within 2 weeks: 3 thiel interviews, 2 critical vc interviews w/ them trying to write a 200k check immediately which i refused, 1 attempt of a multimillion dollar company to either hire me or consider me a future threat, 1 meeting with sun and b to align vision and combine forces for venture scaling, 1 meeting w/ my roommate's dad in PE (so somewhat adjacent to VCs?), 1 consulation w/ an IP attorney, the pending resumption of z as told by baylor adams himself, 1 meeting with a potential mlops hire (suzanne), and 1 meeting w/ kai (a thiel fellow, hs dropout w/ 3 ongoing biotech ventures, putting words for me voluntarily) and wants to bring me to his network of ppl trying to defeat death... things are coming together in some mysterious way that i can't clearly see, but can also see exactly when fallout would take place; everything is happening all at once, yet nothing is happening.
i sat down to myself and analyzed every situation. and here are 3 observations:
-
the experiment goes on. i don't drop out because i have to prove my thesis. to accept money is to accept a legal obligation and carry the wieght of building something that might not work. but right now it requires no financial cost for me to build. the experiment's result is the only concrete thing everyone can understand. ppl's opinions and trust in me will grow or wane, but the result, produced by the data, protected legally by patents, will be the constant real forces acting for me, never against me.
-
critical is backing engram, and engram is doing exactly what im doing from the opposite starting point: robotics. they have huge funding from hf0 and the like, but they lack specialized patented hardware that went through years of research; they despise academia but also desperately tries to hire from academia because they are all 20 year olds and have zero organoid experience; they believe agentic manufacturing will be the endpoint, but agentic manufacturing powered by LLMs will be always prone to mistakes and worse than a systematic, fixed system (post optimization by conducting trials and errors with generative AI ofc when setting up a semi-automated or fully-automated manufacturing infra/pipeline, per Suzanne) that is fully reproducible.
-
from the prev. datapt we can tell what profile the typical SF tech industry views AI integration with biology. they view AI is the end all be all; governments and academia have zero exposure. they are at the opposite spectrum of everything. to have every proven needs a middle ground, where my audience (that would fund me) understands how AI functions, and its most appropriate role to play within industrial biomanufacturing. indiebio (they have changed their name a few days ago to SOSV SF/NY) would be that sweet middle ground. they care about the science, innovation and commercialization. something neither side would fully agree on. academia/grant sources wants "prove everything first, fund maybe never;" and generic tech vcs go "fund everything first, prove everything later." a healthy mindset is in the middle. and i still think indiebio sits in that sweet spot. any research effort directed specifically to curing aging should come after the reassurance that money won't be an issue.
so i wait. i wait for the organoids to grow. and speaking of the organoids, we are at day 32, 3 days away from the period where biological differentiation for fox-g1 happens invisibly (until, hopefully, the model sees something we don't). sun and beatriz expanded the scope of ground truth collection to not just fox-g1 but two other biomarkers with upstream calcium imaging to make this experiment more bulletproof. several organoids show signs of decline or oscillations after not supplying any growth factors. we will see what happens by day 42. the argument to maintain instead of discarding underperforming ones go beyond that of the "being friendly to the model" but just that of the concern of data sample preservations; to instinctively exclude orange wells is an scaled industrial question that this PoC isn't by itself aiming for, and the term "orange" tied to its mathematical definition isn't grounded in biological truth. and that ground to stand on to conduct refinements require truth within biology. to discard is to see declines so drastic that they fall below 40% of the peak after perhaps 2 weeks, physical disintegration, or technical attrition. anyways... is this situation catastrophic? claude code says no. that is some reassurance but let's see what happens anyway.
“事以密成，语以泄败。”
“宁缺毋滥。”
if engram doesn't pivot, it will be an all out war to dominate brain organoid biomanufacturing in the most literal sense. a war that needs to be fought cautiously, patiently and strategically from our end. the market is the true voice, often slower than the FOMO VCs, often faster than the government. use any pragmatic means necessary to achieve monopoly in this space. and the greatest ally, time, is on my side. from what i have heard, even though their founder has inflated himself so much when trying to either recruit or extract info from me, he has no iPSCs. he has no organoids. and both of us know that organoids need a lot of time to grow. even if they started growing organoids today w/ significantly more money, sample size and/or a strategy better than mine, they cannot change the speed of organoid growth, and i will always be at least 32 days ahead. he doens't understand AI, and it seems like everyone on his team aren't too smart to realize there's levels to this. i have real data. they have divisions of ppl trying to conquer OI, drug discovery, generative + agentic integration w/ robotics (which is the only thing they are good at but is not a real moat), and they are trying to engineer individualized tracking without any specialized hardware because they never had the time to build one.
the hour after the call i had w/ their founder is the moment my friend reminded me that they reached me, not the other way around. they never got interviewed by thiel fellowship, and they pivoted so much they have produced zero results from any PoC's besides raising rounds. raising without a concrete, focused goal in sight is thinking too temporally short. but this is a long game. being a monopoly is a long game and it requires dedication on dull tasks, a single division (so no division at all), and pragmatic thinking, not trying everything at once.
"I fear not the man who has practiced 10,000 kicks once, but I fear the man who has practiced one kick 10,000 times." -- Bruce Lee
keep going. thiel is a great add-on. but the original script never had its existence in consideration. and its addition into the script will not change the immediate outcome i need: the accumulation of all the data, and the model that was built from it to prove that my thesis is correct. everything compounds from there.