I stopped typing prompts. Now I just say it out loud and Jarvis runs my whole company.
Jarvis is a voice assistant sitting on top of my business. I talk, it listens, it acts. Hey Jarvis, pull up Slack, and it does. Hey Jarvis, handle the support inbox, and it routes the work to the right agent and comes back with it done.
It's the Iron Man setup, running for real. One voice on top, the whole operation underneath. No keyboard, no tabs, no prompts. You speak and things happen.
The best part is you can build your own, and it's simpler than it looks.
Watch before everyone has one but you.
P.S. full step by step guide on installing your own Jarvis in the article below.
A model that finishes the job while you sleep is back, and its name is Fable 5.
The newest model takes your hardest problems and runs them on its own, with far fewer check-ins than you're used to. It draws down usage faster than Opus 4.8, but it earns it by actually finishing what it starts.
Until July 7 you can put up to 50% of your weekly limit on it. Hit the cap and you keep going on usage credits.
Watch before you waste it doing the small stuff.
P.S. full guide on getting the most out of Fable 5 in the article below.
A model that finishes the job while you sleep is back, and its name is Fable 5.
The newest model takes your hardest problems and runs them on its own, with far fewer check-ins than you're used to. It draws down usage faster than Opus 4.8, but it earns it by actually finishing what it starts.
Until July 7 you can put up to 50% of your weekly limit on it. Hit the cap and you keep going on usage credits.
Watch before you waste it doing the small stuff.
P.S. full guide on getting the most out of Fable 5 in the article below.
An AI agent spent the night building my second brain while I slept.
This is an Obsidian vault run by a Hermes agent. It reads, writes, and links notes on its own. Mental models, startups, CS, AI, all wired into one living graph. The green nodes are what it touched overnight.
It even builds its own dashboards. On the other screen the agent is coding one from scratch, no prompting from me.
You wake up and your knowledge base grew itself. That's the whole idea.
Watch before you go back to organizing notes by hand.
P.S. build guide in the article below.
An AI agent spent the night building my second brain while I slept.
This is an Obsidian vault run by a Hermes agent. It reads, writes, and links notes on its own. Mental models, startups, CS, AI, all wired into one living graph. The green nodes are what it touched overnight.
It even builds its own dashboards. On the other screen the agent is coding one from scratch, no prompting from me.
You wake up and your knowledge base grew itself. That's the whole idea.
Watch before you go back to organizing notes by hand.
P.S. build guide in the article below.
Everyone's running one Hermes agent. I'm running a team of them.
This is Osiris Console, a control room built on Hermes. 4 agents, each with a role and a live status, working your projects at once. Osiris, Helios, Slate, Forge.
The whole thing runs through Telegram, so the agents ping you for a decision and you approve from your phone. No terminal, no laptop at 2am.
Watch before everyone catches on.
P.S. build guide in the article below.
Imagine hiring an assistant who works 24 hours a day, never forgets anything, and costs you nothing after setup. That is what this guy built, and he called it Hermes.
Hermes runs on your own machine. It browses the web, runs code, remembers past work, and schedules its own tasks, so it keeps going long after you close the laptop. You give it a job once and it handles the rest.
The real power is skills. You plug in whatever you need, from building diagrams to running your Obsidian vault to monitoring your computer, and each one becomes a thing Hermes can do on command. The more you teach it, the more it takes off your plate.
Everyone else is still chatting with AI. This is owning one that works for you.
Full build guide in the article below. The video is just a look at it in action.
You can now build your own AI agent that works while you sleep. This one is called Hermes, and it browses, remembers, and finishes tasks on its own.
This is not a chatbot. Hermes has its own tools: it searches the web, clicks pages, writes and reads files, and keeps a memory between tasks. It even carries its own skills, so it runs a job the same way every time. Give it a command and it goes and does it.
The video says it best: building agents like this is the skill that matters now. Not prompting harder. Building a thing that acts on its own.
Most people still type questions. This person built a worker.
Watch it run before this becomes the baseline.
P.S. build guide is in the article below.
I have been staring at this for ten minutes and I still cannot fully believe it. Someone used Claude Code to turn a plain Obsidian vault into an actual brain. Not a metaphor. Real brain regions, real neurons, firing live on the screen.
There is a Sensory Cortex, a Hippocampus, a Prefrontal layer, an Association region sitting on 37,000 neurons, each one pulsing at its own rate. At the bottom there is a little box that says feed the brain, type something. You drop in a thought and watch the right regions light up like you just woke something.
And here is what got me. It learns. Every note you add builds new connections, so it gets smarter on its own, every single day. Change one idea and the whole thing shifts, the way a real memory pulls everything around it. Your notes stop sitting there dead and start thinking back.
I used to be proud of my tidy folders. After this they feel like a filing cabinet next to a living mind.
Go watch it fire. You will feel the same thing I did.
P.S. full how to build guide is in the article below.
I have been staring at this for ten minutes and I still cannot fully believe it. Someone used Claude Code to turn a plain Obsidian vault into an actual brain. Not a metaphor. Real brain regions, real neurons, firing live on the screen.
There is a Sensory Cortex, a Hippocampus, a Prefrontal layer, an Association region sitting on 37,000 neurons, each one pulsing at its own rate. At the bottom there is a little box that says feed the brain, type something. You drop in a thought and watch the right regions light up like you just woke something.
And here is what got me. It learns. Every note you add builds new connections, so it gets smarter on its own, every single day. Change one idea and the whole thing shifts, the way a real memory pulls everything around it. Your notes stop sitting there dead and start thinking back.
I used to be proud of my tidy folders. After this they feel like a filing cabinet next to a living mind.
Go watch it fire. You will feel the same thing I did.
P.S. full how to build guide is in the article below.
Most people use Claude Code to write code. The smartest ones pointed it at their Obsidian vault and turned their notes into a second brain that works for them.
Obsidian always gave you the graph: every note a node, every idea connected, the whole thing growing into something shaped like a brain. But you still did all the work yourself. That part ends here.
Connected to your vault, Claude Code works inside your notes like it works in a codebase. It reads your full history, drafts new notes in your structure, links them to the right places, and pulls exact context on demand. Ask it to plan a project, it builds the notes. Ask what you decided months ago, it finds the thread.
Obsidian stored your thinking. Claude Code makes it move.
Watch the setup before everyone wires their brain like this.
P.S. full how to build guide is in the article below.
Most people use Claude Code to write code. The smartest ones pointed it at their Obsidian vault and turned their notes into a second brain that works for them.
Obsidian always gave you the graph: every note a node, every idea connected, the whole thing growing into something shaped like a brain. But you still did all the work yourself. That part ends here.
Connected to your vault, Claude Code works inside your notes like it works in a codebase. It reads your full history, drafts new notes in your structure, links them to the right places, and pulls exact context on demand. Ask it to plan a project, it builds the notes. Ask what you decided months ago, it finds the thread.
Obsidian stored your thinking. Claude Code makes it move.
Watch the setup before everyone wires their brain like this.
P.S. full how to build guide is in the article below.
Imagine an AI that already knows every project, client, and decision you have ever made, and never forgets a single one. This guy built it, and it runs his entire work life.
That is the part that matters. It is not a prettier notes app. It is a system you tune to yourself. His sidebar is his whole life sorted the way he works: Businesses, Finance, Projects, Clients, Goals, Research. Everything he learns gets dropped in once and stays there for good, so nothing has to be explained twice.
Then the AI sits on top and reads all of it. Ask it something and it answers with your context, not generic advice, because it knows your projects, your decisions, your past notes. You stop starting from zero every morning. The longer you feed it, the more it works like an extension of your own head.
This is the difference between using AI and building an ecosystem that runs your work with you.
Watch how it is set up. Then go build your own.
He opened his Obsidian graph and it loaded like a galaxy.
Thousands of notes. A spiral hub at the center, hundreds of links firing out of one node. The vault so big Obsidian froze to rebuild its index.
Then the outer field came into view. Dot after dot after dot. None of them touching anything.
Most of it was orphans. Notes saved once, never linked, never reopened, floating in the dark.
One bright spiral did all the work. Everything around it was just storage.
A big graph is not a big brain. It is a warehouse with the lights off.
He didn't have a second brain. He had a galaxy of dead stars and one sun.