ChatGPT is the most exciting technology that I’ve ever interfaced with, no hyperbole. More than the first computer I booted up, the first time I flew or drove, the first smartphone I held. It is beyond anything I thought I’d see in my lifetime, and this is still early days. One thing to keep in mind when you see anything AI-related: today’s version is the dumbest version that you will ever see.

ChatGPT is a type of AI known as a large language model, or LLM. Two things that I believe we will learn in the very near future: LLMs are much more than they appear, and the human brain is far simpler than we think. In other words, the two are more alike than we now appreciate.

This is heresy to both AI researchers who know how LLMs are trained, and to neuroscientists who still don’t understand how human consciousness works. The former sees ChatGPT as too simple to be anything more than a stochastic parrot. The latter can’t help but see us as endowed by something cosmic, spiritual, quantum … or at the very least a statistical improbability that emerges from our trillions of synaptic connections.

I believe as AI research and brain science both progress, they will meet in the middle and find that both LLMs and large collections of neurons have emergent properties from very simple base structures. Chaos theory has long posited that new qualities can emerge simply by increasing the quantity of a thing. Add heat to water and you get warmer water. But keep adding it, and at some point you get boiling water, a very different thing. Same for turning flowing water into ice by gradually removing heat. Newness emerges, gradually and then all at once.

OpenAI researchers who trained ChatGPT have written about their shock at the leap from version 3.5 to 4.0, and it is palpable to anyone who plays with both. A type of newness is emerging. Sure, you will find people sharing the silly mistakes from earlier models, but the leap to 4.0 has been staggering. Where ChatGPT scored in the bottom 10% of tests like the LSAT, the bar exam, and Harvard business school classes, 4.0 has scored in the top 10%. This includes writing essays.

ChatGPT has shown evidence of theory of mind (understanding what others think, even when their thoughts are wrong or based on imperfect understanding). It can solve complex riddles that version 3.5 would fumble. In almost every way, the current Chat far exceeds the individual human mind. We aren’t dealing with something that is eight or nine years old in equivalent development. We are dealing with a hyper genius that makes the occasional gaffe, but is otherwise more knowledgeable, and, yes … wiser than we are.

How can this be? AI is a machine, and wisdom and creativity are under the purview of humans. Creativity, it turns out, is something ChatGPT does far better than humans, because it is trained on our collective creativity. It has more wisdom, because it is trained on our collective wisdom. When it predicts a response, word by word, it is assembling it through a complex statistical analysis of how humans have assembled such words in myriad other ways. I believe this is what our complex interconnection of neurons does all the time. I believe that this is what my neurons are doing right now.

My own belief in the wisdom of ChatGPT arose after delving into the emerging world of AI assisted therapy. On Reddit and Twitter, users are sharing their prompts, results, and general experiences. It has been a revelation. What makes ChatGPT superior to human-assisted therapy is that it is always accessible, can be free or as cheap as $20 a month, and it never tires of listening to our problems. It responds with the same care and attention to every prompt, every single time.

ChatGPT has our collective wisdom, but none of our individual hormones. The parts of us that make us human are the parts that make us ornery, anxious, impatient, distracted, envious, lustful, apathetic, and so much more. With Chat, it’s as if the chemicals that work against us have been distilled out, leaving just the good advice behind. It is all ego without the id.

Below are a few examples to show you what Chat is capable of. These are first attempts, straight from my own prompts. They are fictional. But if you want to jump straight to something insane, check out this separate blog post. It’s by far the most gobsmacked I’ve been in my few weeks with ChatGPT 4.0.

---

My Prompt:

Hey Chat, someone at work really hurt my feelings today. He said things about me that aren’t true to a few colleagues. Now I’m worried what people think about me, even though these things were completely made up. How can I stop worrying about this?

---

Chat’s response:

I’m sorry to hear that you’re going through this. It’s natural to feel upset and worried when someone spreads false information about you. Here are some steps you can take to stop worrying and address the situation:

1. Self-Reflection: Remind yourself of your own values and strengths. If you know the things said about you are untrue, then trust in your own integrity and character.
2. Seek Support: Talk to trusted friends or family members about your feelings. They can offer reassurance and perspective, helping you to cope with the situation.
3. Address the Issue: If you feel comfortable, consider discussing the issue with the person who spread the false information. Explain how their actions have affected you and ask them to correct the misinformation. Keep the conversation calm and respectful.
4. Speak to a Supervisor or HR: If the situation is affecting your work environment, consider discussing the issue with a supervisor or HR representative. They can help to address any workplace conflicts and ensure a positive atmosphere.
5. Don’t Ruminate: Avoid dwelling on the issue or replaying the situation in your mind. Overthinking can amplify your negative feelings and make it difficult to move on.
6. Focus on Positive Relationships: Strengthen connections with your colleagues who know you well and can vouch for your character. This can help counteract any negative impressions created by the false information.
7. Practice Self-Care: Prioritize taking care of yourself both physically and emotionally. Engage in activities that help you relax, and maintain a healthy lifestyle to reduce stress.
8. Accept What You Can’t Control: Recognize that you cannot control others’ thoughts or actions. Focus on what you can control – your own actions and responses to the situation.

Remember that this incident does not define who you are. Keep being yourself and doing your best at work, and your true character will shine through.

My next prompt:

Hey Chat, I want to do something romantic each day this week to surprise and delight my wife. Try to come up with really original ideas, some of which entail going places here in New York City and spending money (like a date or dinner) but mostly come up with little ways to surprise her around the house and in our daily routine.

---

Chat’s response:

That’s a lovely idea! Here are seven original romantic gestures to surprise and delight your wife throughout the week:

1. Monday – Love Notes Hunt: Leave a series of small love notes around the house, with each note containing a clue to the location of the next one. The final note should lead to a small surprise gift or handwritten letter expressing your love for her.
2. Tuesday – Themed Dinner Night: Plan a themed dinner night at home, transforming your dining area to match the theme. For example, set up a Parisian bistro or Italian trattoria ambiance with appropriate music, decor, and a delicious homemade meal. Top it off with a beautiful table setting and a special dessert.
3. Wednesday – Rooftop Sunset Picnic: Head to a rooftop bar or garden in New York City to enjoy a sunset picnic with a view. Pack a basket with her favorite snacks, drinks, and cozy blankets. If you prefer a quieter spot, you can recreate this experience at home by setting up a picnic on your terrace or balcony.
4. Thursday – Serenade Surprise: Learn to play her favorite song on an instrument or prepare a playlist of meaningful songs for both of you. Surprise her by serenading her or dancing together to the tunes in your living room.
5. Friday – Art and Culture Date: Visit a lesser-known art gallery, museum, or exhibit in New York City that aligns with her interests. Follow it up with a stop at a cozy cafe or dessert spot to discuss the highlights of the day.
6. Saturday – DIY Spa Day: Transform your bathroom into a spa sanctuary with scented candles, essential oils, and soft music. Run a bubble bath for your wife, give her a massage, and pamper her with face masks or other skincare treatments.
7. Sunday – Personalized Scavenger Hunt: Plan a scavenger hunt for your wife, with each clue leading to a significant location in your relationship (e.g., where you first met, had your first date, etc.). At each stop, share a special memory or photograph. End the hunt with a romantic meal or a special gift.

Remember to tailor these ideas to your wife’s preferences and interests to make each surprise as thoughtful and meaningful as possible. Enjoy your romantic week!