I'm currently situated somewhere next to UMass Amherst via Airbnb, where I can regularly walk to their dining hall, Worcester, as well as the LSL building, and being not so remote from HistoSpring so I can visit there when needed.
Here is the state of everything:
- Since block 21 didn't go through this full six-panel multiplex, the likely outcome of that staining session on block 21, which corresponds to G2, G3, and G4, is that these samples will probably not be in the final data set. We have halved our total possible fit counts. And since we are not using covariates by default, the total fit possibilities halves again and we are down to the original:
- 1,974 configs
- 25 outer folds/repeats × 5 inner folds × 600 hyperparameter settings = 75,000 fits per config
- 1,974 × 75,000 = 148,050,000 total fits
- Conversations on the specific staining protocol and logistics with sectioning, staining, imaging, and quantifications are in discussion on three fronts:
- Beatriz/Sun Lab
- HistoSpring mainly
- and Jim Chambers of the LSL Microscopy core.
- I have switched from Claude code to Codex and am using the Codex macOS app (w/ GPT Plus subscription) as my super app and Obsidian as its grand elaborative memory system. GPT 5.5 is pretty good only in this harness, not even on the web app. Its automation is better than both OpenClaw and Claude Cowork, and I'm very impressed by how many tools are usable and automatable. Further testing is needed to determine whether it can fully replace Claude. Per Dan's suggestion, if I get the ChatGPT Pro plan ($200/month), it is essentially unlimited tokens, which is equivalent to burning $10K on OpenAI's API.
- I live in a nice, quiet house next to a chicken coop. My brain is very washed out and needs a hard reset. I want that reset to come quickly. I don't know how to restore my short-term memory and recover. Bevel says that, with 32.5% confidence, my current biological age is 22 years old, not 19.4 years old. This is very worrisome, and I'm trying to adjust everything in my schedule, my sleep schedule, my dietary schedule, and my workload. I should enjoy some walks in nature and visit the chickens every day. The male chickens are somewhat threatened by my presence, but the female ones seem to be purring. Seems like gender differences are very prevalent even in other areas of the animal kingdom.
- I have a meeting tomorrow with Mohan Iyer and his team. I should prepare for an hour sometime before tomorrow's meeting, and then take a break to hop on the call. My confidence has been gradually diminishing ever since the Thiel Fellowship rejected me. Dan said my confidence isn't high enough because my charisma isn't either. Not sure if he means by appearance or just mental projections on my end. I have to reflect on that. He did say that "confidence that comes out of thin air is worth far less than confidence that comes from a produced outcome" because one is ego and the other is real confidence.
- I will think about moving out or extending my stay in this place if necessary. That decision will probably be taken sometime around Friday of this week, after my call with the people at HistoSpring. It seems like the large remainder of the project is about how to stain, section, image, and quantify everything, depending on people at HistoSpring and Jim Chambers at LSL. I can see that the most likely scenario is that sectioning and staining will occur in HistoSpring, and everything will be transferred back to UMass Amherst for Jim Chambers to assist with both imaging and quantification. There seems to be no need to route Beatriz in for the remaining 24 blocks. We're looking at the final data set size being around 72 organoids (24 x 3), counting down.