OpenAI Dots have memory. Your agent still needs a knowledge base.
OpenAI Dots can remember context and work continuously. Here’s why always-on AI agents still benefit from a durable, user-owned knowledge base.
Posts tagged Essays
Don't put AI in your notes app. Connect your notes to your AI.
People are right to worry that AI will fill their notes with writing they didn't do. But that comes from note apps that generate text on their own. There is another way round: keep the notes app simple, and let the AI you already talk to read and write them.
What AI actually does with write access to your notes: 108,613 tool calls analyzed
We analyzed 30 days of MCP traffic on Hjarni: 108,613 tool calls from 558 people who let ChatGPT and Claude read and write their knowledge base. A quarter of everything the AI does is writing, and it updates old notes far more than it creates new ones.
The best model keeps changing. Your memory shouldn't start over.
GPT-5, then the Claude 5 family, then whatever ships next quarter. The best model changes every few months, and chat history, built-in memory, and Projects don't come with you. A knowledge base over MCP does.
Your Client Will Edit the Prompt. Plan for It.
Handing a client the source code and prompt strings is not a handoff. They will edit the prompt, it will break, and you will get the 8 PM text. Ship the bundle instead: an eval suite, a shared source of truth, and a change trail.
The Team Second Brain
A second brain is personal. In the AI era, teams need a shared one: notes each teammate can connect to their own Claude or ChatGPT. Here's what that looks like.
The Team Knowledge Base for Claude and ChatGPT
Your team already runs on an LLM wiki. It's just single-player. Here's how to build a shared knowledge base every teammate's Claude and ChatGPT can read.
Onboard New Hires With an AI That Knows Your Runbooks
A new hire's AI starts from zero, so they ask a senior engineer instead. Give the whole team's runbooks to a shared knowledge base and their assistant answers on day one.
Claude Projects and ChatGPT Projects Don't Scale to a Team
Projects live inside one assistant, built for a focused task. A team needs shared context both Claude and ChatGPT can read. Here's the gap, and what closes it.
Context engineering is not just for AI teams. It is what your notes should do for you.
AI teams obsess over feeding the model the right context at the right time. You can do the same thing with your notes, so your AI pulls what it needs instead of you re-explaining yourself every session.
What Should Your AI Actually Remember?
Indexing session transcripts sounds like memory. It mostly retrieves scratch work. The useful stuff was already distilled into artifacts. Keep the artifacts.
Why Isn't Your AI Second Brain Working?
Most AI second brains fail because they dump too much into context. A useful AI memory uses retrieval, freshness, provenance, and short curated notes.
Who owns MCP now? Nobody. That is exactly why you can build your AI's memory on it.
Anthropic donated MCP to a vendor-neutral foundation backed by its biggest rivals. Here is what that means, and why it makes a knowledge base built on MCP a safe bet instead of a single-vendor gamble.
Your Notes Aren't Agent-Legible Yet
Context engineering made code legible to coding agents. Your notes need the same thing. Capture is solved. Legibility is the bottleneck.
Is MCP dead? No. You're just pointing it at the wrong job.
The "MCP is dead" posts are right about a real problem and wrong about the cause. Forty-two always-loaded tools is a design smell, not a protocol flaw. Here's where MCP actually wins.
Generation is solved. Finding it again is the problem.
A chat log is a transcript, not a knowledge base. Generation got good. Retrieval is the part that is still broken. Here is why, and what fixes it.
What is a second brain? A guide for the AI age
A second brain is a place outside your head where you store everything worth keeping. Here's what it is, why it matters, and why AI changes what it needs to do.
LLM wiki vs. plain Markdown: when Karpathy's gist stops working
A plain Markdown file works for the LLM wiki pattern. Until it doesn't. Here are the five points where it stops working, and what to use instead.
What is a Second Brain? The AI-Native Take (2026)
A second brain is an external system for what you read, think, and decide. The traditional version is a graveyard. Here's what replaces it in 2026.
My Marketing Team Is Two AIs and a Knowledge Base
I run Hjarni's entire growth program from inside Hjarni itself. Two AIs, three folders, and a built-in MCP server. Here's the real loop, step by step.
Your Tools Do Too Much
Notion has 47 features you'll never touch. Obsidian needs plugins for everything. Both assume you want a system. You don't. You want to write something down, find it later, and let your AI read it too.
Karpathy's LLM Wiki is right. I just didn't want to run it locally.
Andrej Karpathy's LLM wiki gist nails it. Stop dumping documents at LLMs. Build a brain. Hjarni hosts it over MCP, in Claude and ChatGPT, on every device.
Stop Explaining Your Codebase to Your AI Every Time
Your AI remembers a few facts about you, not your codebase. Here are five notes that fix that, so every conversation starts with context instead of from zero.
Never Lose Your Startup's Knowledge
Learn how startups avoid information loss by using Hjarni's shared knowledge base.
Your AI Hit Its Limit. Your Knowledge Shouldn't.
Every AI conversation eventually resets. The real cost isn't the limit. It's re-explaining yourself every time. Your knowledge deserves a home you own.
I gave Claude access to my notes through MCP
I connected Claude to my notes via MCP. Setup took 2 minutes. Now every conversation starts with my project context already loaded — no pasting, no re-explaining. Here is what changed about how I work.
Start here
Write once. You both remember.
25 notes free, forever. No credit card required.
5 out of 5 from
14 reviews
7,600+ notes a day read or updated by AI
Works with Claude, ChatGPT, Cursor and any MCP client.