AI tools like Claude and ChatGPT are indispensable now but they can also feel a little soulless. Use them enough and you start to think the way they do.
I found a way to make the AI experience far less soulless: strike a balance between the liberal arts and AI with Hermes Agent, which is an agentic AI harness for building soulful agents that you can train on your own brand story, voice and values.
That blend is what makes Hermes completely different than just using Claude or ChatGPT. Most AI tools hand you the machine and leave the human part to you as an afterthought. Hermes Agent puts both in one place, so raw intelligence and human judgment work as a single motion.
Why does that matter? Because AI on its own produces slop. It is brilliant and tasteless at the same time. The liberal arts are what supply the taste, the judgment, and the reason to care. Hermes Agent is the first tool I have used that I feel can be designed around that truth instead of pretending the machine can do it all.
Hermes was the Greek god who moved between worlds. He carried messages from the gods to mortals, guided souls from life to death, and stood at every boundary, crossroad, doorway, and border. His work was translation: taking meaning from one place and making it land in another.
That is the exact job of anyone working with AI today. On one side sits a machine with near-infinite knowledge. On the other sits human meaning, purpose, and taste. The work is carrying value across the gap between them.
Hermes Agent is named for the god who did that for a living.
What Makes Hermes Agent Different From Other AI Tools
Every AI agent you have used has the same problem. You teach it something on Monday and by Tuesday morning it has forgotten everything. You are back to square one, re-explaining your project structure and your whole workflow. It is like training a new intern every single morning.
Nous Research built Hermes Agent to kill that cycle. Hermes is an open-source, self-hosted AI agent that is easy to run on your computer or on your own server. It learns from every task it completes and gets measurably better the longer you use it.
Since launching in February 2026, it has collected more than 100,000 GitHub stars and triggered what developers call a migration wave away from OpenClaw, the open-source framework that had dominated the space before it.
The interesting part is not the adoption numbers. It is the architecture underneath, and what it means for anyone who wants an AI that compounds in value over time instead of resetting to zero every session.
Here is what that architecture does. Every task Hermes finishes feeds back into what it knows. It saves what worked as a reusable skill and sharpens that skill the next time. It also builds a steady model of how you work and what you want. Nothing gets thrown away when the session ends.
The agent you use in month three is a better agent than the one you started with, because you are the one who trained it.
If you want to train Hermes to fully understand your brain story, voice and mindset, I have a free training in my Social Creators Community on creating a soul.MD file for Hermes Agent.
What AI Does Well And What It Cannot Do
Modern AI is an autistic savant. Hand it a problem and it returns something near PhD-level command of the field, in seconds, across more subjects than any human could hold at once. The number crunching, the first drafts, the process work, all of it is handled.
What it does not have is a reason to care. It can’t tell you which idea is worth your life. It has no read on what people quietly want, no story of its own, no stake in the outcome. It answers. It does not aspire.
Here is the split I work from:
| The AI carries | You carry |
|---|---|
| Research and synthesis | The question worth asking |
| First drafts and iterations | Taste to pick the right one |
| Number crunching and process | Story, meaning, and point of view |
| Speed and total recall | Purpose and the will to build |
The left column used to be your job. Now it is table stakes. The right column is where you earn your keep.
How to combine AI and the liberal arts with Hermes Agent
Here is the method I use. It works on an essay, a product, a business, or a course.
- Start with a question worth answering. The AI is answer-rich and question-poor. A liberal-arts habit of mind, trained by philosophy and hard reading, is what lets you ask the question that actually matters instead of a shallow one.
- Hand the process work to the model. Research, first drafts, data, structure, the parts that used to eat your week. In Hermes Agent OS this is where the savant intelligence does its best work, at speed.
- Judge the output with taste. The model will hand you twenty options. Choosing the true one is a human act, and taste is built by years of reading great work and understanding why it was great.
- Add the layer a model can’t fake. Your story. Your point of view. What you understand about what people long for right now. Rhetoric and history live in this step.
- Ship it, watch real people react, and adjust. Curiosity closes the loop. A founder who is genuinely curious about people learns faster than any model can.
The difference between AI slop and work that moves people is almost always steps 1, 3, and 4. Those are the liberal-arts steps, and Hermes Agent is built to keep you in charge of them.
The Liberal Arts And The Hermeticism Philosophy
In ancient Greek mythology, Hermes was the messenger, the guide, and the god of language, commerce, and interpretation. The Greek word for interpretation comes from his name, which is where we get hermeneutics, the art of drawing meaning out of a text. He was the one who made the divine legible to human beings.
Over time the Greeks fused him with the Egyptian god Thoth, the inventor of writing and the keeper of wisdom. That combined figure became Hermes Trismegistus, “thrice-greatest Hermes,” the legendary author of the Hermetic writings. Those texts are the root of Hermeticism, an old tradition built on one idea: “as above, so below.” The pattern of the higher world is mirrored in the lower one, and the human mind is what connects them.
You can hear why that fits a tool built to join machine intelligence and human judgment. The AI is the vast store of knowledge. You are the mind that gives it meaning and direction. The design assumes neither one is whole without the other.
There is a second thread, and it runs straight into the liberal arts. In the fifth century, Martianus Capella wrote The Marriage of Philology and Mercury, a book used to teach in European schools for centuries. In it, Mercury, the Roman name for Hermes, weds Philology, the love of learning. Their wedding gifts are the seven classical liberal arts: grammar, logic, rhetoric, arithmetic, geometry, music, and astronomy.
So in the book that handed the West its list of the liberal arts, those arts arrive as the dowry at the wedding of Hermes and learning itself. A tool that puts AI and the liberal arts in one place is drawing on a very old marriage. That is the foundation the name points at.
I did not come to any of this from a tech background. I came to it from a stubborn belief about how people should learn.
I built DIY Genius to help people build an education around the point where their passion meets their purpose, the thing they love set next to the thing they are meant to do. For years, the hard part of getting ahead in life was access to credentials because the best knowledge sat behind expensive institutions and gatekeepers.
That problem is gone. Anyone can reach the world’s knowledge now, and a machine will read all of it for you before breakfast. The new hard part is different. It is knowing what to ask, what to keep, what to build, and why any of it matters.
This is where creative work becomes a liberal arts problem, and I believe combining AI with the liberal arts using AI tools like Hermes is the future of knowledge-based work.
The Soft Skills Are The New Hard Skills
For a generation, we told young people the humanities were a luxury. The real path was technical. Learn to code, study engineering, pick the major that pays, and treat philosophy and literature as something you do on the weekend. The message was that meaning is nice but skill is what feeds you.
That trade is inverting. The technical execution, the exact thing we called safe, is the first thing the machine takes. Writing the code, running the numbers, cleaning the data, producing the competent draft, all of it is becoming cheap and close to infinite.
When competence is infinite, competence stops being the edge. What is left is everything the machine cannot fake. Judgment about what is worth doing. A point of view someone actually holds and can defend.
The word liberal in liberal arts never meant political. It meant free. These were the arts of a free person, the disciplines that make a mind able to think for itself instead of taking orders. For centuries that freedom looked impractical next to a trade. Now it is the most practical thing you can own, because taking orders is exactly what we have automated.
Business runs on this more than people admit. Customers do not buy the most efficient thing. They buy the thing that means something, from someone they trust, told in a way that moves them. AI can generate endless output, but it cannot want anything, and it cannot make you care. The person who supplies the wanting and the caring is the one the market pays.
So the future of business is not colder and more mechanical. It runs the other way. When the mechanical part is free, the human part is the whole game. The people who use AI in creative and soulful ways, who bring a real center of meaning to all that machine power, are the ones who will build what lasts.
Here is the good news under all the anxiety about AI. The age that was supposed to make us less human is going to reward the most human thing about us. Read deeply, ask hard questions and then let the machine carry the weight while you carry the meaning.
I believe the future of education starts with helping people find their "zone of genius" so they can develop their uniquely human ability to create meaning, connection and community (that A.I. can't replicate).
I'm also big on completely unplugging from tech to recharge on nature retreats and trekking adventures. That led me to start my most recent company Sacred Treks.
- 10 Best Home Neurofeedback Devices For Brain Training - July 14, 2026
- 10 Best Online Course Platforms (2026 Pros & Cons) - July 14, 2026
- How To Combine AI And The Liberal Arts With Hermes Agent - July 5, 2026
