dariusmonsef

What does a digital brain look like? How do you visualize its activity? I thnk there is little utility in a digital brain visualization… but hey, they’re fun so here’s how I decided to do take a crack at it for OzBrain.com!

Other robust knowledge wikis like Obsidian have very good brain maps and knowledge node explorer views… but I can’t imagine many people who are looking for a file would try and navigate the tree of knowledge, vs just searching for it.

Agent-first product design enables the highest utility when we just ask agents for things, vs clicking around on our own. I am not the primary worker in my day-to-day getting shit done. My agents are, so the interface should be built first-class for them… and then give me some ways to audit or see at a glance what is going on.

But… Let’s do some visuals!

The OzBrain Signature Visualization. Every digital brain on OzBrain generates a unique signature based on a slice of recent data. Inspired by a brain MRI... I broke this down to Diversity, Density, Activity.

How diverse is the brain and how many collections of knowledge does it hold. - This adds diversity to the number of colors that represent the brain

How many articles build up the depth of each collection. - This adds the focus and clarity of the lines.

How much are agents reading and accessing this knowledge. - This generates the speed of the activity in the brain.

For comparison, the attached video is first my staging environment brain that holds limited knowledge and activity vs my actual brain that I use everyday with my agents. https://youtu.be/Wip9oTxm-II

*OzBrain actually has a built in task system and one user has knocked off hundreds of tasks without even using many articles… gonna have to figure out how to visualize that next!

--

What’s the coolest brain visualization you’ve seen?

If you were working n the OzBrain signature visualization, how would you improve it?

Hi. Just shipped a new version of the site. How’s this feel now? https://ozbrain.com/

Hi. Just shipped a new version of the site. How’s this feel now? https://ozbrain.com/

Hi. Just shipped a new version of the site. How’s this feel now? https://ozbrain.com/

What do you think would improve that? Different language? More visuals?

Also thanks for doing the mental hurdling to get thru! I will work on that image...

Noted! Thanks. I appreciate you getting thru it. What do you think would have made it more clear/faster?

Yeah, I wouldn't put your crypto wallet into OzBrain or anything... this is more the day to day operation of your business or personal life. While that is also valuable, our data is already sitting on a lot of servers so might as well make it useful for you and your agents.

What I'm building has a regular maintenance it runs across the knowledge. A. to chunk long articles down into smaller chunks or information with good routing. B. To review the latest thinking to make sure old thinking is deprecated and linked.

This will be the big challenge with this. Storing/Retrieving works fine across hundreds of articles... I have more than 500 right now just from OzBrain and a bit of the thing I was building before. When it becomes thousands with dozens being added all the time, the maintenance will become more of a necessity.

I lean a little bit towards store everything because the system's abilities are changing so fast right now, that having our previous thinking or reasoning is useful in being able to improve or correct current assumptions based on previous data.

Cool. If I'm reading it right, it only allows read access? How do you get knowledge into the brain?

I'm assuming that the content you put in is all agent driven. Source files (ie, a transcript, a data log, stuff that shouldn't eery change) get recorded but the articles are some kind of agent reasoning on top of it. I don't have any interface to edit an article. Your agent does it.

And yeah I agree about not building for human consumption. The agents create and consume vasly more data than we can.

I don't do any ingestion, I figure out can automate that in a routine or a zap or whatever you want.

Cool. Johnny Decimal sounds badass too, like maybe if an agent loses some of my work it will hunt it down and kill beat it up ;)

Are you seeing engineers use it or more every day people getting the hang of it to?

That sounds like a solid setup for engineering. OzBrain is meant for not just the eng docs, but also the business and product docs. Also my personal stuff, ie my workout log is there, health analysis, etc. And I share our family docs with my wife, so all the kids stuff and our stuff is synced across devices. I think anybody that works on a computer, regardless of industry is going to be a heavy agent user in the future, and I'm building for all of them. Super smart people already have homegrown solutions that work for them.

What a surprisingly lovely comment section. I love the curiosity and positivity. Hope you have a great day.

Thanks! Where did you end up With your thinking? Would love to have you try it and tell me where you think k it could be better.

Yeah, close. But all handled so a user can just focus on connecting the MCP and doing work. Goes to my answer elsewhere here about different between gbrain / llm-wiki… those are highly configurable and powerful, but require technical knowledge and time. I think the vast majority of tech workers will need “that for dummies” ie this just works for them.

While I agree with you on most of what you’re saying, I think there is still an audience outside of your viewpoint that may see value. But thanks for sharing your insights! Super helpful.

I think they’re similar but different. Memory is usually single user focused, a summary of specific facts or instructions. OzBrain is everything that was captured, reasoned, promoted, etc. and auditable record of what the latest thinking is and why. When agents are the primary user you need a way to capture why they ended up at where they ended up. Memory is more like the filtered summary of everything it has saved as important.

A constant challenge. Don't have a perfect solution for it yet, but importantly every change to any article logs who did it, what it did and the reasoning behind it. So I have enough data to work with as I continue to improve things.

Along with my conversations with the 75 founders, there were two other distinct camps that I saw... those that care about the display and UX for them to see their .md files... and those (of which I am in) that don't care, I don't want to see the markdown, it is for my agents to see and relay to me what is important.

OzBrain is text only, and basic markdown formatting (OKF). No fancy PDFS, etc. BUT that is what your agent is for. If you want to generate a fancy PDF, have an agent reference the relevant articles and generate what you want. I've considered a lot of additional services like automatic ingestion, ie every call transcript auto-ingested as a source article in the brain... but you can do that with routines or Zapier. My focus is on the infra of the data getting to and from your agents.

I think Karpathy's llm-wiki was a boost to the wiki-markdown club, and it makes a ton of sense why a technical person would adopt that. They're already veru comfortable with github and moving files with terminal commands. They're not my target audience right now, but maybe when we have more robust brain maintenance they'll decide it's jsut easier to use OzBrain and not maintain their own thing.

P.S. In true move fast break things mode... I spent a good 15min working on the name. :) It works for now and if it really works well for people and they love it, the name won't matter so much.

P.P.S. Thanks, it's either genius or incredibly stupid... hard to know these days as stuff moves so quickly and your AI tells you you're so smart. It's what I HAD to build, because I don't speak any of the current languages. I lost my coding skills long ago, so at a point where it might make sense for a smart engineer to review or approve something... I needed to insert an agent that actually knows what it's talking about and understands how to make a good decision. That thing is constantly improving or breaking... I tried to automate one more step a few days ago and have been paying the price and bug fixing my thing that builds things, instead of just BUILDING THE THINGS! ugh.

P.P.P.S. I mean I used OzBrain to build OzBrain and make that eng flow work. If you think something in there would be useful, I'd just point an agent at it that is connected to your whole workspace/flow and ask it what is useful/dumb.

*Bonus point. Because my whole brain is in OzBrain and it knows what I'm building, why, how... I can take a talk transcript like Garry's from Startup School and just ask an agent "Save this transcript in the brain as a source and then review it and show me where this validates or invalidates some of my thinking. And what else would be interesting for me to consider in my broader work."

Obsidian would fall into the same answer as how is this different than gBrain. Obsidian is a powerful and configurable... but requires more work to maintain and sharing knowledge with others is more difficult. I want this to be a super easy way for agents to connect to my knowledge, and make it easy for me to share my chunks of knowledge with my teammates and their agents. Just MCP in and have your agents get to work.

In a way yes. A hosted llm-wiki, where I handle the diffing, versioning and audit log of what was changed, by what agent and why.

Was very hard to navigate COVID where the business was whiplashed in both directions for a couple years. Money went to opening clinics and staff… in hindsight there were smarter ways to grow the business.

Appreciate the pour. Glad you got great care for your daughter.

There are 25 million ER visits a year for kids where the vast majority of those visits are not life-threatening and don’t require that level of care. An urgent care is 1/10 the cost of an ER and can create a better environment for the kid and family… so better care for 1/10th the cost.

This is something @Asa Miller and I are building: apps for kids under 12 that unlock access to the wonder and power of the web in a fun / safe way.

We’ve started with social messaging for kids under 12, that includes a kid friendly GPT bot they can interact with... Genie.

DM me for early beta access.

Beta site: https://beta.genie.gg

Well thank you very much for the support.

Thanks! We've got more goodness on the roadmap and still working towards the best marketplace possible.

Good luck with your book.

Can't wait to start my coke habit and wreck some hotel rooms! ;)

Wait wait... two people seemed to disagree on the internet and then after a few traded messages they end up in agreement and a better understanding of each other... we might have just broken the internet. :)