My Open Brain breakthrough

You will recall I have been working my way thorough the hefty open brain project Amy assigned me. Organizing and loading all the context in my world was no small challenge - both as measured in tokens (and timeouts) and redirects to re-work short cuts and dumb design choices. I have learned a lot on this journey.

I didn't have any real sense of what I would use this for outside of the obvious use case: a place to ask questions like "who was the person I talked to about neuroscience that pointed me in the direction of this cool podcast?", those third level associations that seem to be harder and harder to navigate these days in my topic cluttered brain.

The reality is that use case is an infrequent one, it probably happens about once or twice a month. But Claude had convinced me to add a Q2 goals to "use Open Brain every day" so I needed to sort out some additional use cases at some point in Q2.

Amy suggested I should use it to help me with scheduling but THAT was an idea that made my stomach churn a bit with anxiety and I told her absolutely not.

Then, I woke up one morning with a new idea to vibe code. I got the idea a bit from Jay Yang's newsletter where he mentioned using an AI tool to send him a daily text message with workout goals. I realized I wanted a daily text message but not about workout goals - about life goals. I had been playing around with having AI build some goal related affirmations but I wasn't very satisfied with the results. I wanted messages that helped me see how my days were lined up with my long term goals.

I already had spent the time documenting my BHAGs, my word of the year and my Q2 goals. What if I had AI take all that in AND look at my calendar and help me see how my time was reflecting my goals, to nudge me to think about what might be missing. I wanted this to spin to the positive as I know from experience that I will naturally do better/more when I feel good about what I already do.

I wanted AI to help me tell the right story to myself about my priorities and I suspected the rest would take care of itself. I edited out specifics but this is generally what it looks like (you will notice it promotes Open Brain back to me. Very meta.)

THIS was a huge unlock for me and it was somewhat orthogonal to my Open Brain project. What it did, was help me re-think what my full context could enable. I could use my context to do the most important part of my energy management -- help me frame the right story. I wanted to create a generative cycle for my energy and this has been a huge step.

Taking that idea one step further, I decided to test the judgement - could it connect the dots of all the things effectively. I had a few things in my inbox that I was struggling to respond to - I realized that it was a massive stack overflow (this not this) challenge for me.

I needed to decide if a) I wanted to make space to take something on and b) was that even possible, based on other commitments and then c) do all the shifts required to make it happen. When I stopped to think for a minute I realized the problem was materially blocked by the first part. I made this pattern discovery when I asked my Truth Council for help deciding how to navigate a conflict between two things I wanted to do. They answered immediately (and unambiguously) making me realize the problem was uniquely hard for me to do but that the answer was often obvious to others. I decided to give this a go and build a "should I do this" tool.

Again, making AI look at my goals and priorities AND my calendar to tell me if it was a good idea. I asked it to give me the rationale and let me know what I would have to reschedule to make it happen.

It was so helpful! Again, not just in solving the specific use case but in building confidence for how to articulate the abstraction of my calendaring anxiety to make AI a useful tool.

Next, I decided to take another half step into the scheduling problem. Help me find a few calendar slots to offer that a) doesn't disrupt existing priorities, b) respects the timezone of the person I am attempting to meet with and c) requires the least shifts to existing plans. Then I have it give me a rationale for the choices, and put calendar holds for the slots so I don't schedule over times I have given someone.

and here is where Open Brain comes back into the story...

I had AI put context information from Open Brain on the person, the topic(s) and other related threads I am holding related to the meeting into the calendar hold it created. This is all work I would do manually and it was taking a fair bit of time. I also had it write up a summary of what action items I need to do when I got back a confirmation so I don't forget to clean up the unnecessary holds.

There are a few things this journey has taught me in addition to general proficiency with Claude.

• Building trust with an agent is very similar with building trust with an EA. The hard part is articulating the vibe bit not finding calendar slots. Managing energy and intention requires both context and judgement. Building up Open Brain helped me create that for my agents. Understanding and articulating the boundary conditions was as hard as defining the task (but more important). Signaling my broader goals and aspirations helped me backstop the context with implicit success criteria.

• Going slow with intention was important - more for my nervous system than for my technical ability. Learning how and what to let go and what kinds of evals I needed to feel confident in the result, was cognitively challenging but iterating through it helped me break it down and tackle it with confidence. It also helped me decide what I was comfortable giving full control to vs. what I still wanted to manage.

• Maintenance is non-trivial. Catching Claude's shortcuts and misses is a daily reality right now. My newsletter ingestion process is a battle of wills, and if I'm honest it's unclear who is winning (see today's conversations)

after fixing this bug I realized that I should check before re-running with the logic fixed

• I've shifted how I think about the value of context from doing this exercise. Of course it's easy to see (feel) the compounding benefit from past work - that is lovely and feels a lot like the satisfaction I would get when improving and optimizing overly complicated processes as an operator. It gives similar value to having a head of operations and/or a chief of staff, but it does a few things that are new as well.

• One of the biggest values it gives me is that it can keep up. If I'm honest, because I read and think so much, I spend a lot of time optimizing my interactions for the audience. This is important and I don't mind, but it's truly amazing to have something that doesn't require me to sub-slice my ideas or requests but can scale to meet my broader cognitive load. This has never happened before and it is both helpful and valuable. I can feel a material weight lifted not just cognitive but emotional. My Open Brain is able to offload both cognitive and emotional labor and anyone who has a lot of that understands how heavy it can be.

• I have learned the trick of not [just] looking to automate the tedious and cumbersome tasks but instead to focus on those things that are emotional burdens. Emotions are data and most things that I feel stressed about have a root cause underneath. Getting curious there, helps me discover places I want to invest next. Leaning into using AI to help get me unstuck and motivated helps with my energy and any energy improvement pays back with interest.

All of this is very nerdy and specific but also likely re-usable. When people use AI to give them a daily summary they are doing something similar - taking the hard bit (getting started) and removing friction. For me the hard bit is different and thus I need a different focus - my hard bit is the deep need to be making progress. Getting agents to help me see the progress I am making has been very motivating - dare I say joyful.

The value of this moment is the opportunity to explore new ways of working. The best way to start that journey is to think about your own work habits and energy systems and ask yourself where are the generative elements for you and your team. How do you let go of what you thought was impossible (or absurd) and see what you discover. While not everyone has an Amy to nudge them forward ,I believe that all of us could benefit from a thought partner to brainstorm who we want to be in the future and help us build a path to get there inclusive of a daily dose of encouragement tuned to your specific needs.

Coming back full circle to the beginning of this post, I have also uncovered a repeating pattern in Amy's Jedi mind tricks. Amy incepts ideas precisely by allowing me to tell her that they can't work. I fall for it every time and even more interestingly, I convince myself it was my own idea when I finally come around to do what she told me.

1. Amy: Meg, you should get Claude to do your calendar stuff for you b/c you are always stressed about that

2. Meg: That would never work because I’m too much of a control freak and there is too many things that freak me out to number

3. Amy: ok

4. Meg: 3 weeks later - AMY I just vibe coded a thing to help me resolve my calendar anxiety…

5. Amy: That’s a great idea Meg - way to go 😬