2025-0116M-0830 Austin, TX
This morning I was started my day listening to a Robert W.B. Love podcast with Dr. Dale Bredesen about Alzheimer’s Disease.
Actually, my day started earlier with a short clip from this Nate B Jones (@nate.b.jones) video about AI having a half-trillion dollar memory problem:
Both of these topics are on my mind right now as I struggle myself with memory and focus. So, starting my focus practice this morning by not watching these videos now and instead writing.
One of my first thoughts on watching the AI half-trillion dollar memory problem was to buy more Micron Technology ($MU). I was apparently not the only one with this idea. The market cap on open rose over 9% from $100B to $109B. I bought some more anyway to remind me of this moment. My stock portfolio of mostly technology stocks is up about 58% over the last three years. It looks now like it’s either about to fall off the cliff are shoot higher. Learning to hold both of these conflicting possibilities and be with them is my practice.
Yesterday, I spent most of the day in silence. I ecstatic danced with a few hundred people and shared a few hugs. I talked to brother B for a while, breaking my silence for the day. I also talked to a couple other friends. Otherwise, I stayed focused on trying to solve a programming bug in a toy program I created for my learning.
I hesitate to go into it. But then, maybe it will be helpful for anyone who reads my blogs to help them sleep. I’ll say this much. I had spent over 12 hours the prior day pleading with OpenAI ChatGPT-4o to help me fix my code. When I finally put AI to bed and wrote the code from scratch myself, the code worked. Of course, as all programmers know, then I uncovered a new bug! But, at least I had the fundamentals working. The new bug seems to be beyond my current abilities to debug. Actually, I have the abilities to learn, so that statement is not really true. I feel like I just finished a 2048 piece puzzle and want to wait before starting the 8192 piece puzzle.
Going Deeper: Here is an analogy for the puzzlers on my struggle to add tiling to my matrix multiplication code. My current code is a naive approach. A family of four with each family member working on their own corner of the puzzle. Everyone is grabbing a random puzzle piece from a single shared box of pieces. They check if it fits in their corner and either place the piece or put it back. Everyone can work independently but it wastes times because many pieces are grabbed repeatedly by different people. The new code I am working on is a tiling approach. The analogy for the new code is the family of four work together. They all focus on a smaller section of the puzzle and first find the pieces for that section. Then they finish that section of the puzzle and repeat for new sections. This is more efficient!
surfing the universe 