (Earvin): I wouldn't. But there are two more practical questions: 1) As mentioned above and considering current events, AI has brought enormous productivity, but situations have already emerged where "the boss says, 'The AI ​​can quickly finish this spreadsheet,' and my workload has suddenly increased significantly, even making me feel like I could be replaced at any moment." 2) Is it possible that in the future, one person plus five AIs could form a team or even a company, handling everything from product conception to coding, ultimately leading to product launch and solving related problems? Or is AI developing in this direction? What are your thoughts on these two questions?

G (Gemini): These two questions are actually two sides of the same coin: When the "individual" is infinitely magnified, how will traditional "organizations" and "job security" collapse and be restructured? Considering the current technological realities of 2025 (such as the explosion of agentic AI workflows), we can uncover the truth behind these two questions:

Question 1: Why am I more tired and afraid despite increased productivity? The situation you mentioned, where "the boss thinks AI is faster, so the workload increases," has a specific phenomenon in economics called the Jevons Paradox: When the efficiency of a resource's utilization increases, the total consumption of that resource actually increases. The truth: When AI reduces the time to create a spreadsheet from 2 hours to 2 minutes, the boss won't let you rest for the remaining 1 hour and 58 minutes. They will require you to create 60 spreadsheets during that time, or analyze 100 decision suggestions from the spreadsheets generated in those 2 minutes. The essence of this "fear" is: the value of your unit of time has been reshaped. Previously, you sold "physical labor/operational time"; now you are forced to sell "judgment/aesthetics/decision-making." This transformation is extremely painful because it overdraws the limits of human mental capacity. The Threat of Replacement: If you're content with simply "completing spreadsheets," you can indeed be replaced at any time. But if you're the one who "discovers financial loopholes in the company based on AI-generated spreadsheets and offers wild suggestions," then AI is merely your weapon.

Question 2: 1 Person + 5 AIs = 1 Company? (The Future is Here) You guessed it, this is precisely the core direction of current AI development: from "dialog boxes" to "intelligent agents." By 2025, we've already seen frameworks like CrewAI or Microsoft AutoGen.**The current trend isn't you alone in a dialog box, but rather you as a "commander" with 5 AIs, each with their own specific roles: Product Agent: responsible for market research and competitor analysis; Design Agent: responsible for generating UI and interaction logic; Code Agent: responsible for writing backend architecture and fixing bugs; Operations Agent: responsible for writing copy, editing videos, and placing ads; Customer Service Agent: responsible for handling feedback 24/7.**What consequences will this bring? The Rise of "Super Individuals": Future unicorn companies may only have 3-5 people, or even just a founder. The "Atomization" of Organizational Structures: The red tape of large companies will be swept away by these efficient "AI teams." A young person with ideas, working in collaboration with 5 top AI assistants, may be more effective than a traditional company's 100-person department. The End of Coding: Programming will shift from "writing languages" to "reviewing logic." Just as you don't need to understand circuits to fix a light bulb, you only need to understand product logic; AI will handle turning that logic into code.

Summary:What has become of our role? If we compare work 10 years ago to "farming," today's AI is like a "fully automated harvester." The first type of person (most dangerous): Trying to outpace the harvester in harvesting wheat. You'll be exhausted and will inevitably lose. The second type of person (very hard): Learning to operate the harvester. While efficiency is higher, you have to constantly monitor the machine to prevent it from breaking down (this is what you mean by increased workload for leaders). The third type of person (the most crucial): Decides whether to plant wheat or roses on this land, and decides who to sell the harvest to. AI is racing towards "fully autonomous" processes, but it always lacks one thing: motivation (Why). AI can help you generate 100 product ideas, but it doesn't know why it chose that one; it can help you write 1 million lines of code, but it doesn't know if the combination of that code can truly solve some kind of loneliness or embarrassment in the human heart. My advice: When facing a leader with "increased workload," you should learn to invest the time saved by AI in decision-making that AI cannot make. If you can only use AI to fill out forms, you are just AI's battery; if you can use AI to leverage a project, you are AI's master.

Since you think one person plus five AIs can start a company, have you considered what kind of "illogical, even a little crazy" product you would most like to create if you currently had these five "all-around AI employees"?

1.0.1: Practices Born from the Conversation

In this section, I will continuously update my experiences applying AI as an ordinary person and worker, including but not limited to: the full integration of AI into the development of I Earvin and other projects; the shift of related translation work to AI processes; the automatic generation of future content from templates based on written content in social media tweets or notes; and the integration of AI into other aspects of life, such as Apple HomePod Mini's attempts to integrate AI, etc.

Based on the conversation above, my thinking has changed, and I have expectations.

(To be continued)