What if the world reset to zero overnight? no machine and people back to prehistoric times? You can see it happened in an anime called Dr. Stone where the present world got cursed (petrification) that every living creature is turned to stone, including human and nature takes over. But one day, a guy called Senku is freed from stoning because of some bat poop liquid (nitric acid) that able to crack the stone by trigger depetrification by corroding the outer stone layer (you’ll know after watch the first episode).

He’s the only person who lived. Senku is a master of science, he can build machines, make medicine and re-invent previous tools like sunglasses, soap, antibiotics, even a radar and telephone machine. It all counts as leaps civilization forward 100 years earlier simply because he knows science. Notice how one domain of knowledge can skipped 100 years time of research and trial & error.

That’s when you realize how valuable science truly is. Science can explain any kind of natural phenomenon, how worlds really works, and even knew something beyond our universe. Our most powerful access to science is books.

A book is the greatest gift our predecessors ever left us. Notice how their 10–20–100 years of research are all condensed into a pile of paper. You can compress years of research into minutes, digging into someone's mind and know exactly what happened in the past, thanks to the book.

A few hours of reading can save you years of research.

Every book is someone else’s years, gifted to you. Every page you turn is years of someone else’s obsession, failure, and breakthrough gifted directly to you. You are not just reading, you are digesting research and knowledge for centuries. We are the wealthiest humans in history, not because of money, but because of access to knowledge.

So, have you absorbed a century of human knowledge today, or in simpler terms, have you read a book?

Tired of recurring billing and limited tokens? Now you can run it with no subscription, no internet, no recurring billing, and unlimited tokens.

We gonna use Ollama for downloading local LLM, there is so many good models you can pick any models like Minimax, GLM, or Kimi, but at this time open-source model that allowed to download is Qwen 3.5. Qwen 3.5 is one of the most advanced open-source models.

1. Download Ollama

You can download Ollama here.

2. Download LLM (Choose Qwen 3.5)

And download the model, at this time we use Qwen 3.5, you can download here.

3. Install OpenCode or Pi.

Both OpenCode and Pi are open-source AI Coding Agent and can integrate with a local LLM. Install OpenCode or Pi.

3.1 Integrate Local LLM with OpenCode.

Basically, Ollama will automatically configure itself for OpenCode or you can directly install with ollama launch opencode --model qwen3.5 But if it doesn’t work i provide the manual here.

3.1.1 Config model with OpenCode.

After install opencode you can open the config with nano ~/.config/opencode/opencode.json and paste this config code.

{
  "$schema": "https://opencode.ai/config.json",
  "provider": {
    "ollama": {
      "npm": "@ai-sdk/openai-compatible",
      "name": "Ollama",
      "options": {
        "baseURL": "http://localhost:11434/v1"
      },
      "models": {
        "qwen3.5:latest": {
          "name": "qwen3.5:latest"
        }
      }
    }
  }
}

3.1.2 Connect the model that we have been config.

The next step is to connect with OpenCode. To open OpenCode, you can type opencode in the terminal. Inside the OpenCode interface, type /models to choose a model.

Choose Qwen 3.5, make sure it’s from Ollama (local).

And it runs!

3.2 Integrate Local LLM with Pi.dev

3.1.1 Config model with Pi.dev.

After install Pi.dev you can open the config with nano ~/.pi/agent/models.jsonand paste this config code.

{
  "providers": {
    "ollama": {
      "baseUrl": "http://localhost:11434/v1",
      "api": "openai-completions",
      "apiKey": "ollama",
      "models": [
        { "id": "qwen3.5:latest" }
      ]
    }
  }
}

3.1.2 Connect the model that you have been config.

The next step is to connect with Pi.dev. Open it with typeopencode in the terminal. Inside the Pi.dev interface, type /model to choose a model.

Choose Qwen 3.5, make sure it’s from Ollama.

Here you go, I do a small task with a specific prompt, and it works!

5. Voila! You can run AI Agent with zero dollar spent.

Benchmark :

*Ubuntu run on CPU only, I still optimize to use my GPU in Ubuntu Workstation

One reason I love Taylor’s songwriting is how she turns feelings into something real, from something abstract into something tangible. What does this mean? How does this happen? Before we dive in, let’s see what science says.

For an overview, our brain has two important parts limbic and neocortex. The limbic system is responsible for emotion and memory it is so powerful, isn’t it? But the limbic system doesn’t have the ability to process our thinking into a word. So that’s where the neocortex comes in, the neocortex is responsible for processing our thinking into a word or translating our feelings into language.

Love is so abstract that we can only think about it, we can’t even explain why we feel it. For example, when someone asks, “Why do you love her?” — you know the answer, but putting it into words feels impossible, it feels like you’re interpreting an entire universe of your emotions into a bunch of words, we’re forced to use words, and sometimes we messed up.

In this part, Taylor really nailed it, she takes limbic memories and turns them into specific, sensory details, for example, “green eyes and freckles and your smile” (Everything Has Changed) or “I miss your tan skin, your sweet smile, so good to me, so right” (Back To December) These aren’t just lyrics it described as raw, unfiltered, natural, and original memory of her stories that’s why her songs feel like memories you never had because she translates feelings into language you can recognize. So when her song hits you right in the feelings — yep, that’s your limbic system doing its thing.

Have you ever try to open a book and after 30 seconds your hand reach your phone, sounds familiar? How is that happened?

In this era, scrolling TikTok has become the norm. It’s a hack to get instant dopamine, just lay on a couch and your brain is filled with a lot of happiness.

When it comes to work or learn it really feels painfully boring, no funny video, no laying around, no sensation to get fun thing in a short amount of time.

That’s why you should be friends with boringness.

A research by Timothy Wilson, University of Virginia in 2014 are experimenting by make people to stay silent in a room for 15 min, they dont have anything except one button that make them strike by electricity.

And guess what they do? 67% of men and 25% of women chose to hit the button and feel pain instead of being stuck in a boring state. When you bored even get shocked is entertaining.

I once experienced it, when I was sitting quietly in a room with no device. I finally reach my pile of books and start read it, and also I finish my read-later article and ended up writing down my expenses report that i always avoid.

If you are able to stay in a boring state, you can read a book that doesn't have instant dopamine. No, it won't give you a quick high dopamine until you get ‘A-HA!’ moment, and that’s when real dopamine cames in, and when you experience it, trust me, you will be addicted, and yes! Reading a book can be addictive.

The best dopamine doesn’t come from short, instant scrolling. It comes from the slow, steady joy of learning and growing. That’s the real satisfying dopamine.

Humans often create errors such as write gramatically error or just typos, but that is a human thing, our weakness is having mistake, but our strength is learn from mistakes.

I don’t blame AI for existing in this world but I’m tired that AI is replacing anything we do, if that’s the case, we don’t need to exist we just need AI to exist in this world.

AI is just a helper, not our replacement, so instead of writing an email full of AI, you can type it from your own and let AI correct it. It is far better, it got your point, it came from your heart, and you can save your time by prompt AI to correct your words.

So the point is, instead of write a perfect long word email, it is okay to write somheting worng like what I just did, I’m more happy if you mistakenly make a grammatical error or typos but it is from your brain, heart, and your own hand, rather than you type perfectly essay, but it is full of AI.